GOD-LEVEL PUBLIC COMPANY FINANCIAL ANALYST JOB GUIDE Zero-to-expert institutional public-equity research operating manual Version 3.0 | 99+ Mastery Rebuild | September 18, 2026 > Purpose: convert raw public information into a defensible, auditable investment view through primary-source verification, accounting reconstruction, driver-based modeling, valuation, falsification, risk control, decision journaling, and continuous review. > This edition consolidates universal controls, deepens technical execution, adds mastery laboratories and casebooks, and requires reproducible calculations, source provenance, explicit edge-case handling, and reviewer-ready completion tests. # How to Use This Manual - Treat each module as a job procedure. Do not merely read it. Produce the required evidence pack, calculations, model changes, and decision record. - Use primary sources first. Secondary research is discovery and triangulation, not a silent replacement for filings, contracts, audited statements, regulator data, and issuer disclosures. - Keep three layers separate: reported fact, analyst calculation, and analyst judgment. Management statements remain claims until corroborated. - Every forecast must have a causal driver, every valuation must have an operating model underneath it, every risk must have a transmission mechanism, and every conclusion must have a falsification condition. - Rules and standards change. Verify the current SEC, FASB, PCAOB, IFRS, exchange, and sector-regulator source before relying on a specific requirement. ## Universal Analyst Quality Standard - Question: write the decision question in one sentence and identify the variables capable of changing value or risk. - Evidence: archive dated primary sources and exact versions. Use secondary sources only with provenance and cross-checks. - Definitions: reconcile KPI definitions across company history and peers before comparing values. - Accounting: tie material conclusions through income statement, balance sheet, cash flow, and footnotes where applicable. - Economics: forecast economic drivers, not template line items. Separate structural economics from cycle, timing, accounting, and one-time effects. - Uncertainty: use ranges, operating scenarios, and sensitivities. Never use a point estimate to hide uncertainty. - Falsification: write the strongest alternative explanation and evidence that would change the view before the next datapoint arrives. - Audit trail: preserve source log, assumption register, model-change log, and decision journal so another analyst can reproduce the work. ## Evidence Hierarchy | Tier | Preferred evidence | Primary use | Main failure mode | | --- | --- | --- | --- | | 1 | Regulatory filings, audited statements, contracts, debt documents, regulator/exchange rules | Core factual foundation | Stale document, missing exhibit, or misread definition | | 2 | Issuer earnings materials, calls, investor-day materials, operating data | Management intent, KPI detail, guidance | Selective framing and non-GAAP emphasis | | 3 | Competitor filings, government data, industry bodies, customer/supplier evidence | Triangulation and market structure | Different definitions or populations | | 4 | Expert calls, channel checks, hiring, web/app/product data | Leading signals and hypothesis testing | Sampling bias, legality/compliance, weak causality | | 5 | Sell-side, media, forums, social | Discovery and sentiment context | Circular sourcing and narrative momentum | ## Master 12-Step Research Workflow 1. Triage - Define the decision question and whether the issue can change intrinsic value, probability, or risk. 1. Primary sources - Pull filings, exhibits, debt documents, presentations, transcripts, ownership forms, and regulator materials. 1. Historical reconstruction - Build clean statements, KPIs, segments, capital structure, and definition history. 1. Business model - Map customer, product, unit economics, pricing, capital, and bottlenecks. 1. Industry - Map value chain, profit pools, market size, competition, regulation, technology, substitutes, and cycle. 1. Forensics - Review accruals, non-GAAP, estimates, working capital, auditor language, related parties, and disclosure drift. 1. Model - Forecast operating drivers, integrated statements, cash, financing, share count, and scenarios. 1. Valuation - Use intrinsic value, relative value, reverse expectations, and scenario analysis appropriate to the economics. 1. Variant view - State what market expectations appear embedded and what evidence could support a different outcome. 1. Risks - Quantify downside, liquidity, balance-sheet resilience, and thesis-break conditions. 1. Memo - Write the conclusion, evidence, assumptions, catalysts, valuation, risks, and disconfirming facts. 1. Monitor - Create dated triggers, KPI checks, filing alerts, estimate-error tracking, and post-event update rules. # Navigation Map ## PART I - ANALYST OPERATING SYSTEM - 1. Mission, Ethics, and the Analyst Charter - Define the analyst's mission, evidence standard, decision rights, escalation paths, and non-negotiable ethical rules. - 2. The Research Audit Trail - Create a source log, assumption register, model-change log, and decision journal that another analyst can reproduce. - 3. Daily, Weekly, Monthly, and Quarterly Cadence - Build a recurring work system that prevents important filings, estimate changes, and thesis drift from being missed. - 4. Hypothesis-Driven Research - Turn vague curiosity into explicit hypotheses, falsification tests, and prioritized research questions. - 5. Time Allocation and Research ROI - Allocate analyst time to the questions most likely to change intrinsic value, probability, or risk. ## PART II - SOURCE INTELLIGENCE AND FILINGS - 6. EDGAR Mastery and Filing Retrieval - Use SEC filing search, accession numbers, exhibits, XBRL data, and amendment history efficiently. - 7. 10-K Deep Read Procedure - Extract business economics, accounting policies, risks, obligations, segment data, and footnote detail from annual reports. - 8. 10-Q and Interim Update Procedure - Reconcile interim results, seasonal patterns, working capital, guidance changes, and emerging risks. - 9. 8-K and Event Filing Triage - Classify current reports by materiality, financial impact, transaction type, and required model action. - 10. Proxy Statement and Governance Mining - Use proxy materials to analyze incentives, ownership, related-party issues, board structure, and pay design. ## PART III - ACCOUNTING FOUNDATION - 11. Three-Statement Mastery - Connect income statement, balance sheet, and cash flow mechanics before attempting forecast modeling. - 12. Revenue Recognition Analysis - Understand contract economics, timing, variable consideration, principal-agent issues, and revenue quality. - 13. Cost Structure and Margin Architecture - Separate variable, fixed, semi-fixed, pass-through, and step-function costs to model operating leverage. - 14. Working Capital Mechanics - Analyze receivables, inventory, payables, deferred revenue, contract assets, and cash conversion. - 15. Capital Expenditures, Depreciation, and Asset Intensity - Distinguish maintenance from growth investment and reconcile capex to productive capacity. ## PART IV - ADVANCED ACCOUNTING - 16. Stock-Based Compensation - Analyze dilution, economic cost, tax effects, share count, buyback offset, and adjusted-metric treatment. - 17. Leases and Off-Balance-Sheet Commitments - Reconstruct lease economics, fixed commitments, and debt-like obligations. - 18. Goodwill, Intangibles, and Impairment - Assess acquisition accounting, amortization, impairment risk, and return on acquired capital. - 19. Taxes and Deferred Taxes - Model cash taxes, effective tax rates, NOLs, valuation allowances, jurisdiction mix, and one-time items. - 20. Pensions and Postretirement Obligations - Analyze funded status, discount rates, expected returns, cash contributions, and hidden leverage. ## PART V - FORENSIC ACCOUNTING - 21. Earnings Quality Framework - Score how much reported profit is supported by cash generation, repeatable economics, and conservative accounting. - 22. Accrual and Cash Conversion Tests - Use accrual ratios, CFO versus net income, working-capital patterns, and multi-year normalization. - 23. Non-GAAP Reconstruction - Rebuild management adjustments and classify recurring, discretionary, acquisition-related, and economically real costs. - 24. Fraud and Manipulation Red Flags - Use professional skepticism, cross-statement inconsistencies, unusual transactions, and disclosure drift. - 25. Accounting Estimates and Management Bias - Identify estimates with subjective assumptions, sensitivity, unobservable inputs, and asymmetric incentives. ## PART VI - BUSINESS QUALITY - 26. Business Model Deconstruction - Reduce the company to customers, value proposition, unit economics, pricing, cost structure, and capital requirements. - 27. Unit Economics and Cohort Thinking - Translate growth into acquisition economics, retention, payback, contribution margin, and lifetime value. - 28. Pricing Power and Price-Volume-Mix - Separate real pricing power from inflation pass-through, mix, channel shift, or temporary shortages. - 29. Recurring Revenue and Retention Quality - Measure renewal behavior, churn, net retention, contract durability, and hidden re-acquisition costs. - 30. Capital Intensity and Reinvestment Runway - Estimate incremental returns on capital, reinvestment capacity, and limits to compounding. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS - 31. Industry Structure Mapping - Map value chains, profit pools, bottlenecks, substitutes, regulators, and bargaining power. - 32. Market Sizing and TAM Discipline - Build bottom-up market sizes and prevent promotional TAM estimates from contaminating valuation. - 33. Competitive Advantage and Moat Testing - Test switching costs, network effects, scale, brand, cost advantage, regulation, and data advantages. - 34. Competitor Benchmarking - Normalize peers across accounting, KPIs, end markets, geography, and capital structure. - 35. Disruption and Technology S-Curves - Analyze adoption curves, cost declines, standards, infrastructure constraints, and incumbent response. ## PART VIII - MANAGEMENT AND GOVERNANCE - 36. Management Quality Assessment - Evaluate capital allocation, operational execution, candor, consistency, and incentives using evidence rather than charisma. - 37. Capital Allocation Scorecard - Evaluate reinvestment, M&A, buybacks, dividends, debt policy, and balance-sheet optionality. - 38. Compensation and Incentive Analysis - Map executive pay metrics to behaviors, accounting choices, time horizons, and shareholder outcomes. - 39. Insider Ownership and Transactions - Interpret insider ownership, buying, selling, grants, 10b5-1 plans, and dilution in context. - 40. Board and Governance Risk - Evaluate independence, tenure, expertise, related parties, committee structure, and control weaknesses. ## PART IX - MODEL BUILDING - 41. Model Architecture and Standards - Build models that are transparent, auditable, modular, scenario-ready, and resistant to hard-coded errors. - 42. Historical Data Normalization - Recast reported data into consistent periods, segments, KPIs, and economic definitions. - 43. Revenue Forecasting by Driver - Forecast units, customers, capacity, utilization, price, mix, or other business-specific drivers. - 44. Margin and Cost Forecasting - Model contribution margin, fixed-cost absorption, operating leverage, and cost actions. - 45. Balance Sheet and Cash Flow Forecasting - Forecast working capital, capex, financing, share count, and cash without circular mistakes. ## PART X - VALUATION - 46. DCF from First Principles - Value the operating asset base using explicit cash flows, reinvestment, terminal economics, and transparent discounting. - 47. Multiples and Relative Valuation - Choose metrics that fit economic reality, normalize peer differences, and avoid mechanically applying averages. - 48. Reverse DCF and Expectations Investing - Infer the growth, margin, capital intensity, and duration assumptions embedded in the stock price. - 49. Sum-of-the-Parts and Conglomerate Analysis - Value distinct businesses separately and account for central costs, taxes, minority interests, and capital structure. - 50. Scenario, Sensitivity, and Probability Weighting - Replace false precision with explicit bull, base, bear, stress, and thesis-break cases. ## PART XI - CATALYSTS AND EVENTS - 51. Earnings Preview and Post-Earnings Procedure - Define expectations, key debate variables, scenario reactions, and model update rules before the print. - 52. Guidance Analysis and Estimate Revisions - Translate management guidance into implied quarterly paths, margins, cash flow, and consensus risk. - 53. M&A and Strategic Transactions - Analyze purchase price, synergies, financing, accretion, integration, antitrust, and value transfer. - 54. Activism, Spin-Offs, and Restructurings - Evaluate separation economics, stranded costs, incentive resets, and capital structure changes. - 55. Bankruptcy, Distress, and Liquidity Events - Analyze runway, covenants, collateral, priority, recovery, dilution, and restructuring pathways. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK - 56. Expert Calls and Channel Checks - Design compliant interviews that test hypotheses without soliciting material nonpublic information. - 57. Customer and Supplier Diligence - Triangulate demand, pricing, product quality, switching behavior, inventory, and competitive share. - 58. Web, App, Hiring, and Product Data - Use digital signals as noisy indicators that require baselines, controls, and definition discipline. - 59. Pricing and Inventory Scraping - Track availability, discounting, lead times, SKU breadth, and channel inventory without overfitting. - 60. Alternative Data Validation - Backtest data against reported outcomes and measure false positives, revisions, survivorship, and coverage bias. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING - 61. Risk Register and Thesis-Break Conditions - Maintain explicit operational, financial, regulatory, competitive, valuation, and governance risks. - 62. Balance Sheet and Liquidity Stress Testing - Stress cash burn, maturities, covenants, refinancing rates, collateral, and working-capital shocks. - 63. Position Sizing Inputs for Research Teams - Translate conviction, downside, liquidity, catalyst path, and correlation into research inputs without confusing analysis with mandate. - 64. Decision Journaling and Pre-Mortems - Record the decision, evidence, disconfirming facts, expected path, and what would change the view. - 65. Post-Mortems and Error Taxonomy - Separate thesis errors, sizing errors, timing errors, data errors, process errors, and luck. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY - 66. Investment Memo Writing - Write concise, decision-useful research that separates facts, assumptions, variant views, valuation, and risks. - 67. Charts, Tables, and Evidence Design - Use visuals to reveal relationships rather than decorate the narrative. - 68. Investment Committee Defense - Prepare for adversarial questions, alternative explanations, and explicit uncertainty. - 69. Analyst Training and Deliberate Practice - Build a 12-month progression from filings and accounting to modeling, industry expertise, and judgment. - 70. Research System Automation and AI Assistance - Use automation and AI for retrieval, extraction, QA, and scenario generation while keeping human verification and accountability. ## PART XV - SECTOR PLAYBOOKS - 71. Software and SaaS Analyst Playbook - 72. Semiconductors Analyst Playbook - 73. AI Accelerators and Compute Analyst Playbook - 74. Cloud and Data Centers Analyst Playbook - 75. Industrial Machinery Analyst Playbook - 76. Aerospace and Defense Analyst Playbook - 77. Airlines Analyst Playbook - 78. Automotive OEMs Analyst Playbook - 79. EV and Battery Manufacturers Analyst Playbook - 80. BESS Electrical Balance-of-System Analyst Playbook - 81. Grid Equipment and Electrification Analyst Playbook - 82. Electric Utilities Analyst Playbook - 83. Renewable Developers Analyst Playbook - 84. Oil and Gas E&P Analyst Playbook - 85. Midstream Energy Analyst Playbook - 86. Refiners Analyst Playbook - 87. Chemicals Analyst Playbook - 88. Mining and Metals Analyst Playbook - 89. Banks Analyst Playbook - 90. Property and Casualty Insurance Analyst Playbook - 91. Life Insurance Analyst Playbook - 92. Asset Managers and Brokers Analyst Playbook - 93. Payments and Fintech Analyst Playbook - 94. REITs Analyst Playbook - 95. Homebuilders Analyst Playbook - 96. Restaurants Analyst Playbook - 97. Retail Analyst Playbook - 98. Consumer Packaged Goods Analyst Playbook - 99. Pharmaceuticals Analyst Playbook - 100. Biotechnology Analyst Playbook - 101. Medical Devices Analyst Playbook - 102. Managed Care Analyst Playbook - 103. Telecom Analyst Playbook - 104. Internet Platforms and Marketplaces Analyst Playbook - 105. Cybersecurity Analyst Playbook - 106. Railroads and Logistics Analyst Playbook ## PART I - ANALYST OPERATING SYSTEM | MODULE 001 # Mission, Ethics, and the Analyst Charter > Mission. Define the analyst's mission, evidence standard, decision rights, escalation paths, and non-negotiable ethical rules. ## Decision output Objective: Define the analyst's mission, evidence standard, decision rights, escalation paths, and non-negotiable ethical rules. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Write the mandate in one sentence: what securities/markets are in scope, what decisions the analyst supports, what time horizon matters, and what the analyst is not authorized to do. 1. Define evidence classes and minimum proof standards before research begins. Separate public fact, management claim, analyst calculation, external estimate, and judgment in every material work product. 1. Establish non-negotiable legal/compliance rules for MNPI, selective disclosure, expert networks, personal trading, conflicts, gifts, source licensing, and record retention. Escalate ambiguity rather than improvising. 1. Define decision rights and review gates: who may change model assumptions, publish a view, approve alternative data, or communicate externally. 1. Set materiality and stop rules so research effort is proportional to potential value/risk impact and does not continue merely because more information exists. 1. Write the analyst charter as a signed checklist that can be audited after a mistake or conflict. ## Required evidence and model bridge - Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in mission, ethics, and the analyst charter. - For each key concept - mandate boundaries, MNPI stop rules, conflicts, personal trading, evidence standards, decision rights - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as mission, ethics, and the analyst charter may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | evidence coverage | Value-weighted % of material thesis claims supported by current primary or independently corroborated evidence; weight each claim by estimated valuation/risk impact. | evidence coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | unresolved-risk PV | Probability-weighted present value of unresolved adverse outcomes: sum of scenario loss × probability, discounted to the valuation date. | unresolved-risk PV: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | source freshness | Age of the evidence supporting each material claim, measured in days from source date to analysis date; report median and oldest material source. | source freshness: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | signoff completeness | % of required research, model, compliance, and senior-review checkpoints completed before publication or decision. | signoff completeness: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | ## Worked application > Case: supplier accidentally hints at an unreleased production shortfall. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For mission, ethics, and the analyst charter, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through mandate boundaries, MNPI stop rules, conflicts, personal trading, then identify which link is directly observed and which link remains an assumption. - Calculate evidence coverage, unresolved-risk PV, source freshness, signoff completeness from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for mission, ethics, and the analyst charter: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: design compliance into the workflow before research begins, not as a final gate. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the mission, ethics, and the analyst charter conclusion. ## Failure tests - FAIL if mandate boundaries cannot be defined and reproduced from the source pack. - FAIL if decision rights, evidence standards, MNPI boundaries, or escalation rules are undefined, contradictory, or cannot be applied to a realistic research conflict. - FAIL if the mission, ethics, and the analyst charter conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the mission, ethics, and the analyst charter conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the mission, ethics, and the analyst charter conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART I - ANALYST OPERATING SYSTEM | MODULE 002 # The Research Audit Trail > Mission. Create a source log, assumption register, model-change log, and decision journal that another analyst can reproduce. ## Decision output Objective: Create a source log, assumption register, model-change log, and decision journal that another analyst can reproduce. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Create a source log containing issuer, document, form/type, filing or publication date, period covered, accession/URL, exact section/table, retrieved timestamp, and version. 1. Create an assumption register with assumption name, historical basis, source, range, confidence, owner, valuation sensitivity, and next review date. 1. Maintain a model-change log that records prior value/formula, new value/formula, reason, source, affected outputs, reviewer, and timestamp. 1. Maintain a decision journal that freezes the conclusion, expected path, key uncertainties, disconfirming evidence, and thesis-break conditions before later outcomes are known. 1. Version raw sources, extracted data, model files, and memos so a reviewer can reproduce any past conclusion. Never silently replace an amended filing or revised dataset. 1. At each review, test whether the audit trail is sufficient for another analyst to rebuild the material conclusion without asking the original author. ## Required evidence and model bridge - Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in the research audit trail. - For each key concept - source IDs, accession numbers, raw versus transformed data, assumption taxonomy, model lineage, decision journal - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as the research audit trail may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | claim traceability | % of material factual claims that link to a dated source, exact page/section, and model or memo location. | claim traceability: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | model-change coverage | % of material model changes with a documented source, rationale, date, author, and affected output/sensitivity. | model-change coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | reproduction time | Elapsed analyst time for an independent reviewer to reproduce the key conclusion from the archived sources, assumptions, and model. | reproduction time: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | ## Worked application > Case: management changes a KPI definition and recasts only part of history. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For the research audit trail, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through source IDs, accession numbers, raw versus transformed data, assumption taxonomy, then identify which link is directly observed and which link remains an assumption. - Calculate claim traceability, model-change coverage, reproduction time from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for the research audit trail: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: preserve rejected hypotheses and prior model versions so hindsight cannot rewrite the record. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the the research audit trail conclusion. ## Failure tests - FAIL if source IDs cannot be defined and reproduced from the source pack. - FAIL if a reviewer cannot reproduce a material claim, assumption, model change, and decision from dated source/version records without asking the original analyst. - FAIL if the the research audit trail conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the the research audit trail conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the the research audit trail conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART I - ANALYST OPERATING SYSTEM | MODULE 003 # Daily, Weekly, Monthly, and Quarterly Cadence > Mission. Build a recurring work system that prevents important filings, estimate changes, and thesis drift from being missed. ## Decision output Objective: Build a recurring work system that prevents important filings, estimate changes, and thesis drift from being missed. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Daily: triage new 8-Ks, filings, guidance, material news, regulator actions, price/volume anomalies, and thesis-break alerts. Record only items that can affect assumptions, risk, or evidence quality. 1. Weekly: review estimate changes, peer developments, channel/alternative-data signals, valuation drift, upcoming catalysts, and unresolved source questions. 1. Monthly: refresh industry dashboards, risk register, liquidity/capital-structure watch, competitor benchmarking, and major assumption confidence. 1. Quarterly: freeze the pre-earnings model, reconcile the reported period, update historicals, perform forecast-error attribution, revise valuation, and write a short post-mortem. 1. Event-driven: immediately elevate amendments, restatements, auditor/control issues, financing, M&A, litigation, covenant events, management departures, or regulatory changes that can alter the thesis. 1. Use a coverage calendar with named owner, due date, completion evidence, and escalation rule so cadence does not depend on memory. ## Required evidence and model bridge - Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in daily, weekly, monthly, and quarterly cadence. - For each key concept - event triage, filing alerts, thesis reviews, estimate rolls, open-question queue, stale-assumption control - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as daily, weekly, monthly, and quarterly cadence may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | alert latency | Minutes or hours from a material filing/event becoming public to triage, classification, and assignment of required model action. | alert latency: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | stale assumption count | Count of material assumptions whose source date or review date exceeds the predefined freshness threshold for that variable. | stale assumption count: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | research backlog aging | Age distribution of open material research questions; track median days open, 90th percentile, and valuation-weighted overdue items. | research backlog aging: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | ## Worked application > Case: a competitor changes pricing six weeks before the covered company reports. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For daily, weekly, monthly, and quarterly cadence, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through event triage, filing alerts, thesis reviews, estimate rolls, then identify which link is directly observed and which link remains an assumption. - Calculate alert latency, stale assumption count, research backlog aging from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for daily, weekly, monthly, and quarterly cadence: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: treat cadence as an operating system with exception alerts, not a calendar checklist. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the daily, weekly, monthly, and quarterly cadence conclusion. ## Failure tests - FAIL if event triage cannot be defined and reproduced from the source pack. - FAIL if a material filing, guidance change, thesis-break signal, or stale assumption can occur without an assigned owner, alert path, and review deadline. - FAIL if the daily, weekly, monthly, and quarterly cadence conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the daily, weekly, monthly, and quarterly cadence conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the daily, weekly, monthly, and quarterly cadence conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART I - ANALYST OPERATING SYSTEM | MODULE 004 # Hypothesis-Driven Research > Mission. Turn vague curiosity into explicit hypotheses, falsification tests, and prioritized research questions. ## Decision output Objective: Turn vague curiosity into explicit hypotheses, falsification tests, and prioritized research questions. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Convert the decision question into explicit competing hypotheses, not one favored story. Each hypothesis must predict observable evidence. 1. List the evidence that would be expected if each hypothesis were true and the evidence that would be surprising. Prioritize tests with the highest ability to discriminate between explanations. 1. Assign a starting confidence range or qualitative prior based on base rates and known evidence. Do not manufacture numerical precision when no defensible probability exists. 1. Sequence research so cheap, high-information tests occur before expensive channel work or complex modeling. Stop a line of inquiry once it can no longer change the decision. 1. Record disconfirming evidence with the same prominence as confirming evidence. A hypothesis that cannot be falsified is not a usable research hypothesis. 1. At conclusion, state which hypothesis best fits the evidence, what remains unresolved, and the next observation that could reverse the ranking. ## Required evidence and model bridge - Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in hypothesis-driven research. - For each key concept - falsifiable claims, base rates, alternative explanations, Bayesian updating, variant perception, precommitment - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as hypothesis-driven research may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | hypothesis specificity | % of research hypotheses stated with a measurable variable, expected direction/magnitude, time horizon, and explicit falsification condition. | hypothesis specificity: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | expected value of information | Expected decision-value improvement from resolving a question: sum of probability of each possible answer × change in decision value, less research cost. | expected value of information: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | probability update magnitude | Absolute change in assigned scenario or thesis probability after new evidence, measured in percentage points and linked to the triggering evidence. | probability update magnitude: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | ## Worked application > Case: replace "pricing power is strong" with a measurable price, volume, and timing claim. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For hypothesis-driven research, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through falsifiable claims, base rates, alternative explanations, Bayesian updating, then identify which link is directly observed and which link remains an assumption. - Calculate hypothesis specificity, expected value of information, probability update magnitude from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for hypothesis-driven research: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: rank questions by how much evidence can move value, probability, or downside. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the hypothesis-driven research conclusion. ## Failure tests - FAIL if falsifiable claims cannot be defined and reproduced from the source pack. - FAIL if the research question is not falsifiable, has no competing explanation, or evidence collection is selected after the analyst already knows which answer it supports. - FAIL if the hypothesis-driven research conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the hypothesis-driven research conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the hypothesis-driven research conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART I - ANALYST OPERATING SYSTEM | MODULE 005 # Time Allocation and Research ROI > Mission. Allocate analyst time to the questions most likely to change intrinsic value, probability, or risk. ## Decision output Objective: Allocate analyst time to the questions most likely to change intrinsic value, probability, or risk. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Rank research questions by expected decision impact, uncertainty, ability to resolve the uncertainty, time/cost required, and time sensitivity. 1. Translate material questions into valuation or risk sensitivity before allocating senior analyst hours. A fascinating issue with immaterial value impact should remain low priority. 1. Distinguish information gathering from analysis. Stop collecting sources once marginal evidence is unlikely to change the range, probability, or risk assessment. 1. Use deadlines around filings, catalysts, and committee decisions to prevent low-value perfectionism from crowding out decision-critical work. 1. Track forecast errors and past research outcomes to learn which activities actually improved decisions and which created false confidence. 1. Reallocate coverage time toward variables where the analyst has both an information edge and a meaningful path from evidence to value. ## Required evidence and model bridge - Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in time allocation and research roi. - For each key concept - valuation sensitivity, uncertainty, information value, time cost, catalyst deadline, stopping rules - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as time allocation and research roi may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | research ROI | Expected value of information or avoided error divided by analyst time/cost devoted to the research question. | research ROI: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | sensitivity-weighted uncertainty | For each uncertain input, absolute valuation sensitivity × plausible input range; sum or rank across inputs to identify uncertainty concentration. | sensitivity-weighted uncertainty: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | deadline-adjusted information value | Expected value of information multiplied by the probability the answer arrives before the decision deadline, net of acquisition cost. | deadline-adjusted information value: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: a terminal-margin question can move value 30% while a tax detail moves value 0.5%. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For time allocation and research roi, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through valuation sensitivity, uncertainty, information value, time cost, then identify which link is directly observed and which link remains an assumption. - Calculate research ROI, sensitivity-weighted uncertainty, deadline-adjusted information value from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for time allocation and research roi: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: allocate deep work to the assumptions with the largest decision impact, not the most interesting questions. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the time allocation and research roi conclusion. ## Failure tests - FAIL if valuation sensitivity cannot be defined and reproduced from the source pack. - FAIL if senior research time is spent on low-value questions while high-sensitivity, weak-evidence assumptions remain unresolved or unowned. - FAIL if the time allocation and research roi conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the time allocation and research roi conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the time allocation and research roi conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART II - SOURCE INTELLIGENCE AND FILINGS | MODULE 006 # EDGAR Mastery and Filing Retrieval > Mission. Use SEC filing search, accession numbers, exhibits, XBRL data, and amendment history efficiently. ## Decision output Objective: Use SEC filing search, accession numbers, exhibits, XBRL data, and amendment history efficiently. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Resolve the issuer CIK and legal registrant before retrieval, especially for dual listings, holding-company structures, renamed issuers, and multiple share classes. 1. Use accession number, form type, filing date, report period, amendment status, and exhibit index as the canonical filing identifiers. Archive the exact filed version used in analysis. 1. Retrieve the complete filing package, not only the rendered main document. Inspect exhibits, material contracts, debt documents, certifications, XBRL facts, and amendments when relevant. 1. Use SEC submissions/company-facts APIs or structured filing data for scalable retrieval, then reconcile machine-extracted values to the actual filing table and footnotes before modeling. 1. Compare current and prior filing text for inserted qualifiers, removed language, definition changes, new risk factors, and changed segment/KPI presentation. 1. Build filing alerts by form and material item so retrieval is systematic, then log the filing-to-model action or explicit no-action conclusion. ## Required evidence and model bridge - Primary-source set: EDGAR filings, exhibits, prior-period filings, debt documents, ownership forms. Preserve exact document/version, date, period, and source location for every material factual input used in edgar mastery and filing retrieval. - For each key concept - CIK mapping, accession numbers, amendments, filing packages, exhibits, XBRL - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as edgar mastery and filing retrieval may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | filing completeness | % of required filing components, exhibits, amendments, and referenced documents retrieved and reviewed for the relevant period/event. | filing completeness: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | amendment control | Binary/version-control check that the analysis uses the latest amended filing and records superseded accession numbers and changed disclosures. | amendment control: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | XBRL-to-filing reconciliation | Difference between machine-readable XBRL facts and the filed financial statements/footnotes after unit, sign, scale, and taxonomy mapping; target zero unexplained difference. | XBRL-to-filing reconciliation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: XBRL revenue differs from the model because one fact is YTD and the other is quarterly. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For edgar mastery and filing retrieval, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through CIK mapping, accession numbers, amendments, filing packages, then identify which link is directly observed and which link remains an assumption. - Calculate filing completeness, amendment control, XBRL-to-filing reconciliation from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for edgar mastery and filing retrieval: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: automate retrieval but keep human control of periods, dimensions, custom tags, and materiality. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the edgar mastery and filing retrieval conclusion. ## Failure tests - FAIL if CIK mapping cannot be defined and reproduced from the source pack. - FAIL if the analyst cannot prove the filing form, period, accession, amendment status, and exhibit version used for a material conclusion. - FAIL if the edgar mastery and filing retrieval conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the edgar mastery and filing retrieval conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the edgar mastery and filing retrieval conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART II - SOURCE INTELLIGENCE AND FILINGS | MODULE 007 # 10-K Deep Read Procedure > Mission. Extract business economics, accounting policies, risks, obligations, segment data, and footnote detail from annual reports. ## Decision output Objective: Extract business economics, accounting policies, risks, obligations, segment data, and footnote detail from annual reports. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. First pass: map the business, segments, customers, geography, seasonality, competition, regulation, material properties, and stated risks without updating the model. 1. Second pass: reconstruct accounting policies and footnotes for revenue, segment reporting, leases, debt, taxes, stock compensation, pensions, acquisitions, goodwill/intangibles, contingencies, fair value, and related parties. 1. Reconcile MD&A explanations to actual statement and footnote movements. Quantify price/volume/mix, margin drivers, working capital, capex, liquidity, and known commitments. 1. Read the auditor report, critical audit matters where applicable, internal-control disclosures, and any material weaknesses or restatements. 1. Compare the 10-K with the prior year line by line for disclosure drift and with prior guidance/promises for accountability. 1. Produce a 10-K delta memo listing new facts, changed definitions, model changes, unresolved questions, and thesis implications. ## Required evidence and model bridge - Primary-source set: EDGAR filings, exhibits, prior-period filings, debt documents, ownership forms. Preserve exact document/version, date, period, and source location for every material factual input used in 10-k deep read procedure. - For each key concept - business sections, segment note, accounting policies, commitments, auditor report, CAMs - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as 10-k deep read procedure may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | footnote coverage | % of material balance-sheet, income-statement, cash-flow, segment, commitment, tax, and accounting-policy areas traced to their supporting footnotes. | footnote coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | disclosure drift | Count and severity of substantive wording, definition, scope, or omission changes versus the prior comparable filing, classified by likely economic significance. | disclosure drift: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | obligation coverage | Present value or undiscounted amount of debt, leases, purchase commitments, guarantees, pensions, and other fixed obligations captured in the model divided by disclosed obligations. | obligation coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: inventory rises 35% while sales rise 8% and new risk language appears. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For 10-k deep read procedure, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through business sections, segment note, accounting policies, commitments, then identify which link is directly observed and which link remains an assumption. - Calculate footnote coverage, disclosure drift, obligation coverage from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for 10-k deep read procedure: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: read transaction and debt exhibits when they define economics that summarized MD&A omits. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the 10-k deep read procedure conclusion. ## Failure tests - FAIL if business sections cannot be defined and reproduced from the source pack. - FAIL if the annual review omits a material accounting policy, segment change, contractual obligation, risk change, or footnote needed to reconcile the model. - FAIL if the 10-k deep read procedure conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the 10-k deep read procedure conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the 10-k deep read procedure conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART II - SOURCE INTELLIGENCE AND FILINGS | MODULE 008 # 10-Q and Interim Update Procedure > Mission. Reconcile interim results, seasonal patterns, working capital, guidance changes, and emerging risks. ## Decision output Objective: Reconcile interim results, seasonal patterns, working capital, guidance changes, and emerging risks. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Separate quarter-only economics from year-to-date presentation and reconcile current-quarter values when the filing presents cumulative cash-flow or tax data. 1. Compare the quarter with the same seasonal quarter, sequential quarter, trailing periods, and the latest annual run rate before inferring a trend. 1. Reconcile revenue/margin drivers, working capital, capex, cash, debt, taxes, diluted shares, and segment/KPI changes to the 10-Q footnotes and earnings materials. 1. Identify guidance changes, new contingencies, subsequent events, covenant/liquidity changes, customer concentration, inventory/channel movement, and disclosure wording changes. 1. Update historical actuals before forecasts, then attribute every material forecast revision to new evidence rather than mechanical rolling. 1. Write a quarter delta: what changed versus the pre-quarter expectation, whether the thesis moved, and what evidence the next quarter must show. ## Required evidence and model bridge - Primary-source set: EDGAR filings, exhibits, prior-period filings, debt documents, ownership forms. Preserve exact document/version, date, period, and source location for every material factual input used in 10-q and interim update procedure. - For each key concept - quarter/YTD reconciliation, seasonality, working capital, guidance math, wording changes, debt/cash/share update - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as 10-q and interim update procedure may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | quarter bridge | Current-quarter result minus prior comparable quarter, decomposed into volume, price, mix, FX, acquisitions/divestitures, accounting, and cost/margin effects. | quarter bridge: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | | cash conversion drift | Change in CFO or FCF conversion versus historical/peer baseline, with working-capital, capex, tax, and one-time cash drivers isolated. | cash conversion drift: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | implied remaining-period guidance | Full-year guidance midpoint or range minus actual year-to-date results, adjusted for seasonality, to derive the implied remaining-quarter revenue/earnings/cash-flow requirement. | implied remaining-period guidance: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | ## Worked application > Case: full-year guidance is reiterated after a weak first half. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For 10-q and interim update procedure, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through quarter/YTD reconciliation, seasonality, working capital, guidance math, then identify which link is directly observed and which link remains an assumption. - Calculate quarter bridge, cash conversion drift, implied remaining-period guidance from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for 10-q and interim update procedure: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: convert annual guidance into required remaining-quarter performance before accepting the headline. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the 10-q and interim update procedure conclusion. ## Failure tests - FAIL if quarter/YTD reconciliation cannot be defined and reproduced from the source pack. - FAIL if quarterly performance is interpreted without reconciling year-to-date math, seasonality, working capital, guidance, and changes from the prior comparable filing. - FAIL if the 10-q and interim update procedure conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the 10-q and interim update procedure conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the 10-q and interim update procedure conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART II - SOURCE INTELLIGENCE AND FILINGS | MODULE 009 # 8-K and Event Filing Triage > Mission. Classify current reports by materiality, financial impact, transaction type, and required model action. ## Decision output Objective: Classify current reports by materiality, financial impact, transaction type, and required model action. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Identify the 8-K item number, event date, filing date, incorporated exhibits, and whether the filing furnishes or files information with different legal consequences. 1. Classify the event into operating results, material agreement, acquisition/disposition, financing/default, impairment/restructuring, auditor/control, management/board, Reg FD disclosure, or other material category. 1. Read the exhibit or underlying agreement when economics depend on contract terms; do not rely on the 8-K summary alone. 1. Quantify immediate and contingent effects on revenue, cost, assets/liabilities, cash, debt, covenants, shares, taxes, and valuation. 1. Determine whether the event changes the model now, creates a monitoring item, or has no material analytical effect. Record the reason. 1. For earnings 8-Ks, reconcile furnished tables with the later 10-Q/10-K and preserve any definition or non-GAAP differences. ## Required evidence and model bridge - Primary-source set: EDGAR filings, exhibits, prior-period filings, debt documents, ownership forms. Preserve exact document/version, date, period, and source location for every material factual input used in 8-k and event filing triage. - For each key concept - item numbers, exhibits, operating versus financing events, liquidity, transaction terms, model action - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as 8-k and event filing triage may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | event value impact | Change in probability-weighted equity value attributable to the event after updating operating, financing, dilution, and timing assumptions. | event value impact: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | liquidity impact | Change in unrestricted cash plus committed undrawn liquidity less forecast cash needs and near-term maturities caused by the event. | liquidity impact: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | model-action latency | Elapsed time from verified material event to completion of the required model, valuation, risk-register, and memo update. | model-action latency: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | ## Worked application > Case: an acquisition is financed with cash and new debt. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For 8-k and event filing triage, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through item numbers, exhibits, operating versus financing events, liquidity, then identify which link is directly observed and which link remains an assumption. - Calculate event value impact, liquidity impact, model-action latency from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for 8-k and event filing triage: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: quantify first-order financial effects and second-order covenant, customer, and integration effects. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the 8-k and event filing triage conclusion. ## Failure tests - FAIL if item numbers cannot be defined and reproduced from the source pack. - FAIL if a material current report or exhibit is treated as immaterial before quantifying the event, contractual obligation, financing, or required model action. - FAIL if the 8-k and event filing triage conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the 8-k and event filing triage conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the 8-k and event filing triage conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART II - SOURCE INTELLIGENCE AND FILINGS | MODULE 010 # Proxy Statement and Governance Mining > Mission. Use proxy materials to analyze incentives, ownership, related-party issues, board structure, and pay design. ## Decision output Objective: Use proxy materials to analyze incentives, ownership, related-party issues, board structure, and pay design. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct executive compensation using salary, annual incentive, equity grants, grant-date value, realized/realizable outcomes, vesting conditions, performance metrics, thresholds, caps, and change-in-control terms. 1. Map compensation metrics to the economic variables that create or destroy long-term value. Identify incentives that can favor growth, adjusted earnings, buybacks, or short horizons at the expense of ROIC/cash economics. 1. Review beneficial ownership, pledging, hedging, related-party transactions, family/business relationships, and dual-class or other control structures. 1. Evaluate board independence, tenure, refreshment, relevant operating/financial expertise, committee composition, attendance, overboarding, and lead-independent-director structure. 1. Read shareholder proposals, voting outcomes, auditor ratification, say-on-pay results, and governance changes for signs of unresolved owner concerns. 1. Compare at least three years of proxy disclosures to detect metric changes, target resets, retention awards, one-off grants, and shifts in accountability. ## Required evidence and model bridge - Primary-source set: EDGAR filings, exhibits, prior-period filings, debt documents, ownership forms. Preserve exact document/version, date, period, and source location for every material factual input used in proxy statement and governance mining. - For each key concept - comp metrics, grant versus realized pay, ownership, dilution, board skills, related parties - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as proxy statement and governance mining may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | pay alignment | % of executive incentive opportunity tied to economically relevant long-duration outcomes, adjusted for metric quality, measurement period, and discretion. | pay alignment: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | dilution burden | Net dilution = ending diluted share count / beginning diluted share count - 1, adjusted for major capital actions | dilution burden: Recalculate dilution burden from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | control concentration | Share of board voting power, management authority, related-party influence, or key control functions concentrated in a single person/group. | control concentration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: performance pay is tied to adjusted EBITDA while dilution rises 4% per year. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For proxy statement and governance mining, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through comp metrics, grant versus realized pay, ownership, dilution, then identify which link is directly observed and which link remains an assumption. - Calculate pay alignment, dilution burden, control concentration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for proxy statement and governance mining: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: judge incentives by the behaviors they reward, not by the percentage labeled performance-based. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the proxy statement and governance mining conclusion. ## Failure tests - FAIL if comp metrics cannot be defined and reproduced from the source pack. - FAIL if compensation, ownership, board control, related-party transactions, or voting rights are assessed from summary language without reading the underlying proxy tables and footnotes. - FAIL if the proxy statement and governance mining conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the proxy statement and governance mining conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the proxy statement and governance mining conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART III - ACCOUNTING FOUNDATION | MODULE 011 # Three-Statement Mastery > Mission. Connect income statement, balance sheet, and cash flow mechanics before attempting forecast modeling. ## Decision output Objective: Connect income statement, balance sheet, and cash flow mechanics before attempting forecast modeling. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Start with the economic transaction and write the journal-entry logic conceptually before tracing it through the statements. 1. For every modeled balance, build beginning balance plus additions/minus uses plus noncash/reclassification effects equals ending balance. Key schedules include cash, PP&E, debt, equity, deferred taxes, and working capital. 1. Understand retained earnings as beginning retained earnings plus net income minus dividends and other applicable adjustments; reconcile equity changes separately from OCI and share issuance/repurchase. 1. Use the cash-flow statement as a reconciliation of changes in cash, not as an independent forecast. Forecast operating and balance-sheet drivers first, then let cash emerge. 1. Test noncash items by identifying both the accounting offset and future cash consequence. D&A, SBC, deferred tax, impairment, and fair-value changes cannot be understood from the income statement alone. 1. Require the balance sheet to balance in every forecast period without an unexplained plug and require the cash roll-forward to tie exactly. ## Required evidence and model bridge - Primary-source set: audited statements, footnotes, accounting policies, roll-forwards, segment disclosures. Preserve exact document/version, date, period, and source location for every material factual input used in three-statement mastery. - For each key concept - double entry, retained earnings, cash roll-forward, working capital, PP&E, debt - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as three-statement mastery may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | balance check | Total assets minus total liabilities and equity for every historical and forecast period; required to equal zero within immaterial rounding tolerance. | balance check: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | cash roll-forward | Beginning cash + CFO + CFI + CFF + FX/other = ending cash; reconcile exactly to the balance sheet for every period. | cash roll-forward: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | | retained-earnings roll-forward | Beginning retained earnings + net income attributable to common - dividends ± prior-period/other equity adjustments = ending retained earnings. | retained-earnings roll-forward: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | ## Three-statement propagation laboratory - Assume revenue rises by $10, cash collection is delayed, gross margin is 60%, tax is 25%, and no other item changes. Build the income statement effect first, then receivables, retained earnings, taxes, and cash flow. The balance sheet must still balance without a plug. - Create a roll-forward for cash, debt, retained earnings, PP&E, deferred taxes, and diluted shares. Every ending balance must equal beginning balance plus sourced movements. - For each noncash expense, identify the balance-sheet account created or consumed and the later cash consequence. ## Worked application > Case: SBC reduces EBIT, is added back in CFO, and is offset by a cash buyback. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For three-statement mastery, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through double entry, retained earnings, cash roll-forward, working capital, then identify which link is directly observed and which link remains an assumption. - Calculate balance check, cash roll-forward, retained-earnings roll-forward from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for three-statement mastery: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: build transaction-level mini-ledgers for complex accounting before inserting them into a model. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the three-statement mastery conclusion. ## Failure tests - FAIL if double entry cannot be defined and reproduced from the source pack. - FAIL if a material transaction cannot be traced through income statement, balance sheet, cash flow, and retained earnings without an unexplained plug. - FAIL if the three-statement mastery conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the three-statement mastery conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the three-statement mastery conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART III - ACCOUNTING FOUNDATION | MODULE 012 # Revenue Recognition Analysis > Mission. Understand contract economics, timing, variable consideration, principal-agent issues, and revenue quality. ## Decision output Objective: Understand contract economics, timing, variable consideration, principal-agent issues, and revenue quality. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Identify each material revenue stream, customer contract, performance obligation, transaction price, variable consideration, and timing of control transfer. 1. Separate bookings/order intake, billings, cash collections, contract assets, contract liabilities/deferred revenue, RPO/backlog, and recognized revenue. Do not treat them as interchangeable demand measures. 1. Test principal-versus-agent presentation, gross-to-net deductions, rebates, returns, warranties, concessions, usage/consumption, and contract modifications where material. 1. Reconcile reported growth into organic scope, acquisition/divestiture, FX, price, volume, mix, and accounting/presentation effects. 1. Inspect changes in contract assets/liabilities and unbilled receivables for timing shifts that can make revenue growth diverge from cash or underlying delivery. 1. Forecast revenue from economic delivery drivers and contract terms, then reconcile the forecast to backlog/RPO only after adjusting for cancellation, renewal, and timing risk. ## Required evidence and model bridge - Primary-source set: audited statements, footnotes, accounting policies, roll-forwards, segment disclosures. Preserve exact document/version, date, period, and source location for every material factual input used in revenue recognition analysis. - For each key concept - performance obligations, over-time versus point-in-time, variable consideration, principal-agent, contract assets, deferred revenue - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as revenue recognition analysis may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | deferred-revenue conversion | Revenue recognized from beginning deferred revenue or billings cohort divided by the applicable opening deferred-revenue balance; track by term/cohort where available. | deferred-revenue conversion: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | contract-asset growth versus revenue | contract-asset growth versus revenue = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | contract-asset growth versus revenue: Recalculate contract-asset growth versus revenue from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | RPO coverage | Remaining performance obligations divided by next-twelve-month expected revenue, with current RPO separated from long-dated commitments. | RPO coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Revenue recognition laboratory - For a multi-year contract, separate signed contract value, billings, cash collections, deferred revenue/contract liability, contract asset, and GAAP revenue. A booking is not revenue and a billing is not necessarily revenue. - Build a remaining-performance-obligation schedule only when issuer definitions are understood. Model variable consideration, returns/rebates, principal-agent presentation, and contract modification separately when material. - Reconcile organic growth both as reported and on a constant-definition basis when acquisitions, FX, or gross/net presentation change comparability. ## Worked application > Case: a marketplace changes presentation from net to gross. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For revenue recognition analysis, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through performance obligations, over-time versus point-in-time, variable consideration, principal-agent, then identify which link is directly observed and which link remains an assumption. - Calculate deferred-revenue conversion, contract-asset growth versus revenue, RPO coverage from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for revenue recognition analysis: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: separate billings, cash collection, and GAAP revenue because each can move on a different timeline. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the revenue recognition analysis conclusion. ## Failure tests - FAIL if performance obligations cannot be defined and reproduced from the source pack. - FAIL if bookings, billings, cash collections, backlog/RPO, contract balances, and recognized revenue are treated as interchangeable without contract-timing reconciliation. - FAIL if the revenue recognition analysis conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the revenue recognition analysis conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the revenue recognition analysis conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART III - ACCOUNTING FOUNDATION | MODULE 013 # Cost Structure and Margin Architecture > Mission. Separate variable, fixed, semi-fixed, pass-through, and step-function costs to model operating leverage. ## Decision output Objective: Separate variable, fixed, semi-fixed, pass-through, and step-function costs to model operating leverage. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Classify costs as unit-variable, revenue-variable, semi-variable, fixed, step-function, pass-through, or capacity-creating. Use the classification that predicts behavior, not the financial-statement label. 1. Build gross-margin and operating-margin bridges for price, volume, mix, input costs, labor, utilization, logistics, FX, productivity, restructuring, and acquisitions. 1. Estimate incremental margin over comparable intervals and explain why it differs from average margin. Capacity additions and mix shifts can reverse apparent operating leverage. 1. Model price-cost lag explicitly where contracts, commodities, freight, fuel, wages, or annual pricing cycles delay recovery. 1. Separate temporary under-absorption/over-absorption and shortage economics from sustainable structural margin. 1. Stress margin using interacting drivers such as volume plus utilization or price plus input cost rather than an arbitrary basis-point haircut. ## Required evidence and model bridge - Primary-source set: audited statements, footnotes, accounting policies, roll-forwards, segment disclosures. Preserve exact document/version, date, period, and source location for every material factual input used in cost structure and margin architecture. - For each key concept - fixed/variable/step costs, contribution margin, utilization, absorption, input inflation, productivity - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as cost structure and margin architecture may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | incremental margin | incremental margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | incremental margin: Recalculate incremental margin from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | contribution margin | contribution margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | contribution margin: Recalculate contribution margin from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | breakeven revenue | Fixed operating costs divided by contribution margin percentage, after separating truly variable costs from semi-fixed and step-function costs. | breakeven revenue: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | ## Worked application > Case: factory utilization falls to 55% and gross margin compresses sharply. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For cost structure and margin architecture, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through fixed/variable/step costs, contribution margin, utilization, absorption, then identify which link is directly observed and which link remains an assumption. - Calculate incremental margin, contribution margin, breakeven revenue from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for cost structure and margin architecture: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: model the cost driver and capacity state rather than extrapolating a reported margin. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the cost structure and margin architecture conclusion. ## Failure tests - FAIL if fixed/variable/step costs cannot be defined and reproduced from the source pack. - FAIL if the forecast margin can change without an explicit volume, price/mix, input-cost, utilization, headcount, or fixed-cost mechanism. - FAIL if the cost structure and margin architecture conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the cost structure and margin architecture conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the cost structure and margin architecture conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART III - ACCOUNTING FOUNDATION | MODULE 014 # Working Capital Mechanics > Mission. Analyze receivables, inventory, payables, deferred revenue, contract assets, and cash conversion. ## Decision output Objective: Analyze receivables, inventory, payables, deferred revenue, contract assets, and cash conversion. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build separate schedules for receivables, inventory, payables, deferred revenue/contract liabilities, contract assets, prepaid/accrued items, and other material operating balances. 1. Calculate DSO, inventory days, DPO, and other turnover measures using average balances and economically matched flow denominators; adjust for seasonality and acquisitions. 1. Explain changes through growth, mix, billing terms, collections, production timing, channel inventory, supplier terms, customer advances, and deliberate cash management. 1. Identify factoring, securitization, supplier finance, reverse factoring, extended terms, and other financing-like arrangements that can flatter operating cash flow. 1. Forecast balances from operating drivers rather than forcing historical percentages when business mix or contract structure is changing. 1. Convert days/turn changes into cash dollars and test whether favorable working-capital release is repeatable or merely pulls cash forward. ## Required evidence and model bridge - Primary-source set: audited statements, footnotes, accounting policies, roll-forwards, segment disclosures. Preserve exact document/version, date, period, and source location for every material factual input used in working capital mechanics. - For each key concept - receivables, inventory, payables, deferred revenue, supplier finance, factoring - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as working capital mechanics may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | DSO | DSO = average accounts receivable / revenue x days in period | DSO: Recalculate DSO from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | inventory days | Inventory days = average inventory / COGS x days in period | inventory days: Recalculate inventory days from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | DPO | DPO = average accounts payable / COGS x days in period | DPO: Recalculate DPO from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | cash conversion cycle | Cash conversion cycle = DSO + inventory days - DPO | cash conversion cycle: Recalculate cash conversion cycle from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Working-capital laboratory - Forecast receivables from revenue and DSO, inventory from COGS and inventory days, and payables from COGS or purchases and DPO. Use average balances when calculating historical days. - Convert a five-day DSO deterioration into incremental cash tied up: approximate impact = revenue per day x five days, then refine for seasonality and mix. - Test whether supplier finance, factoring, securitization, customer advances, or contract assets are shifting cash flows without changing underlying economics. ## Worked application > Case: CFO jumps because DPO rises after adoption of supplier finance. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For working capital mechanics, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through receivables, inventory, payables, deferred revenue, then identify which link is directly observed and which link remains an assumption. - Calculate DSO, inventory days, DPO, cash conversion cycle from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for working capital mechanics: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: reconcile working-capital cash flows to balance-sheet changes including M&A, FX, and reclassifications. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the working capital mechanics conclusion. ## Failure tests - FAIL if receivables cannot be defined and reproduced from the source pack. - FAIL if receivables, inventory, payables, contract assets, or deferred revenue are forecast from a ratio that no longer matches the commercial process or business mix. - FAIL if the working capital mechanics conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the working capital mechanics conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the working capital mechanics conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART III - ACCOUNTING FOUNDATION | MODULE 015 # Capital Expenditures, Depreciation, and Asset Intensity > Mission. Distinguish maintenance from growth investment and reconcile capex to productive capacity. ## Decision output Objective: Distinguish maintenance from growth investment and reconcile capex to productive capacity. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconcile PP&E beginning balance, capex, acquisitions, disposals, FX/reclassifications, depreciation, and ending balance by major asset class when available. 1. Separate maintenance capex from growth capex using asset age, replacement cycles, capacity, utilization, management plans, physical footprint, and peer evidence rather than management labels alone. 1. Map capex to the productive capacity or service capability it creates and estimate lag from cash spend to revenue and mature utilization. 1. Test useful lives, residual values, depreciation method, capitalized interest, and asset impairments for changes that alter reported margin without equivalent economics. 1. Calculate asset turns, capex/depreciation, capex/revenue, and incremental ROIC through a full investment cycle. 1. In valuation, model the reinvestment required to sustain forecast growth and terminal earning power instead of assuming depreciation automatically equals maintenance capex. ## Required evidence and model bridge - Primary-source set: audited statements, footnotes, accounting policies, roll-forwards, segment disclosures. Preserve exact document/version, date, period, and source location for every material factual input used in capital expenditures, depreciation, and asset intensity. - For each key concept - PP&E roll-forward, maintenance versus growth capex, useful lives, capacity, capitalized software, construction-in-progress - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as capital expenditures, depreciation, and asset intensity may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | capex intensity | Capital expenditures divided by revenue, or by capacity/output for asset-heavy sectors; separate maintenance and growth capex. | capex intensity: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | capex/depreciation | Capital expenditures divided by depreciation and amortization for the same asset base and period; interpret with asset age, inflation, and growth investment. | capex/depreciation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | incremental ROIC | ROIC = NOPAT / average invested capital | incremental ROIC: Recalculate incremental ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: a data-center company spends heavily on new capacity while reporting strong FCF. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For capital expenditures, depreciation, and asset intensity, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through PP&E roll-forward, maintenance versus growth capex, useful lives, capacity, then identify which link is directly observed and which link remains an assumption. - Calculate capex intensity, capex/depreciation, incremental ROIC from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for capital expenditures, depreciation, and asset intensity: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: tie capex to physical capacity, commissioning, utilization, and the return on the new investment. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the capital expenditures, depreciation, and asset intensity conclusion. ## Failure tests - FAIL if PP&E roll-forward cannot be defined and reproduced from the source pack. - FAIL if maintenance and growth investment are not separated where material, or if the capex/depreciation roll-forward cannot reconcile to productive capacity and cash. - FAIL if the capital expenditures, depreciation, and asset intensity conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the capital expenditures, depreciation, and asset intensity conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the capital expenditures, depreciation, and asset intensity conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IV - ADVANCED ACCOUNTING | MODULE 016 # Stock-Based Compensation > Mission. Analyze dilution, economic cost, tax effects, share count, buyback offset, and adjusted-metric treatment. ## Decision output Objective: Analyze dilution, economic cost, tax effects, share count, buyback offset, and adjusted-metric treatment. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct total SBC by employee/function and award type, then reconcile income-statement expense, cash-flow add-back, tax effects, and equity-account movements. 1. Build basic-to-diluted share mechanics including RSUs, options, performance awards, employee purchase plans, convertibles, issuance, repurchases, and treasury-stock treatment as applicable. 1. Measure multi-year net dilution and compare cumulative SBC with cash repurchases used to offset issuance. A noncash expense can still transfer economic value to employees. 1. Evaluate grant cadence, vesting, forfeiture assumptions, option exercise prices, performance conditions, and share-price sensitivity. 1. Normalize peer profitability both including SBC and, if useful, an explicitly reconstructed cash/ownership view. Never treat company-adjusted exclusion as automatically economic. 1. Use diluted per-share valuation and, where material, value outstanding options/awards consistently rather than relying only on a period share-count average. ## Required evidence and model bridge - Primary-source set: footnotes, valuation inputs, tax notes, compensation tables, lease and pension schedules. Preserve exact document/version, date, period, and source location for every material factual input used in stock-based compensation. - For each key concept - grant accounting, vesting, RSUs/PSUs/options, dilution, treasury-stock method, repurchases - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as stock-based compensation may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | net dilution | Net dilution = ending diluted share count / beginning diluted share count - 1, adjusted for major capital actions | net dilution: Recalculate net dilution from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | SBC/revenue | Stock-based compensation expense divided by revenue; also track on a per-employee and diluted-share basis when material. | SBC/revenue: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | SBC/FCF | Stock-based compensation expense divided by reported free cash flow; use to show how much cash-flow presentation relies on a non-cash but economically dilutive cost. | SBC/FCF: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | buyback offset efficiency | Shares retired from repurchases divided by gross shares issued from SBC, options, acquisitions, and other equity issuance; also compare repurchase dollars with net share reduction. | buyback offset efficiency: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Stock compensation and dilution laboratory - Build a share-count waterfall from basic shares to diluted shares, including RSUs, options, performance awards, convertibles, issuance, repurchases, and treasury-stock mechanics as applicable. - Treat SBC as an economic cost even when noncash in the period. Show separately the expense add-back in cash flow, the dilution created, the tax effect, and the cash used to offset dilution through repurchases. - Calculate net dilution over three to five years and compare cumulative SBC expense with repurchase spending and change in diluted share count. ## Worked application > Case: FCF adds back SBC while buybacks fail to prevent dilution. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For stock-based compensation, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through grant accounting, vesting, RSUs/PSUs/options, dilution, then identify which link is directly observed and which link remains an assumption. - Calculate net dilution, SBC/revenue, SBC/FCF, buyback offset efficiency from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for stock-based compensation: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: treat SBC consistently across profit, cash flow, and per-share value so it is neither ignored nor double counted. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the stock-based compensation conclusion. ## Failure tests - FAIL if grant accounting cannot be defined and reproduced from the source pack. - FAIL if the analysis treats SBC only as a noncash add-back or only as dilution, rather than reconciling expense, awards, tax effects, share count, and repurchase offset. - FAIL if the stock-based compensation conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the stock-based compensation conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the stock-based compensation conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IV - ADVANCED ACCOUNTING | MODULE 017 # Leases and Off-Balance-Sheet Commitments > Mission. Reconstruct lease economics, fixed commitments, and debt-like obligations. ## Decision output Objective: Reconstruct lease economics, fixed commitments, and debt-like obligations. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct operating and finance lease assets/liabilities, maturities, discount rates, lease expense, cash payments, variable rent, and renewal/termination assumptions. 1. Normalize peers with different own-versus-lease strategies before comparing EBITDA, EBIT, margins, leverage, and returns on capital. 1. Map purchase commitments, take-or-pay contracts, guarantees, minimum volumes, supplier commitments, data-center/energy contracts, and other fixed obligations not captured by headline debt. 1. Decide whether each obligation is operating, financing, or both for the analytical question, and avoid double counting in cash forecasts and EV-to-equity adjustments. 1. Stress fixed commitments under revenue or utilization downside to determine true operating leverage and liquidity risk. 1. Reconcile note disclosures to cash-flow forecasts and debt-like claims used in valuation. ## Required evidence and model bridge - Primary-source set: footnotes, valuation inputs, tax notes, compensation tables, lease and pension schedules. Preserve exact document/version, date, period, and source location for every material factual input used in leases and off-balance-sheet commitments. - For each key concept - ROU assets, lease liabilities, maturity tables, variable rent, renewal options, take-or-pay - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as leases and off-balance-sheet commitments may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | lease-adjusted leverage | Debt plus lease liabilities and other debt-like fixed obligations divided by lease-adjusted EBITDA or another consistent cash-earnings measure. | lease-adjusted leverage: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | fixed-charge coverage | Cash earnings available for fixed charges divided by cash interest, required lease/rent payments, preferred dividends, and other contractual fixed charges included in the definition. | fixed-charge coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | commitment burden | Undiscounted or present value of contractual commitments due over the selected horizon divided by liquidity or normalized annual FCF. | commitment burden: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: two retailers have identical EBITDA but one owns stores and one leases them. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For leases and off-balance-sheet commitments, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through ROU assets, lease liabilities, maturity tables, variable rent, then identify which link is directly observed and which link remains an assumption. - Calculate lease-adjusted leverage, fixed-charge coverage, commitment burden from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for leases and off-balance-sheet commitments: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: normalize EV and earnings consistently so leases are not counted twice or ignored. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the leases and off-balance-sheet commitments conclusion. ## Failure tests - FAIL if ROU assets cannot be defined and reproduced from the source pack. - FAIL if lease and contractual commitments are omitted from liquidity or valuation while related operating expenses or debt-like claims are treated inconsistently. - FAIL if the leases and off-balance-sheet commitments conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the leases and off-balance-sheet commitments conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the leases and off-balance-sheet commitments conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IV - ADVANCED ACCOUNTING | MODULE 018 # Goodwill, Intangibles, and Impairment > Mission. Assess acquisition accounting, amortization, impairment risk, and return on acquired capital. ## Decision output Objective: Assess acquisition accounting, amortization, impairment risk, and return on acquired capital. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct acquisition purchase price allocation into tangible assets, identifiable intangibles, deferred taxes, goodwill, contingent consideration, and later changes. 1. Track intangible type, useful life, amortization, renewal/replacement needs, and whether reported amortization approximates economic consumption. 1. Evaluate impairment testing assumptions including reporting units, projected cash flows, discount rates, long-term growth, and headroom where disclosed. 1. Measure acquisition returns from actual cash invested and post-deal operating cash generation. Impairment is a lagging accounting signal, not the date value destruction necessarily began. 1. Normalize peer comparisons for acquisitive versus organic strategies without pretending acquired customer relationships, technology, or brands are costless. 1. Include contingent consideration, minority stakes, deferred payments, and other acquisition-related claims in the valuation bridge where material. ## Required evidence and model bridge - Primary-source set: footnotes, valuation inputs, tax notes, compensation tables, lease and pension schedules. Preserve exact document/version, date, period, and source location for every material factual input used in goodwill, intangibles, and impairment. - For each key concept - purchase accounting, goodwill, identifiable intangibles, amortization, reporting units, impairment assumptions - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as goodwill, intangibles, and impairment may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | acquisition ROIC | ROIC = NOPAT / average invested capital | acquisition ROIC: Recalculate acquisition ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | goodwill/invested capital | Goodwill divided by average invested capital; pair with acquired-intangible balances and post-deal ROIC to assess acquisition dependence. | goodwill/invested capital: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | impairment headroom | Estimated fair value of the reporting unit or asset less carrying value, expressed in dollars and as a % of carrying value; use disclosed sensitivity where available. | impairment headroom: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Worked application > Case: a serial acquirer grows adjusted EPS while goodwill keeps rising. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For goodwill, intangibles, and impairment, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through purchase accounting, goodwill, identifiable intangibles, amortization, then identify which link is directly observed and which link remains an assumption. - Calculate acquisition ROIC, goodwill/invested capital, impairment headroom from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for goodwill, intangibles, and impairment: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: evaluate economic impairment before accounting impairment and score deals against original underwriting. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the goodwill, intangibles, and impairment conclusion. ## Failure tests - FAIL if purchase accounting cannot be defined and reproduced from the source pack. - FAIL if acquisition value, identifiable intangibles, goodwill, amortization, impairment indicators, and acquired-capital returns cannot be reconciled by transaction or reporting unit. - FAIL if the goodwill, intangibles, and impairment conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the goodwill, intangibles, and impairment conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the goodwill, intangibles, and impairment conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IV - ADVANCED ACCOUNTING | MODULE 019 # Taxes and Deferred Taxes > Mission. Model cash taxes, effective tax rates, NOLs, valuation allowances, jurisdiction mix, and one-time items. ## Decision output Objective: Model cash taxes, effective tax rates, NOLs, valuation allowances, jurisdiction mix, and one-time items. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Bridge statutory tax rate to reported effective rate by jurisdiction, credits, permanent items, stock-compensation effects, discrete items, valuation allowances, and other major reconciling items. 1. Reconcile tax expense to current tax payable/cash taxes and deferred tax changes. Identify temporary differences and the events that cause reversal. 1. Build NOL and credit schedules with jurisdiction, amount, expiration, usage assumptions, valuation allowance, and material legal limitations when available. 1. Model cash taxes from forecast taxable economics rather than extrapolating a noisy historical effective rate. 1. Evaluate uncertain tax positions, audits, repatriation/global minimum tax effects, withholding, and acquisition tax attributes where material. 1. In valuation, distinguish recurring cash-tax rate from one-time benefits and avoid capitalizing tax assets that cannot realistically be used. ## Required evidence and model bridge - Primary-source set: footnotes, valuation inputs, tax notes, compensation tables, lease and pension schedules. Preserve exact document/version, date, period, and source location for every material factual input used in taxes and deferred taxes. - For each key concept - statutory-to-effective reconciliation, current versus deferred tax, cash taxes, NOLs, valuation allowance, jurisdictions - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as taxes and deferred taxes may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | cash tax rate | Cash taxes paid, adjusted for refunds and material non-operating/discrete items, divided by pre-tax cash earnings or normalized pre-tax income. | cash tax rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | effective tax rate | Income tax expense divided by pre-tax book income, with discrete items and jurisdictional mix separately identified. | effective tax rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | usable DTA ratio | Deferred tax assets expected to be realizable before expiration divided by gross deferred tax assets, net of valuation allowance considerations. | usable DTA ratio: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Cash-tax laboratory - Reconcile statutory rate to effective tax rate, then bridge tax expense to cash taxes using deferred taxes, NOLs, valuation allowances, uncertain tax positions, discrete items, and jurisdiction mix. - Do not capitalize a tax benefit into valuation without identifying the legal/economic mechanism and period in which cash tax actually changes. - For DCF, forecast cash taxes on operating income with explicit treatment of NOL usage and limits rather than mechanically applying the historical effective rate. ## Worked application > Case: a one-time tax benefit lowers ETR to 12% while cash taxes stay near 22%. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For taxes and deferred taxes, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through statutory-to-effective reconciliation, current versus deferred tax, cash taxes, NOLs, then identify which link is directly observed and which link remains an assumption. - Calculate cash tax rate, effective tax rate, usable DTA ratio from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for taxes and deferred taxes: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: forecast cash tax economics, not a copied GAAP tax rate. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the taxes and deferred taxes conclusion. ## Failure tests - FAIL if statutory-to-effective reconciliation cannot be defined and reproduced from the source pack. - FAIL if a reported effective tax rate is used as normalized cash tax without reconciling current/deferred tax, NOLs, credits, valuation allowances, and material discrete items. - FAIL if the taxes and deferred taxes conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the taxes and deferred taxes conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the taxes and deferred taxes conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IV - ADVANCED ACCOUNTING | MODULE 020 # Pensions and Postretirement Obligations > Mission. Analyze funded status, discount rates, expected returns, cash contributions, and hidden leverage. ## Decision output Objective: Analyze funded status, discount rates, expected returns, cash contributions, and hidden leverage. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct plan assets, projected/accumulated benefit obligations, funded status, discount rate, expected return assumptions, service cost, interest cost, actuarial changes, contributions, and benefit payments. 1. Separate operating compensation economics from financing/asset-return components when comparing operating performance across issuers. 1. Assess asset allocation, duration mismatch, discount-rate sensitivity, longevity assumptions, and expected required contributions under stress. 1. Treat material underfunding and required contributions as claims on enterprise cash flow and reconcile treatment in EV-to-equity value. 1. Normalize one-time settlements, curtailments, pension income, and OCI movements before using earnings or cash flow in valuation. 1. Use note sensitivities and independent rate movement to test whether reported funded status can change materially without changes in core operations. ## Required evidence and model bridge - Primary-source set: footnotes, valuation inputs, tax notes, compensation tables, lease and pension schedules. Preserve exact document/version, date, period, and source location for every material factual input used in pensions and postretirement obligations. - For each key concept - PBO, plan assets, funded status, service cost, interest cost, expected return - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as pensions and postretirement obligations may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | funded ratio | Fair value of pension plan assets divided by projected benefit obligation (or comparable disclosed obligation measure). | funded ratio: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | pension deficit/EBITDA | Underfunded pension obligation, PBO less plan assets, divided by normalized EBITDA. | pension deficit/EBITDA: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | contribution burden | Expected required pension/postretirement cash contributions over the selected horizon divided by CFO, FCF, or liquidity. | contribution burden: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: modest stated leverage hides a large underfunded pension. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For pensions and postretirement obligations, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through PBO, plan assets, funded status, service cost, then identify which link is directly observed and which link remains an assumption. - Calculate funded ratio, pension deficit/EBITDA, contribution burden from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for pensions and postretirement obligations: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: stress asset returns, discount rates, and required contributions separately. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the pensions and postretirement obligations conclusion. ## Failure tests - FAIL if PBO cannot be defined and reproduced from the source pack. - FAIL if funded status, discount rate, plan assets, cash contributions, benefit payments, and debt-like valuation treatment are not analyzed consistently. - FAIL if the pensions and postretirement obligations conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the pensions and postretirement obligations conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the pensions and postretirement obligations conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART V - FORENSIC ACCOUNTING | MODULE 021 # Earnings Quality Framework > Mission. Score how much reported profit is supported by cash generation, repeatable economics, and conservative accounting. ## Decision output Objective: Score how much reported profit is supported by cash generation, repeatable economics, and conservative accounting. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Start with multi-year cash generation and balance-sheet change, then evaluate whether reported earnings are supported by repeatable customer economics and conservative recognition. 1. Decompose earnings into recurring operating profit, cyclical/mix effects, working-capital timing, accounting estimates, non-operating items, tax effects, and genuinely unusual events. 1. Compare net income, EBIT, EBITDA, CFO, and FCF over a full cycle while reconciling capex, SBC, acquisitions, restructuring, and financing-like working-capital actions. 1. Score revenue quality, margin quality, cash conversion, balance-sheet quality, estimate sensitivity, and disclosure consistency separately rather than collapsing everything into one ratio. 1. Use peer and historical baselines to distinguish business-model-normal accruals from deterioration. 1. Translate quality concerns into normalized earnings/cash flow and valuation rather than using a qualitative red flag with no financial consequence. ## Required evidence and model bridge - Primary-source set: multi-year statements, auditor reports, non-GAAP reconciliations, reserve disclosures, filing changes. Preserve exact document/version, date, period, and source location for every material factual input used in earnings quality framework. - For each key concept - cash support, accruals, working capital, capitalization, non-GAAP, reserves - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as earnings quality framework may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | CFO/net income | Cash conversion = cash flow from operations / net income, interpreted over multiple years | CFO/net income: Recalculate CFO/net income from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | FCF conversion | Free cash flow divided by the relevant earnings base, typically net income, EBIT, or EBITDA; define consistently and reconcile every adjustment. | FCF conversion: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | accrual ratio | Total accruals proxy = net income - cash flow from operations; scale consistently, commonly by average assets | accrual ratio: Recalculate accrual ratio from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: adjusted EBITDA rises while CFO is flat and capitalized costs accelerate. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For earnings quality framework, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through cash support, accruals, working capital, capitalization, then identify which link is directly observed and which link remains an assumption. - Calculate CFO/net income, FCF conversion, accrual ratio from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for earnings quality framework: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: score quality over a multi-year cycle and convert concerns into normalized cash flow. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the earnings quality framework conclusion. ## Failure tests - FAIL if cash support cannot be defined and reproduced from the source pack. - FAIL if the quality conclusion relies on headline earnings without multi-year cash conversion, accruals, balance-sheet growth, recurring adjustments, and accounting-estimate review. - FAIL if the earnings quality framework conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the earnings quality framework conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the earnings quality framework conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART V - FORENSIC ACCOUNTING | MODULE 022 # Accrual and Cash Conversion Tests > Mission. Use accrual ratios, CFO versus net income, working-capital patterns, and multi-year normalization. ## Decision output Objective: Use accrual ratios, CFO versus net income, working-capital patterns, and multi-year normalization. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Calculate CFO/net income, accrual proxies, working-capital components, and multi-year cumulative earnings versus cash flow using consistent definitions. 1. Decompose CFO changes into underlying profit, receivables, inventory, payables, deferred revenue/customer advances, taxes, restructuring, and other operating balances. 1. Use average assets or another consistent scale when comparing accrual intensity and recognize that high-growth or subscription models can have structurally different working capital. 1. Test persistence across at least twelve quarters and a business cycle before labeling a divergence abnormal. 1. Investigate factoring, supplier finance, classification changes, acquisitions, and one-time working-capital releases that can mechanically improve cash conversion. 1. Normalize valuation cash flow only after explaining the causal source of the accrual/cash gap. ## Required evidence and model bridge - Primary-source set: multi-year statements, auditor reports, non-GAAP reconciliations, reserve disclosures, filing changes. Preserve exact document/version, date, period, and source location for every material factual input used in accrual and cash conversion tests. - For each key concept - total accruals, working-capital accruals, reversals, acquisitions, supplier finance, factoring - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as accrual and cash conversion tests may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | net income minus CFO scaled by assets | (Net income - cash flow from operations) divided by average total assets; positive values indicate accruals exceeding operating cash generation. | net income minus CFO scaled by assets: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | CFO/NI | Cash flow from operations divided by net income over the same period; evaluate over multiple years to reduce working-capital timing noise. | CFO/NI: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | working-capital accrual ratios | Total accruals proxy = net income - cash flow from operations; scale consistently, commonly by average assets | working-capital accrual ratios: Recalculate working-capital accrual ratios from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Accrual and cash-conversion laboratory - Calculate multi-year CFO/net income, total accrual proxy, DSO, inventory days, DPO, and cash conversion cycle. Flag only persistent or causally unexplained divergences. - When CFO improves, decompose the improvement into earnings, receivables, inventory, payables, deferred revenue, tax, and other working-capital movements. Supplier or customer financing can make CFO temporarily flattering. - Compare the same analysis with at least two peers because business-model structure determines normal accrual intensity. ## Worked application > Case: inventory builds before a launch and must be validated by later sell-through. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For accrual and cash conversion tests, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through total accruals, working-capital accruals, reversals, acquisitions, then identify which link is directly observed and which link remains an assumption. - Calculate net income minus CFO scaled by assets, CFO/NI, working-capital accrual ratios from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for accrual and cash conversion tests: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use accrual screens as prompts for account-level investigation, never as standalone accusations. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the accrual and cash conversion tests conclusion. ## Failure tests - FAIL if total accruals cannot be defined and reproduced from the source pack. - FAIL if an accrual signal is labeled aggressive or benign without identifying the balance-sheet accounts, economic driver, reversal pattern, and cash consequence. - FAIL if the accrual and cash conversion tests conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the accrual and cash conversion tests conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the accrual and cash conversion tests conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART V - FORENSIC ACCOUNTING | MODULE 023 # Non-GAAP Reconstruction > Mission. Rebuild management adjustments and classify recurring, discretionary, acquisition-related, and economically real costs. ## Decision output Objective: Rebuild management adjustments and classify recurring, discretionary, acquisition-related, and economically real costs. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Rebuild every major adjusted measure from GAAP for at least twelve quarters using the issuer reconciliation and your own consistent classification. 1. Classify exclusions as recurring operating cost, noncash but economic cost, acquisition/integration cost, restructuring, litigation, tax/financing item, or genuinely unusual event. 1. Run a recurrence test: repeated restructuring, acquisition costs, stock compensation, or other essential costs should not disappear from economic analysis simply because management excludes them. 1. Check naming, prominence, consistency, reconciliation, individually tailored accounting, and cash-cost treatment against current SEC non-GAAP guidance. 1. Rebuild adjusted FCF from actual cash movements and include recurring capex-like expenditures and economically recurring cash charges. 1. Show valuation under GAAP/fully economic normalization and management-adjusted presentation so the effect of adjustments is transparent. ## Required evidence and model bridge - Primary-source set: multi-year statements, auditor reports, non-GAAP reconciliations, reserve disclosures, filing changes. Preserve exact document/version, date, period, and source location for every material factual input used in non-gaap reconstruction. - For each key concept - adjustment ledger, recurring exclusions, SBC, restructuring, M&A costs, tax effects - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as non-gaap reconstruction may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | adjustment persistence | Number of consecutive periods a supposedly non-recurring adjustment appears, plus cumulative adjustment amount as a % of GAAP earnings. | adjustment persistence: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | GAAP-to-non-GAAP gap | Non-GAAP earnings or operating income minus GAAP equivalent, expressed in dollars and as a % of the GAAP measure; reconcile by adjustment category. | GAAP-to-non-GAAP gap: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | cash adjustment ratio | Cash costs excluded from non-GAAP results divided by total non-GAAP add-backs; higher values imply more economically recurring exclusions. | cash adjustment ratio: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Non-GAAP reconstruction laboratory - Rebuild adjusted operating income and EPS from GAAP for at least twelve quarters. Classify each exclusion as recurring operating cost, acquisition-related cost, restructuring, noncash but economic cost, or genuinely unusual item. - Create a recurrence test: if the same category appears repeatedly or is necessary to operate the acquisition model, do not treat it as economically absent merely because management labels it nonrecurring. - Compare management-adjusted FCF with cash actually available after recurring capex-like uses, restructuring cash, acquisition integration, and SBC-related dilution. ## Worked application > Case: restructuring is excluded in seven of eight years. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For non-gaap reconstruction, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through adjustment ledger, recurring exclusions, SBC, restructuring, then identify which link is directly observed and which link remains an assumption. - Calculate adjustment persistence, GAAP-to-non-GAAP gap, cash adjustment ratio from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for non-gaap reconstruction: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: create an analyst-normalized policy that does not change to fit each quarter. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the non-gaap reconstruction conclusion. ## Failure tests - FAIL if adjustment ledger cannot be defined and reproduced from the source pack. - FAIL if recurring excluded costs, cash costs, acquisition-related items, and favorable adjustments are not classified consistently across multiple periods. - FAIL if the non-gaap reconstruction conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the non-gaap reconstruction conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the non-gaap reconstruction conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART V - FORENSIC ACCOUNTING | MODULE 024 # Fraud and Manipulation Red Flags > Mission. Use professional skepticism, cross-statement inconsistencies, unusual transactions, and disclosure drift. ## Decision output Objective: Use professional skepticism, cross-statement inconsistencies, unusual transactions, and disclosure drift. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Do not start by alleging fraud. Build a chronology of anomalies and test benign, aggressive-accounting, control-failure, and intentional-manipulation explanations separately. 1. Cross-check revenue, receivables/contract assets, inventory, payables, reserves, cash, taxes, shares, segment data, and cash flow for inconsistencies that cannot be explained by the business model. 1. Test revenue cutoff, side agreements, channel stuffing indicators, unusual end-of-period working capital, reserve releases, capitalization, related parties, acquisitions, and non-GAAP changes where evidence warrants. 1. Read auditor changes, internal-control weaknesses, restatements, delayed filings, critical audit matters, legal/regulatory disclosures, and management departures as a connected timeline. 1. Compare disclosure wording and KPI definitions across filings for drift that coincides with deteriorating economics. 1. Escalate only after quantifying materiality and documenting what evidence remains unresolved; screening ratios are prompts for investigation, never proof. ## Required evidence and model bridge - Primary-source set: multi-year statements, auditor reports, non-GAAP reconciliations, reserve disclosures, filing changes. Preserve exact document/version, date, period, and source location for every material factual input used in fraud and manipulation red flags. - For each key concept - cross-statement inconsistencies, revenue cut-off, reserves, related parties, controls, auditor changes - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as fraud and manipulation red flags may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | red-flag density | Count of independent forensic red flags per reporting period or per 10 material accounting areas, weighted by severity and corroboration. | red-flag density: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | restatement severity | Restated cumulative pre-tax/net-income/equity effect divided by prior reported amount, supplemented by number of periods affected and control implications. | restatement severity: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | control-remediation persistence | Number of reporting periods a material weakness/significant deficiency remains unresolved, with recurrence and scope expansion separately flagged. | control-remediation persistence: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: receivables, distributor inventory, terms, and incentive pressure deteriorate together. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For fraud and manipulation red flags, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through cross-statement inconsistencies, revenue cut-off, reserves, related parties, then identify which link is directly observed and which link remains an assumption. - Calculate red-flag density, restatement severity, control-remediation persistence from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for fraud and manipulation red flags: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: differentiate poor controls, aggressive accounting, and fraud and keep language proportional to evidence. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the fraud and manipulation red flags conclusion. ## Failure tests - FAIL if cross-statement inconsistencies cannot be defined and reproduced from the source pack. - FAIL if a fraud or manipulation conclusion is drawn from a screening ratio without primary-source evidence, competing explanations, and quantified financial effect. - FAIL if the fraud and manipulation red flags conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the fraud and manipulation red flags conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the fraud and manipulation red flags conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART V - FORENSIC ACCOUNTING | MODULE 025 # Accounting Estimates and Management Bias > Mission. Identify estimates with subjective assumptions, sensitivity, unobservable inputs, and asymmetric incentives. ## Decision output Objective: Identify estimates with subjective assumptions, sensitivity, unobservable inputs, and asymmetric incentives. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Inventory estimates with high judgment: credit/loss reserves, warranty/returns, gross-to-net, useful lives, impairment, fair value, contingent consideration, tax allowances, pensions, legal provisions, and revenue estimates. 1. For each estimate, identify management inputs, observable inputs, model method, historical error/revision pattern, and the direction in which optimistic assumptions affect earnings or capital. 1. Build roll-forwards that separate new-period provision, use/write-off, releases, acquisitions/FX, and ending balance. 1. Compare estimate ratios with peers and the company's own risk indicators rather than judging reserve adequacy from the balance alone. 1. Use sensitivity disclosures and independent market inputs to create a reasonable alternative estimate and quantify EPS, FCF, capital, and valuation impact. 1. Track whether estimate changes consistently occur near compensation targets, guidance objectives, financing events, or periods of operational weakness without assuming motive. ## Required evidence and model bridge - Primary-source set: multi-year statements, auditor reports, non-GAAP reconciliations, reserve disclosures, filing changes. Preserve exact document/version, date, period, and source location for every material factual input used in accounting estimates and management bias. - For each key concept - reserves, fair value, useful lives, credit losses, warranty, returns - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as accounting estimates and management bias may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | estimate sensitivity | Change in earnings, cash flow, or equity value caused by a specified change in a key accounting estimate such as reserve rate, useful life, discount rate, or fair-value input. | estimate sensitivity: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | reserve coverage | Balance of the relevant reserve divided by the exposure it is intended to absorb, such as receivables, claims, returns, warranties, or credit losses. | reserve coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | revision bias | Average signed forecast or accounting-estimate revision over time; compare upward versus downward revisions and forecast errors for directional bias. | revision bias: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | ## Worked application > Case: warranty reserves fall despite worse early failure data. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For accounting estimates and management bias, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through reserves, fair value, useful lives, credit losses, then identify which link is directly observed and which link remains an assumption. - Calculate estimate sensitivity, reserve coverage, revision bias from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for accounting estimates and management bias: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: track estimate revisions over time and compare assumptions with observable outside evidence. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the accounting estimates and management bias conclusion. ## Failure tests - FAIL if reserves cannot be defined and reproduced from the source pack. - FAIL if a thesis-relevant estimate is accepted without sensitivity to plausible alternative assumptions and evidence on management incentives, controls, and historical accuracy. - FAIL if the accounting estimates and management bias conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the accounting estimates and management bias conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the accounting estimates and management bias conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VI - BUSINESS QUALITY | MODULE 026 # Business Model Deconstruction > Mission. Reduce the company to customers, value proposition, unit economics, pricing, cost structure, and capital requirements. ## Decision output Objective: Reduce the company to customers, value proposition, unit economics, pricing, cost structure, and capital requirements. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Identify the paying customer, end user, problem solved, alternative, purchase process, decision maker, contract/transaction mechanism, delivery channel, and switching behavior. 1. Map the revenue equation into units/customers/usage, realized price, mix, frequency, and retention; map cost into the resources required to deliver and support that revenue. 1. Locate the profit pool in the value chain and determine why this company captures it rather than suppliers, distributors, customers, or substitutes. 1. Identify the practical growth constraint: demand, sales capacity, manufacturing, qualified supply, power, permits, labor, capital, channel, or customer implementation. 1. Build unit economics and incremental returns before relying on consolidated margin or growth. 1. Write the business model in one causal page that another analyst can use to explain what creates customer value, what converts it into cash, and what can break the conversion. ## Required evidence and model bridge - Primary-source set: KPI history, customer cohorts, pricing evidence, cost structure, capital requirements. Preserve exact document/version, date, period, and source location for every material factual input used in business model deconstruction. - For each key concept - payer/user/decision maker, customer problem, value proposition, revenue mechanism, unit cost, capital - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as business model deconstruction may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | customer ROI | Customer economic benefit attributable to the product minus total customer cost, divided by total customer cost; use customer-specific operating data where possible. | customer ROI: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | gross profit per constraint unit | Incremental or total gross profit divided by the binding scarce resource, such as machine hour, MW, wafer, sales rep, bed, or square foot. | gross profit per constraint unit: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | incremental ROIC | ROIC = NOPAT / average invested capital | incremental ROIC: Recalculate incremental ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: software bookings grow but implementation capacity delays revenue. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For business model deconstruction, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through payer/user/decision maker, customer problem, value proposition, revenue mechanism, then identify which link is directly observed and which link remains an assumption. - Calculate customer ROI, gross profit per constraint unit, incremental ROIC from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for business model deconstruction: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: build the model from the customer transaction backward, not from accounting lines forward. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the business model deconstruction conclusion. ## Failure tests - FAIL if payer/user/decision maker cannot be defined and reproduced from the source pack. - FAIL if the analyst cannot state the customer, problem, purchase decision, price, delivery mechanism, gross economics, capital requirement, and binding constraint in plain language. - FAIL if the business model deconstruction conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the business model deconstruction conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the business model deconstruction conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VI - BUSINESS QUALITY | MODULE 027 # Unit Economics and Cohort Thinking > Mission. Translate growth into acquisition economics, retention, payback, contribution margin, and lifetime value. ## Decision output Objective: Translate growth into acquisition economics, retention, payback, contribution margin, and lifetime value. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define the economic unit first: customer, location, device, seat, transaction, MW, policy, account, route, well, project, or another unit that maps to value creation. 1. Build acquisition/onboarding cost from the resources required to acquire the unit, then calculate contribution or gross profit after directly attributable servicing costs. 1. Measure retention/churn by cohort and age, including contraction, expansion, reactivation, and reacquisition. Avoid mixing mature and immature cohorts in a single average. 1. Calculate CAC payback and LTV using observed cohort cash/gross-profit curves and discounting; do not rely on perpetual 1/churn shortcuts when retention is nonlinear. 1. Segment cohorts by channel, geography, customer size, product, vintage, or other economically meaningful dimension to detect deteriorating new-customer quality. 1. Translate unit economics into a company growth model: number of new units, acquisition capacity/cost, mature contribution, reinvestment needs, and incremental ROIC. ## Required evidence and model bridge - Primary-source set: KPI history, customer cohorts, pricing evidence, cost structure, capital requirements. Preserve exact document/version, date, period, and source location for every material factual input used in unit economics and cohort thinking. - For each key concept - cohort definition, CAC, payback, gross retention, net retention, contribution margin - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as unit economics and cohort thinking may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | CAC payback | CAC payback months = customer acquisition cost / monthly gross profit from new customer | CAC payback: Recalculate CAC payback from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | gross retention | GRR = beginning-cohort recurring revenue retained before expansion / beginning-cohort recurring revenue | gross retention: Recalculate gross retention from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | net retention | Recurring revenue from the opening customer cohort after churn, contraction, and expansion divided by that cohort's opening recurring revenue. | net retention: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | LTV/CAC with explicit assumptions | LTV should be constructed from cohort gross profit, retention curve, servicing cost, and discounting rather than a perpetual shortcut | LTV/CAC with explicit assumptions: Recalculate LTV/CAC with explicit assumptions from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: new cohorts weaken while mature cohorts hide the deterioration. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For unit economics and cohort thinking, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through cohort definition, CAC, payback, gross retention, then identify which link is directly observed and which link remains an assumption. - Calculate CAC payback, gross retention, net retention, LTV/CAC with explicit assumptions from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for unit economics and cohort thinking: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use cohort and survival analysis to find aggregate metrics that are masking deterioration. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the unit economics and cohort thinking conclusion. ## Failure tests - FAIL if cohort definition cannot be defined and reproduced from the source pack. - FAIL if LTV, CAC, payback, retention, or contribution margin is calculated from mismatched cohorts, periods, or cost definitions, or if acquisition growth hides deteriorating cohorts. - FAIL if the unit economics and cohort thinking conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the unit economics and cohort thinking conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the unit economics and cohort thinking conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VI - BUSINESS QUALITY | MODULE 028 # Pricing Power and Price-Volume-Mix > Mission. Separate real pricing power from inflation pass-through, mix, channel shift, or temporary shortages. ## Decision output Objective: Separate real pricing power from inflation pass-through, mix, channel shift, or temporary shortages. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build a consistent price-volume-mix bridge from source data; isolate FX, acquisition, channel, product, and geographic mix so they do not masquerade as price. 1. Distinguish announced/list price from realized net price after incentives, rebates, discounting, contract escalation, churn, and mix. 1. Test elasticity by customer cohort or segment: what happened to units, retention, competitive share, and usage after prior price changes? 1. Separate structural pricing power from inflation pass-through, supply shortage, capacity scarcity, product launch mix, and temporary promotional pullback. 1. Identify the customer's economic alternative and switching cost to determine the ceiling on sustainable price. 1. Forecast price and volume jointly, including competitive response and time lag, rather than assuming historical price increases can repeat without demand consequences. ## Required evidence and model bridge - Primary-source set: KPI history, customer cohorts, pricing evidence, cost structure, capital requirements. Preserve exact document/version, date, period, and source location for every material factual input used in pricing power and price-volume-mix. - For each key concept - realized price, units, mix, rebates, promotions, elasticity - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as pricing power and price-volume-mix may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | real price growth | real price growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | real price growth: Recalculate real price growth from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | elasticity | % change in quantity demanded divided by % change in price, measured over a comparable period and adjusted for mix, promotions, and supply constraints. | elasticity: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | PVM bridge | Revenue change decomposed into price, volume, and mix effects using a consistent base-period volume/price convention and an explicit treatment of cross terms. | PVM bridge: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | ## Worked application > Case: headline price rises 6% but mix and volume explain most of reported revenue growth. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For pricing power and price-volume-mix, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through realized price, units, mix, rebates, then identify which link is directly observed and which link remains an assumption. - Calculate real price growth, elasticity, PVM bridge from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for pricing power and price-volume-mix: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: measure realized same-product price and customer response, not list-price announcements. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the pricing power and price-volume-mix conclusion. ## Failure tests - FAIL if realized price cannot be defined and reproduced from the source pack. - FAIL if revenue growth attributed to pricing cannot be separated from inflation pass-through, volume, mix, channel, geography, incentives, and customer elasticity. - FAIL if the pricing power and price-volume-mix conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the pricing power and price-volume-mix conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the pricing power and price-volume-mix conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VI - BUSINESS QUALITY | MODULE 029 # Recurring Revenue and Retention Quality > Mission. Measure renewal behavior, churn, net retention, contract durability, and hidden re-acquisition costs. ## Decision output Objective: Measure renewal behavior, churn, net retention, contract durability, and hidden re-acquisition costs. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define what actually recurs: contractual subscription, usage relationship, maintenance, repeat purchase, installed-base service, policy renewal, or another mechanism. 1. Reconstruct gross retention, net retention, logo/customer retention, expansion, contraction, renewal term, cohort age, and price/usage effects using stable definitions. 1. Separate contractual backlog/RPO from expected economic retention; cancellation rights, renewal options, consumption risk, and implementation delays change durability. 1. Measure customer concentration, cohort maturity, product/module dependence, and reacquisition cost to understand whether recurring revenue is diversified and self-sustaining. 1. Test whether expansion revenue is genuine customer value growth or merely price increases, acquisitions, seat inflation, or product bundling. 1. Forecast installed-base revenue from beginning cohort, churn, contraction, expansion, and new additions so retention assumptions are visible and falsifiable. ## Required evidence and model bridge - Primary-source set: KPI history, customer cohorts, pricing evidence, cost structure, capital requirements. Preserve exact document/version, date, period, and source location for every material factual input used in recurring revenue and retention quality. - For each key concept - ARR/MRR, renewals, cancellation rights, usage, gross retention, net retention - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as recurring revenue and retention quality may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | gross retention | GRR = beginning-cohort recurring revenue retained before expansion / beginning-cohort recurring revenue | gross retention: Recalculate gross retention from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | net retention | Recurring revenue from the opening customer cohort after churn, contraction, and expansion divided by that cohort's opening recurring revenue. | net retention: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | renewal exposure | Recurring revenue or contract value scheduled for renewal within the selected horizon divided by total recurring revenue or contract value. | renewal exposure: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: 120% net retention hides 83% gross retention and customer concentration. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For recurring revenue and retention quality, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through ARR/MRR, renewals, cancellation rights, usage, then identify which link is directly observed and which link remains an assumption. - Calculate gross retention, net retention, renewal exposure from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for recurring revenue and retention quality: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: separate expansion from durability and track renewal calendars by cohort. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the recurring revenue and retention quality conclusion. ## Failure tests - FAIL if ARR/MRR cannot be defined and reproduced from the source pack. - FAIL if recurring revenue is judged from aggregate growth without gross retention, expansion/contraction, contract durability, reacquisition cost, and cohort behavior. - FAIL if the recurring revenue and retention quality conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the recurring revenue and retention quality conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the recurring revenue and retention quality conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VI - BUSINESS QUALITY | MODULE 030 # Capital Intensity and Reinvestment Runway > Mission. Estimate incremental returns on capital, reinvestment capacity, and limits to compounding. ## Decision output Objective: Estimate incremental returns on capital, reinvestment capacity, and limits to compounding. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Map the capital required to acquire customers, build capacity, fund working capital, develop technology, obtain licenses, maintain assets, and replace depreciating productive resources. 1. Separate maintenance from growth investment and identify off-P&L investments such as capitalized software, commissions, R&D economics, customer incentives, or acquisition spending where relevant. 1. Calculate ROIC and incremental ROIC using consistent operating profit and invested-capital definitions over a period long enough for investment to mature. 1. Estimate reinvestment capacity: how much attractive capital can the company deploy before market size, execution, regulation, infrastructure, or returns become limiting? 1. Model growth as reinvestment multiplied by incremental return, with timing lags and capacity ramps. Do not assume high historical ROIC automatically persists as the asset base expands. 1. Use the analysis to distinguish compounders with long high-return runways from businesses whose growth consumes capital at falling returns. ## Required evidence and model bridge - Primary-source set: KPI history, customer cohorts, pricing evidence, cost structure, capital requirements. Preserve exact document/version, date, period, and source location for every material factual input used in capital intensity and reinvestment runway. - For each key concept - invested capital, maintenance investment, growth investment, incremental ROIC, reinvestment rate, market runway - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as capital intensity and reinvestment runway may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | ROIC | ROIC = NOPAT / average invested capital | ROIC: Recalculate ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | incremental ROIC | ROIC = NOPAT / average invested capital | incremental ROIC: Recalculate incremental ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | reinvestment rate | Growth investment divided by after-tax operating profit, or equivalently long-run growth divided by incremental ROIC when using a value-creation framework. | reinvestment rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | sustainable growth | sustainable growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | sustainable growth: Recalculate sustainable growth from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: a 15% ROIC business can create more value than a 25% ROIC business if its reinvestment runway is much longer. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For capital intensity and reinvestment runway, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through invested capital, maintenance investment, growth investment, incremental ROIC, then identify which link is directly observed and which link remains an assumption. - Calculate ROIC, incremental ROIC, reinvestment rate, sustainable growth from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for capital intensity and reinvestment runway: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: reconcile terminal growth with required reinvestment and terminal returns. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the capital intensity and reinvestment runway conclusion. ## Failure tests - FAIL if invested capital cannot be defined and reproduced from the source pack. - FAIL if growth is forecast without the working capital, fixed assets, R&D, acquisition, customer acquisition, or other economic investment needed to support it. - FAIL if the capital intensity and reinvestment runway conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the capital intensity and reinvestment runway conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the capital intensity and reinvestment runway conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS | MODULE 031 # Industry Structure Mapping > Mission. Map value chains, profit pools, bottlenecks, substitutes, regulators, and bargaining power. ## Decision output Objective: Map value chains, profit pools, bottlenecks, substitutes, regulators, and bargaining power. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Draw the full value chain from raw input/capital to end customer, including suppliers, manufacturers/service providers, distributors, platforms, regulators, complements, and substitutes. 1. Estimate revenue and profit pools by layer and identify where bargaining power causes economics to migrate over time. 1. Map concentration, switching costs, capacity, cost curves, contract terms, standards, regulation, and customer procurement to determine the marginal price setter. 1. Identify bottlenecks and scarce complements that can capture disproportionate value even when the visible end market grows rapidly. 1. Separate structural industry economics from the current cycle by showing what changes when capacity is tight versus abundant. 1. Update the map when technology, vertical integration, new regulation, or business-model innovation changes who controls the customer or critical resource. ## Required evidence and model bridge - Primary-source set: peer filings, industry data, regulator data, channel evidence, technology roadmaps. Preserve exact document/version, date, period, and source location for every material factual input used in industry structure mapping. - For each key concept - value chain, profit pools, marginal producer, customer/supplier power, capacity, bottlenecks - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as industry structure mapping may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | profit-pool share | Company operating profit or economic profit divided by total estimated industry/value-chain profit pool for the same scope and period. | profit-pool share: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | demand/capacity | Expected demand units divided by effective industry capacity units after utilization, downtime, yields, and committed additions; track forward by period. | demand/capacity: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | concentration | Top-customer, supplier, geography, product, or channel exposure divided by the relevant total; report top-1, top-5, and HHI when data permit. | concentration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: the bottleneck in AI infrastructure moves from chips to power equipment. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For industry structure mapping, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through value chain, profit pools, marginal producer, customer/supplier power, then identify which link is directly observed and which link remains an assumption. - Calculate profit-pool share, demand/capacity, concentration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for industry structure mapping: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: track where the constraint moves because that is where incremental economics can migrate. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the industry structure mapping conclusion. ## Failure tests - FAIL if value chain cannot be defined and reproduced from the source pack. - FAIL if the value-chain map omits a layer that controls a material bottleneck, standard, customer relationship, input, or profit pool. - FAIL if the industry structure mapping conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the industry structure mapping conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the industry structure mapping conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS | MODULE 032 # Market Sizing and TAM Discipline > Mission. Build bottom-up market sizes and prevent promotional TAM estimates from contaminating valuation. ## Decision output Objective: Build bottom-up market sizes and prevent promotional TAM estimates from contaminating valuation. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define the exact product/service and buyer population before calculating TAM. Separate current market, serviceable available market, and realistic obtainable market. 1. Build bottom-up demand from number of buyers/units, penetration, replacement/adoption rate, usage, and price; show all units and conversions. 1. Use top-down industry estimates only as a cross-check and reconcile scope, geography, calendar year, nominal/real price, and channel definitions. 1. Model adoption constraints including customer ROI, budgets, infrastructure, regulation, implementation capacity, supply, and replacement cycles. 1. Avoid double counting overlapping categories, platform revenue and underlying transaction value, or multiple value-chain layers. 1. Translate company forecast into implied market share and installed-base penetration to test whether valuation assumptions require an impossible TAM path. ## Required evidence and model bridge - Primary-source set: peer filings, industry data, regulator data, channel evidence, technology roadmaps. Preserve exact document/version, date, period, and source location for every material factual input used in market sizing and tam discipline. - For each key concept - bottom-up units, penetration, replacement, normalized price, SAM, geography - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as market sizing and tam discipline may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | bottom-up TAM | Sum of addressable customers/units × realistic annual spend or units per customer, segmented by geography, use case, and adoption constraints. | bottom-up TAM: Rebuild the market denominator bottom-up from independent sources, align geography/product/time scope with the company numerator, and sensitivity-test uncertain adoption or pricing assumptions. | | serviceable market | Portion of TAM reachable with the company's current or planned product, geography, channel, regulatory approvals, and capacity within the forecast horizon. | serviceable market: Rebuild the market denominator bottom-up from independent sources, align geography/product/time scope with the company numerator, and sensitivity-test uncertain adoption or pricing assumptions. | | implied terminal share | Terminal-year company revenue or units divided by estimated terminal addressable market revenue or units. | implied terminal share: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: a $100B promotional TAM becomes a $9B serviceable market after real constraints. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For market sizing and tam discipline, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through bottom-up units, penetration, replacement, normalized price, then identify which link is directly observed and which link remains an assumption. - Calculate bottom-up TAM, serviceable market, implied terminal share from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for market sizing and tam discipline: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use terminal implied share and capacity as a valuation consistency check. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the market sizing and tam discipline conclusion. ## Failure tests - FAIL if bottom-up units cannot be defined and reproduced from the source pack. - FAIL if market size cannot be reconstructed from countable units, price/usage, adoption, and replacement without double counting or implausible customer budget assumptions. - FAIL if the market sizing and tam discipline conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the market sizing and tam discipline conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the market sizing and tam discipline conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS | MODULE 033 # Competitive Advantage and Moat Testing > Mission. Test switching costs, network effects, scale, brand, cost advantage, regulation, and data advantages. ## Decision output Objective: Test switching costs, network effects, scale, brand, cost advantage, regulation, and data advantages. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Name the claimed advantage specifically: switching cost, network effect, scale/cost, brand/trust, regulation/license, proprietary data/technology, distribution, or another mechanism. 1. Identify the measurable output the advantage should produce: retention, share stability/gain, price premium, lower acquisition cost, lower unit cost, superior incremental ROIC, or resilience in downturns. 1. Test whether the outcome persists after controlling for cycle, product age, scarcity, customer captivity, favorable mix, and accounting. 1. Estimate replication requirements: capital, time, data, distribution, certification, installed base, ecosystem, or customer migration cost needed by a competitor. 1. Identify moat decay mechanisms such as interoperability, standardization, open source, channel change, regulation, customer multi-sourcing, or technology leapfrogging. 1. Do not assign a moat label in valuation. Forecast the economic consequences and duration of the advantage explicitly. ## Required evidence and model bridge - Primary-source set: peer filings, industry data, regulator data, channel evidence, technology roadmaps. Preserve exact document/version, date, period, and source location for every material factual input used in competitive advantage and moat testing. - For each key concept - switching costs, network effects, scale, cost advantage, brand, regulation - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as competitive advantage and moat testing may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | excess ROIC | ROIC = NOPAT / average invested capital | excess ROIC: Recalculate excess ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | share stability | Dispersion and trend of market share across periods and end markets; quantify standard deviation, drawdowns, and recovery rather than relying on one point. | share stability: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | switching friction | Quantified customer cost/time/risk to migrate, using implementation expense, downtime, retraining, data migration, contract penalties, or observed churn after competitive entry. | switching friction: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: high margins exist even though customers can dual-source and switch quickly. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For competitive advantage and moat testing, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through switching costs, network effects, scale, cost advantage, then identify which link is directly observed and which link remains an assumption. - Calculate excess ROIC, share stability, switching friction from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for competitive advantage and moat testing: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: state the causal defense mechanism and define the evidence that would show it is weakening. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the competitive advantage and moat testing conclusion. ## Failure tests - FAIL if switching costs cannot be defined and reproduced from the source pack. - FAIL if a claimed moat is based on market share, brand, patents, or margins without a causal barrier to replication and evidence of durable excess incremental returns. - FAIL if the competitive advantage and moat testing conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the competitive advantage and moat testing conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the competitive advantage and moat testing conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS | MODULE 034 # Competitor Benchmarking > Mission. Normalize peers across accounting, KPIs, end markets, geography, and capital structure. ## Decision output Objective: Normalize peers across accounting, KPIs, end markets, geography, and capital structure. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Choose peers based on business economics and customer alternatives, not only industry classification or management-selected comps. 1. Create a normalization sheet for fiscal calendar, currency, acquisitions/divestitures, lease treatment, SBC, capitalization, segment definitions, non-GAAP adjustments, and capital structure. 1. Compare operating KPIs, price/volume/mix, margins, incremental margins, working capital, reinvestment, ROIC, balance sheet, and valuation using matched definitions. 1. Separate structural difference from lifecycle/cycle position. A higher margin may reflect mix, geography, utilization, or underinvestment rather than superior execution. 1. Use a time series, not one snapshot, to identify who gains share, who converts growth to cash, and who revises definitions when performance deteriorates. 1. End with causal differences that explain valuation dispersion rather than a table of ratios with no mechanism. ## Required evidence and model bridge - Primary-source set: peer filings, industry data, regulator data, channel evidence, technology roadmaps. Preserve exact document/version, date, period, and source location for every material factual input used in competitor benchmarking. - For each key concept - peer selection, KPI definitions, accounting normalization, geography/product mix, cycle position, capital structure - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as competitor benchmarking may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | normalized margin spread | normalized margin spread = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | normalized margin spread: Recalculate normalized margin spread from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | organic growth quality | organic growth quality = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | organic growth quality: Recalculate organic growth quality from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | valuation residual | Observed valuation premium/discount remaining after controlling for growth, margins, capital intensity, risk, and balance-sheet differences versus peers. | valuation residual: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: two 30% growers have very different organic growth and retention after normalization. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For competitor benchmarking, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through peer selection, KPI definitions, accounting normalization, geography/product mix, then identify which link is directly observed and which link remains an assumption. - Calculate normalized margin spread, organic growth quality, valuation residual from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for competitor benchmarking: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: keep raw and normalized peer views side by side so every adjustment can be reversed. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the competitor benchmarking conclusion. ## Failure tests - FAIL if peer selection cannot be defined and reproduced from the source pack. - FAIL if peer comparisons mix incompatible KPI definitions, accounting policies, capital structures, geography, product mix, or cycle positions without normalization. - FAIL if the competitor benchmarking conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the competitor benchmarking conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the competitor benchmarking conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VII - INDUSTRY AND COMPETITIVE ANALYSIS | MODULE 035 # Disruption and Technology S-Curves > Mission. Analyze adoption curves, cost declines, standards, infrastructure constraints, and incumbent response. ## Decision output Objective: Analyze adoption curves, cost declines, standards, infrastructure constraints, and incumbent response. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Break adoption into technical readiness, cost/performance competitiveness, supply capacity, infrastructure, standards, regulation, customer workflow change, and replacement cycle. 1. Estimate the addressable installed base and realistic annual conversion capacity; an S-curve cannot exceed physical supply, skilled labor, power, permits, customer budgets, or implementation capacity. 1. Track learning curves and cost decline using cumulative volume where evidence supports it, but identify floors set by materials, energy, labor, or mature-process economics. 1. Model incumbent response: price cuts, bundling, self-cannibalization, acquisition, standards influence, distribution leverage, or capital redeployment. 1. Distinguish revenue adoption from profit-pool migration. A rapidly growing technology can destroy industry profit if supply enters faster than differentiation. 1. Create early, middle, and late adoption scenarios with explicit triggers that move the probability or economic outcome. ## Required evidence and model bridge - Primary-source set: peer filings, industry data, regulator data, channel evidence, technology roadmaps. Preserve exact document/version, date, period, and source location for every material factual input used in disruption and technology s-curves. - For each key concept - performance, TCO, cost parity, adoption, standards, infrastructure - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as disruption and technology s-curves may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | adoption penetration | Installed/active units or users divided by the realistically addressable population or installed base for the relevant use case. | adoption penetration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | cost-parity gap | Incumbent alternative total cost of ownership minus new-technology TCO, expressed in absolute terms and as a % of incumbent TCO. | cost-parity gap: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | learning rate | Percentage decline in unit cost for each doubling of cumulative production; estimate from a log-log regression of cost on cumulative output. | learning rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: superior battery energy density fails to produce adoption because yield and infrastructure lag. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For disruption and technology s-curves, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through performance, TCO, cost parity, adoption, then identify which link is directly observed and which link remains an assumption. - Calculate adoption penetration, cost-parity gap, learning rate from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for disruption and technology s-curves: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: separate technical readiness, manufacturability, bankability, and customer adoption. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the disruption and technology s-curves conclusion. ## Failure tests - FAIL if performance cannot be defined and reproduced from the source pack. - FAIL if adoption is forecast from technical improvement alone without customer economics, infrastructure, standards, supply, incumbent response, and substitution risk. - FAIL if the disruption and technology s-curves conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the disruption and technology s-curves conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the disruption and technology s-curves conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VIII - MANAGEMENT AND GOVERNANCE | MODULE 036 # Management Quality Assessment > Mission. Evaluate capital allocation, operational execution, candor, consistency, and incentives using evidence rather than charisma. ## Decision output Objective: Evaluate capital allocation, operational execution, candor, consistency, and incentives using evidence rather than charisma. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build a five-to-ten-year chronology of guidance, strategic promises, acquisitions, divestitures, capital returns, restructurings, and operating targets. 1. Score outcomes against what management controlled, separating industry tailwinds/headwinds from execution. 1. Compare stated capital-allocation priorities with actual cash deployment and subsequent returns, including opportunity cost. 1. Track candor: whether management quantifies misses, preserves definitions, discloses bad news promptly, and changes strategy with evidence rather than narrative. 1. Evaluate organization depth, succession, key-person dependence, operating metrics, and evidence that management can scale systems as the company grows. 1. Use documented behavior, not charisma, conference-call style, or share-price performance, as the basis for the assessment. ## Required evidence and model bridge - Primary-source set: proxy statements, ownership forms, capital-allocation history, board records, guidance history. Preserve exact document/version, date, period, and source location for every material factual input used in management quality assessment. - For each key concept - guidance history, target changes, capital allocation, candor, KPI stability, execution - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as management quality assessment may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | guidance calibration | Historical actual result minus management guidance midpoint/range, scaled by guidance width or revenue/earnings; track bias and hit rate. | guidance calibration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | allocation returns | Incremental NOPAT or FCF attributable to a capital-allocation decision divided by capital deployed, measured over an appropriate post-investment horizon. | allocation returns: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | definition stability | Count and materiality of KPI/accounting-definition changes across periods; flag restatements, exclusions, denominator changes, and loss of comparability. | definition stability: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: quarterly EPS beats coexist with poor long-term acquisition ROIC. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For management quality assessment, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through guidance history, target changes, capital allocation, candor, then identify which link is directly observed and which link remains an assumption. - Calculate guidance calibration, allocation returns, definition stability from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for management quality assessment: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: score what management controlled and separate charisma from decision quality. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the management quality assessment conclusion. ## Failure tests - FAIL if guidance history cannot be defined and reproduced from the source pack. - FAIL if management is rated from charisma or stock performance rather than a dated record of controllable decisions, outcomes, disclosure behavior, and accountability. - FAIL if the management quality assessment conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the management quality assessment conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the management quality assessment conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VIII - MANAGEMENT AND GOVERNANCE | MODULE 037 # Capital Allocation Scorecard > Mission. Evaluate reinvestment, M&A, buybacks, dividends, debt policy, and balance-sheet optionality. ## Decision output Objective: Evaluate reinvestment, M&A, buybacks, dividends, debt policy, and balance-sheet optionality. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct sources and uses of capital over at least five years: operating cash, debt/equity issuance, organic reinvestment, R&D, acquisitions, divestitures, repurchases, dividends, and debt reduction. 1. Estimate returns on major reinvestment programs and acquisitions after maturation, using incremental NOPAT/FCF and capital invested rather than management-adjusted accretion alone. 1. Evaluate repurchases against intrinsic value, dilution offset, leverage, and alternative uses; a lower share count does not prove value creation. 1. Assess dividend policy and debt targets against business cyclicality, fixed commitments, refinancing windows, and investment opportunities. 1. Quantify optionality: cash, borrowing capacity, valuable non-operating assets, and ability to invest during downturns without destructive financing. 1. Score process and economics separately so a good outcome caused by luck does not mask a poor capital-allocation decision. ## Required evidence and model bridge - Primary-source set: proxy statements, ownership forms, capital-allocation history, board records, guidance history. Preserve exact document/version, date, period, and source location for every material factual input used in capital allocation scorecard. - For each key concept - organic reinvestment, R&D, M&A, buybacks, dividends, debt - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as capital allocation scorecard may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | incremental ROIC | ROIC = NOPAT / average invested capital | incremental ROIC: Recalculate incremental ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | realized deal return | Post-acquisition incremental NOPAT/FCF plus realized synergies minus integration/restructuring costs, divided by purchase price and incremental capital invested. | realized deal return: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | net share retirement | Beginning diluted shares minus ending diluted shares, adjusted for splits, divided by beginning diluted shares; reconcile gross repurchases and issuance. | net share retirement: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | leverage through cycle | Net debt or debt-like obligations divided by normalized EBITDA/FCF across peak, mid-cycle, and trough scenarios. | leverage through cycle: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Worked application > Case: buybacks mostly offset SBC while attractive internal projects are underfunded. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For capital allocation scorecard, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through organic reinvestment, R&D, M&A, buybacks, then identify which link is directly observed and which link remains an assumption. - Calculate incremental ROIC, realized deal return, net share retirement, leverage through cycle from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for capital allocation scorecard: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: measure allocation over full cycles and compare actions with contemporaneous value, not hindsight. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the capital allocation scorecard conclusion. ## Failure tests - FAIL if organic reinvestment cannot be defined and reproduced from the source pack. - FAIL if reinvestment, M&A, debt, dividends, buybacks, and cash are not compared on expected risk-adjusted return, resilience, and opportunity cost. - FAIL if the capital allocation scorecard conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the capital allocation scorecard conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the capital allocation scorecard conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VIII - MANAGEMENT AND GOVERNANCE | MODULE 038 # Compensation and Incentive Analysis > Mission. Map executive pay metrics to behaviors, accounting choices, time horizons, and shareholder outcomes. ## Decision output Objective: Map executive pay metrics to behaviors, accounting choices, time horizons, and shareholder outcomes. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Map annual and long-term incentive metrics, weights, thresholds, targets, maximums, vesting periods, relative measures, and board discretion. 1. Compare compensation metrics with the variables that actually create long-term value, such as ROIC, durable FCF per share, retention, or risk-adjusted growth, not merely revenue or adjusted EPS. 1. Identify ways a metric can be optimized cosmetically through buybacks, acquisitions, cost capitalization, target resets, one-time exclusions, or end-of-period working-capital actions. 1. Evaluate equity award dilution, grant timing, retention awards, change-in-control provisions, severance, and ownership/holding requirements. 1. Compare realizable outcomes with business outcomes over multiple cycles and inspect whether poor results produce real downside for management. 1. Model how the incentive plan could rationally influence behavior under both upside and downside scenarios. ## Required evidence and model bridge - Primary-source set: proxy statements, ownership forms, capital-allocation history, board records, guidance history. Preserve exact document/version, date, period, and source location for every material factual input used in compensation and incentive analysis. - For each key concept - metric selection, target curves, horizon, discretion, TSR, ROIC - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as compensation and incentive analysis may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | economic alignment | Fraction of incentive pay and insider wealth exposure tied to per-share value creation, ROIC, FCF, or other durable economics rather than easily managed volume metrics. | economic alignment: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | payout asymmetry | Expected compensation upside for exceeding targets relative to downside for missing them, measured around threshold, target, and maximum payout points. | payout asymmetry: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | realized dilution cost | Net dilution = ending diluted share count / beginning diluted share count - 1, adjusted for major capital actions | realized dilution cost: Recalculate realized dilution cost from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## Worked application > Case: adjusted EBITDA compensation rewards acquisitions without a capital charge. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For compensation and incentive analysis, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through metric selection, target curves, horizon, discretion, then identify which link is directly observed and which link remains an assumption. - Calculate economic alignment, payout asymmetry, realized dilution cost from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for compensation and incentive analysis: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: model the behavior each metric can rationally encourage. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the compensation and incentive analysis conclusion. ## Failure tests - FAIL if metric selection cannot be defined and reproduced from the source pack. - FAIL if pay metrics are accepted at face value without reconstructing definitions, thresholds, discretion, and the behaviors the plan economically rewards. - FAIL if the compensation and incentive analysis conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the compensation and incentive analysis conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the compensation and incentive analysis conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VIII - MANAGEMENT AND GOVERNANCE | MODULE 039 # Insider Ownership and Transactions > Mission. Interpret insider ownership, buying, selling, grants, 10b5-1 plans, and dilution in context. ## Decision output Objective: Interpret insider ownership, buying, selling, grants, 10b5-1 plans, and dilution in context. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconcile beneficial ownership from proxy/13D/13G where applicable with Forms 3, 4, and 5 and the equity compensation tables. 1. Classify transactions into open-market purchases/sales, option exercises, tax withholding, gifts, grants, estate/planning transfers, and 10b5-1 plan transactions before interpreting signal. 1. Measure transaction size relative to existing ownership, compensation, liquidity needs, and historical pattern rather than focusing on raw dollars. 1. Track adoption/modification of trading plans and transaction timing around material corporate events while avoiding unsupported claims about motive. 1. Calculate net ownership and dilution over time for founders, executives, board, and significant holders. 1. Treat insider activity as contextual evidence, not a standalone valuation thesis; open-market buying can be more informative than routine compensation-related selling, but all cases require context. ## Required evidence and model bridge - Primary-source set: proxy statements, ownership forms, capital-allocation history, board records, guidance history. Preserve exact document/version, date, period, and source location for every material factual input used in insider ownership and transactions. - For each key concept - Forms 3/4/5, 10b5-1, open-market trades, grants, tax withholding, option exercises - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as insider ownership and transactions may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | net discretionary flow | Discretionary capital returned or invested, buybacks + dividends + M&A + growth investments, net of financing, classified by expected return and timing. | net discretionary flow: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | ownership at risk | Executive/director beneficial ownership value exposed to common-stock performance, net of immediately hedgeable or near-term vesting amounts, relative to annual compensation/net worth where disclosed. | ownership at risk: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | sale concentration | Insider shares sold over the selected window divided by beginning beneficial ownership, with 10b5-1, tax sales, option exercises, and discretionary sales separated. | sale concentration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: a large sale is mostly tax and option-related under a preexisting plan. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For insider ownership and transactions, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through Forms 3/4/5, 10b5-1, open-market trades, grants, then identify which link is directly observed and which link remains an assumption. - Calculate net discretionary flow, ownership at risk, sale concentration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for insider ownership and transactions: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: classify transaction mechanics before interpreting signaling. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the insider ownership and transactions conclusion. ## Failure tests - FAIL if Forms 3/4/5 cannot be defined and reproduced from the source pack. - FAIL if insider buying or selling is interpreted without separating purchased shares, compensation, tax sales, plans, dilution, transaction size, and context. - FAIL if the insider ownership and transactions conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the insider ownership and transactions conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the insider ownership and transactions conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART VIII - MANAGEMENT AND GOVERNANCE | MODULE 040 # Board and Governance Risk > Mission. Evaluate independence, tenure, expertise, related parties, committee structure, and control weaknesses. ## Decision output Objective: Evaluate independence, tenure, expertise, related parties, committee structure, and control weaknesses. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Map board independence, tenure, refreshment, diversity of relevant expertise, committee memberships, chair/CEO structure, lead independent director, and director workload. 1. Match board skills to company-specific risks such as regulated operations, technology, cybersecurity, capital intensity, complex accounting, M&A, or global supply chains. 1. Review audit, compensation, and nominating/governance committee charters and evidence of oversight failures, material weaknesses, related parties, or repeated target resets. 1. Assess control structure, dual-class shares, shareholder rights, poison pills, supermajority provisions, special meeting/written consent rights, and takeover defenses. 1. Track related-party transactions, family/business relationships, director interlocks, and significant shareholder influence. 1. Use governance as a risk multiplier or mitigant tied to concrete decisions and controls, not as a generic ESG score. ## Required evidence and model bridge - Primary-source set: proxy statements, ownership forms, capital-allocation history, board records, guidance history. Preserve exact document/version, date, period, and source location for every material factual input used in board and governance risk. - For each key concept - skills, independence, tenure, interlocks, dual class, control - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as board and governance risk may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | skill coverage | % of board/management critical competencies represented by at least one demonstrably qualified member, weighted by company-specific importance. | skill coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | voting/economic wedge | Voting power percentage minus economic ownership percentage for controlling holders or share classes. | voting/economic wedge: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | governance incident rate | Material governance/control incidents per year or reporting period, classified by severity and recurrence. | governance incident rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: founder controls 70% of votes with 12% economic ownership. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For board and governance risk, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through skills, independence, tenure, interlocks, then identify which link is directly observed and which link remains an assumption. - Calculate skill coverage, voting/economic wedge, governance incident rate from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for board and governance risk: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: translate governance into economic scenarios rather than a generic governance score. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the board and governance risk conclusion. ## Failure tests - FAIL if skills cannot be defined and reproduced from the source pack. - FAIL if formal independence is treated as sufficient without evaluating voting control, tenure, expertise, related parties, committees, succession, and actual decision rights. - FAIL if the board and governance risk conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the board and governance risk conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the board and governance risk conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IX - MODEL BUILDING | MODULE 041 # Model Architecture and Standards > Mission. Build models that are transparent, auditable, modular, scenario-ready, and resistant to hard-coded errors. ## Decision output Objective: Build models that are transparent, auditable, modular, scenario-ready, and resistant to hard-coded errors. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Design the workbook/data model in a one-way flow from sources and historicals to assumptions, operating schedules, integrated statements, scenarios, valuation, dashboard, and checks. 1. Separate reported data, analyst normalization, forecast assumptions, formulas, and outputs visually and structurally. Every material input needs source/date/comment. 1. Use dedicated schedules for revenue drivers, costs, fixed assets, working capital, debt/interest, taxes, shares/dilution, acquisitions, and other material mechanics. 1. Eliminate unexplained plugs and accidental circularity. Document intentional circular calculations and use algebraic or controlled iterative solutions where necessary. 1. Build visible error checks for statement balance, cash roll-forward, segment totals, share roll-forward, debt maturities, signs/units, scenario selection, and stale inputs. 1. Optimize for auditability and change control, not clever formula compression. A reviewer should trace any output to its source and assumption quickly. ## Required evidence and model bridge - Primary-source set: normalized historicals, KPI bridges, driver assumptions, debt/share schedules, source notes. Preserve exact document/version, date, period, and source location for every material factual input used in model architecture and standards. - For each key concept - raw data, normalized history, assumptions, calculations, statements, schedules - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as model architecture and standards may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | unexplained hardcode count | Count of forecast/model cells containing hard-coded values without an identified source, assumption label, or documented rationale. | unexplained hardcode count: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | check coverage | % of material model linkages and outputs covered by automated balance, cash, share-count, debt, tax, scenario, and valuation error checks. | check coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | trace depth | Number of auditable steps from a decision output back to primary-source data; lower and fully documented is better, with no hidden transformations. | trace depth: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | ## Model architecture standard - Recommended workbook flow: Cover/Control -> Sources -> Historical statements -> KPI/segments -> Assumptions -> Operating forecast -> IS -> BS -> CF -> Debt/interest -> Shares -> Taxes -> Scenarios -> Valuation -> Dashboard -> Checks. - Inputs must be visually and structurally separated from formulas. Each input needs source/date/comment. Never bury an assumption inside a formula that reviewers cannot identify. - Add error flags for balance-sheet balance, cash roll-forward, segment-consolidated reconciliation, share roll-forward, debt maturity, circularity, sign conventions, scenario selection, and valuation bridge. ## Worked application > Case: revenue is hardcoded independently in three places and silently diverges. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For model architecture and standards, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through raw data, normalized history, assumptions, calculations, then identify which link is directly observed and which link remains an assumption. - Calculate unexplained hardcode count, check coverage, trace depth from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for model architecture and standards: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: one driver should flow through all downstream schedules and outputs. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the model architecture and standards conclusion. ## Failure tests - FAIL if raw data cannot be defined and reproduced from the source pack. - FAIL if source data, assumptions, formulas, outputs, scenarios, and checks are mixed so a reviewer cannot trace or safely change the model. - FAIL if the model architecture and standards conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the model architecture and standards conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the model architecture and standards conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IX - MODEL BUILDING | MODULE 042 # Historical Data Normalization > Mission. Recast reported data into consistent periods, segments, KPIs, and economic definitions. ## Decision output Objective: Recast reported data into consistent periods, segments, KPIs, and economic definitions. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build a reported historical layer exactly as filed before creating any normalized series. Preserve amendments and document which filing version is authoritative. 1. Create a mapping from filing rows/footnotes and XBRL concepts to model rows, including units, signs, fiscal period, segment, and scope. 1. Bridge acquisitions, divestitures, discontinued operations, fiscal-year changes, stock splits, segment reorganizations, accounting-standard adoption, and definition changes. 1. Normalize nonrecurring items only in a separate layer, preserving both GAAP/reporting history and the analyst adjustment with rationale. 1. Reconcile segment totals, KPI history, share count, cash, debt, and equity to consolidated statements and footnotes. 1. Use the normalized history to calculate driver relationships only after confirming the series is comparable over time. ## Required evidence and model bridge - Primary-source set: normalized historicals, KPI bridges, driver assumptions, debt/share schedules, source notes. Preserve exact document/version, date, period, and source location for every material factual input used in historical data normalization. - For each key concept - reported history, restatements, recasts, M&A, discontinued ops, FX - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as historical data normalization may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | normalization bridge | Reported historical metric plus/minus explicitly itemized accounting, one-time, acquisition, FX, and definition adjustments to reach the normalized series. | normalization bridge: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | | organic growth | organic growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | organic growth: Recalculate organic growth from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | history integrity | % of historical periods that tie to filed/reported data after restatements and definition changes, with zero unexplained gaps in material series. | history integrity: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Historical normalization laboratory - Create a reported layer that exactly matches filings before creating an adjusted layer. Never overwrite reported history with analyst normalization. - Bridge acquisitions, divestitures, discontinued operations, fiscal-calendar changes, segment reorganizations, stock splits, accounting-standard adoption, and definition changes explicitly. - Preserve a mapping table from each modeled row to filing statement/footnote and source date so another analyst can reproduce the history. ## Worked application > Case: segment definitions change after an acquisition and only two years are recast. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For historical data normalization, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through reported history, restatements, recasts, M&A, then identify which link is directly observed and which link remains an assumption. - Calculate normalization bridge, organic growth, history integrity from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for historical data normalization: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: preserve raw reported history and make every normalization reversible. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the historical data normalization conclusion. ## Failure tests - FAIL if reported history cannot be defined and reproduced from the source pack. - FAIL if reported history is overwritten by adjusted data or comparability changes are made without a reversible bridge to original filings. - FAIL if the historical data normalization conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the historical data normalization conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the historical data normalization conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IX - MODEL BUILDING | MODULE 043 # Revenue Forecasting by Driver > Mission. Forecast units, customers, capacity, utilization, price, mix, or other business-specific drivers. ## Decision output Objective: Forecast units, customers, capacity, utilization, price, mix, or other business-specific drivers. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Choose the shortest causal revenue equation available: units x price, customers x ARPU, capacity x utilization x price, stores x sales/store, users x monetization, MW x realization, or another native model. 1. Rebuild historical revenue from the selected drivers and explain residuals through mix, FX, acquisitions, geography, accounting scope, or measurement error. 1. Forecast demand and the practical ability to serve it separately. Constrain revenue by salesforce, manufacturing, qualified supply, installed capacity, power, permits, channel, customer implementation, or capital where applicable. 1. Model price, volume, mix, churn, and new business as separate assumptions when they respond differently to competition and macro conditions. 1. Use backlog/RPO/orders only after modeling cancellation, timing, conversion, renewal, and capacity. 1. Tie each forecast driver to a dated evidence source and create a high/low range based on historical error and scenario conditions. ## Required evidence and model bridge - Primary-source set: normalized historicals, KPI bridges, driver assumptions, debt/share schedules, source notes. Preserve exact document/version, date, period, and source location for every material factual input used in revenue forecasting by driver. - For each key concept - units/customers, price, mix, usage, retention, capacity - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as revenue forecasting by driver may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | driver identity | % of forecast revenue/cost lines linked to explicit economic drivers rather than simple growth-rate extrapolation or plugs. | driver identity: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capacity headroom | Effective available capacity less forecast demand, expressed in units and as a % of effective capacity after utilization/yield constraints. | capacity headroom: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | implied share | Forecast company units or revenue divided by forecast market units or revenue for the same definition, geography, and period. | implied share: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Driver-based revenue laboratory - Use the most causal available architecture: units x price, customers x ARPU, capacity x utilization x price, stores x sales/store, users x engagement x monetization, or project MW x realization. Avoid consolidated percentage guesses when drivers exist. - Reconcile the driver model back to reported revenue for every historical period. The residual should be explained as mix, FX, M&A, scope, or measurement error. - Forecast bottlenecks before demand. A demand forecast is not executable if sales capacity, manufacturing, power, permits, supply, or financing cannot support it. ## Worked application > Case: semiconductor revenue is units x ASP constrained by fab, yield, packaging, and demand. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For revenue forecasting by driver, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through units/customers, price, mix, usage, then identify which link is directly observed and which link remains an assumption. - Calculate driver identity, capacity headroom, implied share from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for revenue forecasting by driver: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: build bull/base/bear by changing drivers, not arbitrary CAGR percentages. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the revenue forecasting by driver conclusion. ## Failure tests - FAIL if units/customers cannot be defined and reproduced from the source pack. - FAIL if revenue is forecast from a top-line percentage when units, customers, capacity, price, mix, retention, or another observable economic driver is available and material. - FAIL if the revenue forecasting by driver conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the revenue forecasting by driver conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the revenue forecasting by driver conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IX - MODEL BUILDING | MODULE 044 # Margin and Cost Forecasting > Mission. Model contribution margin, fixed-cost absorption, operating leverage, and cost actions. ## Decision output Objective: Model contribution margin, fixed-cost absorption, operating leverage, and cost actions. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Rebuild historical gross and operating margin using price, volume, mix, utilization, input cost, labor, freight, FX, productivity, capacity additions, and restructuring. 1. Classify costs by economic behavior and model the driver of each material cost rather than applying a flat percent of revenue where causality is observable. 1. Estimate incremental margins over comparable periods and explicitly model step costs as capacity or organizational layers are added. 1. Separate structural productivity from temporary under-absorption, over-absorption, shortage premiums, favorable mix, and delayed hiring/spend. 1. Build price-cost lag and contract escalators for businesses exposed to volatile inputs or annual repricing. 1. Run margin scenarios jointly with demand and mix so downside cases do not assume impossible cost flexibility. ## Required evidence and model bridge - Primary-source set: normalized historicals, KPI bridges, driver assumptions, debt/share schedules, source notes. Preserve exact document/version, date, period, and source location for every material factual input used in margin and cost forecasting. - For each key concept - variable cost, fixed cost, headcount, utilization, input prices, mix - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as margin and cost forecasting may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | incremental gross margin | Gross margin = gross profit / revenue | incremental gross margin: Recalculate incremental gross margin from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | incremental EBIT margin | incremental EBIT margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | incremental EBIT margin: Recalculate incremental EBIT margin from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | utilization sensitivity | Change in revenue, margin, EBITDA, or FCF for a specified change in utilization, holding price/mix and other drivers constant. | utilization sensitivity: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Margin architecture laboratory - Separate truly variable cost, semi-variable cost, fixed cost, pass-through cost, stock compensation, depreciation, and step-function capacity cost. - Estimate incremental gross and operating margin from historical volume/mix changes and management capacity actions. Do not force consolidated historical margin onto a materially different mix. - Build price-cost lag explicitly for businesses with commodity inputs, labor contracts, fuel surcharges, or contractual escalators. ## Worked application > Case: revenue growth assumes implausible SG&A leverage despite required sales hiring. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For margin and cost forecasting, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through variable cost, fixed cost, headcount, utilization, then identify which link is directly observed and which link remains an assumption. - Calculate incremental gross margin, incremental EBIT margin, utilization sensitivity from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for margin and cost forecasting: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: link cost pools to the same operating drivers that create revenue. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the margin and cost forecasting conclusion. ## Failure tests - FAIL if variable cost cannot be defined and reproduced from the source pack. - FAIL if margin expansion or contraction is modeled without explicit variable cost, fixed cost, utilization, labor, input, mix, or cost-action mechanics. - FAIL if the margin and cost forecasting conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the margin and cost forecasting conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the margin and cost forecasting conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART IX - MODEL BUILDING | MODULE 045 # Balance Sheet and Cash Flow Forecasting > Mission. Forecast working capital, capex, financing, share count, and cash without circular mistakes. ## Decision output Objective: Forecast working capital, capex, financing, share count, and cash without circular mistakes. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Forecast receivables, inventory, payables, deferred revenue, and other operating balances from driver days/turns or contract mechanics; do not use cash as a plug. 1. Roll PP&E and intangibles from beginning balances through capex/acquisitions, depreciation/amortization, disposals, FX, and impairments. 1. Build a debt schedule by instrument with maturity, rate/floating benchmark, amortization, revolver, covenants, and refinancing assumption; link interest to average balances and rates. 1. Forecast cash taxes, dividends, repurchases/issuance, SBC dilution, leases/debt-like commitments, and acquisition consideration explicitly when material. 1. Let ending cash emerge from integrated operating/investing/financing flows, then trigger revolver/equity/other financing only under an explicit minimum-cash policy. 1. Require balance sheet balance and cash roll-forward in every scenario, including stress. ## Required evidence and model bridge - Primary-source set: normalized historicals, KPI bridges, driver assumptions, debt/share schedules, source notes. Preserve exact document/version, date, period, and source location for every material factual input used in balance sheet and cash flow forecasting. - For each key concept - working capital, PP&E, depreciation, debt, rates, cash taxes - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as balance sheet and cash flow forecasting may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | minimum cash headroom | Lowest forecast unrestricted cash plus committed undrawn liquidity minus minimum operating cash requirement across the modeled horizon. | minimum cash headroom: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | net leverage | Net leverage = net debt / normalized EBITDA, with leases and other debt-like items treated consistently | net leverage: Recalculate net leverage from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | FCF conversion | Free cash flow divided by the relevant earnings base, typically net income, EBIT, or EBITDA; define consistently and reconcile every adjustment. | FCF conversion: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | ## Integrated forecast laboratory - Forecast working-capital balances from operating drivers, PP&E from beginning balance plus capex less depreciation/disposals, debt from contractual maturities and financing needs, and shares from awards/issuance/repurchases. - Interest expense should be linked to average debt and rates; cash interest income to average cash and yields; taxes to taxable economics; cash should be the residual output of integrated statements, not a plug. - Use a revolver or explicit financing decision only when cash would fall below a defined minimum. The model must explain how the company survives a stress case. ## Worked application > Case: positive EBITDA company runs out of cash because working capital and maturities consume liquidity. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For balance sheet and cash flow forecasting, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through working capital, PP&E, depreciation, debt, then identify which link is directly observed and which link remains an assumption. - Calculate minimum cash headroom, net leverage, FCF conversion from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for balance sheet and cash flow forecasting: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: cash is an output of the integrated model, not a plug that makes the balance sheet balance. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the balance sheet and cash flow forecasting conclusion. ## Failure tests - FAIL if working capital cannot be defined and reproduced from the source pack. - FAIL if forecast cash, working capital, PP&E, debt, taxes, and equity do not roll from operating assumptions and the statements require a balancing plug. - FAIL if the balance sheet and cash flow forecasting conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the balance sheet and cash flow forecasting conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the balance sheet and cash flow forecasting conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART X - VALUATION | MODULE 046 # DCF from First Principles > Mission. Value the operating asset base using explicit cash flows, reinvestment, terminal economics, and transparent discounting. ## Decision output Objective: Value the operating asset base using explicit cash flows, reinvestment, terminal economics, and transparent discounting. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Forecast FCFF from operating drivers: NOPAT plus noncash operating charges less fixed-asset, working-capital, and other operating reinvestment. Do not start from a target multiple. 1. Set the explicit forecast horizon long enough for key growth, margin, capital intensity, and competitive advantages to move toward a defensible steady state. 1. Estimate cost of capital using market-value capital weights and risk inputs consistent with the cash-flow currency, business risk, and leverage; avoid arbitrary premia or double counting. 1. Discount cash flows using timing appropriate to the valuation date, including stub periods and midyear convention when cash is earned through the year. 1. Make terminal economics internally consistent: terminal growth requires reinvestment when returns are finite, and terminal ROIC should reflect mature competitive economics. 1. Reconcile terminal value to an implied multiple and complete the enterprise-to-equity bridge for debt, leases/debt-like claims, pensions, preferred, minority interests, excess cash, investments, options/awards, convertibles, and other material claims. ## Required evidence and model bridge - Primary-source set: normalized forecasts, capital structure, market data, peer definitions, scenario assumptions. Preserve exact document/version, date, period, and source location for every material factual input used in dcf from first principles. - For each key concept - normalized FCFF, WACC, explicit horizon, midyear convention, reinvestment, ROIC fade - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as dcf from first principles may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | FCFF | FCFF = EBIT x (1 - cash tax rate) + D&A - capex - change in NWC - other required operating investment | FCFF: Recalculate FCFF from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | WACC | WACC = E/(D+E) x cost of equity + D/(D+E) x after-tax cost of debt | WACC: Recalculate WACC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | g=reinvestment x ROIC | ROIC = NOPAT / average invested capital | g=reinvestment x ROIC: Recalculate g=reinvestment x ROIC from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | terminal value concentration | Present value of terminal value divided by total enterprise value in the DCF. | terminal value concentration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## DCF from first principles laboratory - Build FCFF from operating economics, not from a shortcut EBITDA multiple: NOPAT + noncash operating charges - reinvestment in fixed assets, working capital, and other operating assets. - Estimate cost of capital using current market value weights and risk inputs that match currency and cash-flow risk. Do not add arbitrary company-specific premia without explaining double counting. - Use midyear discounting when cash flows occur through the year. For a stub period, discount by actual fraction of year rather than pretending the valuation date is fiscal year-end. - Terminal consistency is mandatory. If terminal growth is 3% and terminal ROIC is 12%, the implied reinvestment rate is 25%. A model that assumes 3% perpetual growth with zero reinvestment and finite ROIC is internally inconsistent. - Reconcile terminal value using both perpetuity-growth economics and an implied terminal multiple. If the implied multiple is implausible versus mature economics, revisit the operating assumptions. - Complete the EV-to-equity bridge with debt, leases/debt-like claims, pension deficits, minority interests, non-operating investments, excess cash, preferred claims, options/awards, convertibles, and other material claims. ## Worked application > Case: 3% terminal growth requires reinvestment if terminal ROIC is finite. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For dcf from first principles, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through normalized FCFF, WACC, explicit horizon, midyear convention, then identify which link is directly observed and which link remains an assumption. - Calculate FCFF, WACC, g=reinvestment x ROIC, terminal value concentration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for dcf from first principles: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: make terminal growth, reinvestment, returns, and discounting mathematically consistent. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the dcf from first principles conclusion. ## Failure tests - FAIL if normalized FCFF cannot be defined and reproduced from the source pack. - FAIL if terminal growth, reinvestment, mature returns, discount rate, timing, enterprise-to-equity bridge, or dilution are internally inconsistent or unsupported. - FAIL if the dcf from first principles conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the dcf from first principles conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the dcf from first principles conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART X - VALUATION | MODULE 047 # Multiples and Relative Valuation > Mission. Choose metrics that fit economic reality, normalize peer differences, and avoid mechanically applying averages. ## Decision output Objective: Choose metrics that fit economic reality, normalize peer differences, and avoid mechanically applying averages. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Choose valuation metrics that match the business and capital providers: enterprise value with pre-interest operating metrics, equity value with after-interest equity metrics. 1. Normalize peers for cycle, fiscal period, accounting, leases, SBC, capitalized investment, acquisitive growth, taxes, leverage, and non-operating assets before comparing headline multiples. 1. Use forward estimates only when the forecast definition and horizon are comparable; preserve dispersion and source date rather than treating consensus as one precise fact. 1. Explain multiple differences through growth, margin, ROIC, reinvestment, duration, balance-sheet risk, and earnings quality. 1. Use historical trading ranges carefully, adjusting for changed business mix, rates, capital structure, and growth economics. 1. Never average peer multiples mechanically. Derive a justified range and reconcile it with DCF/reverse-expectations evidence. ## Required evidence and model bridge - Primary-source set: normalized forecasts, capital structure, market data, peer definitions, scenario assumptions. Preserve exact document/version, date, period, and source location for every material factual input used in multiples and relative valuation. - For each key concept - metric selection, normalization, fiscal alignment, growth, margin, ROIC - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as multiples and relative valuation may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | EV/EBITDA | EV/EBITDA = enterprise value / normalized EBITDA | EV/EBITDA: Recalculate EV/EBITDA from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | EV/EBIT | EV/EBITDA = enterprise value / normalized EBITDA | EV/EBIT: Recalculate EV/EBIT from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | P/E | P/E = diluted equity value per share / normalized diluted EPS | P/E: Recalculate P/E from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | FCF yield | FCF yield = normalized free cash flow to equity / current equity value | FCF yield: Recalculate FCF yield from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | residual multiple after fundamentals | Peer valuation premium/discount remaining after adjusting for measured differences in growth, margins, ROIC, capital intensity, leverage, and risk. | residual multiple after fundamentals: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Relative valuation laboratory - Choose multiples whose numerator and denominator refer to the same capital providers. Enterprise multiples pair EV with pre-interest operating metrics; equity multiples pair equity value with after-interest metrics. - Normalize peer accounting before comparing multiples. Lease policy, SBC, capitalized R&D, acquisitive growth, pension accounting, fiscal dates, and cyclicality can make headline multiples non-comparable. - Build a regression or matrix against growth, margin, ROIC, leverage, and durability only when sample size and economics justify it. Peer average is not intrinsic value. ## Worked application > Case: same 20x EBITDA multiple means different value for capital-light and capital-heavy businesses. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For multiples and relative valuation, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through metric selection, normalization, fiscal alignment, growth, then identify which link is directly observed and which link remains an assumption. - Calculate EV/EBITDA, EV/EBIT, P/E, FCF yield from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for multiples and relative valuation: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: explain the premium or discount before calling a multiple cheap or expensive. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the multiples and relative valuation conclusion. ## Failure tests - FAIL if metric selection cannot be defined and reproduced from the source pack. - FAIL if a peer multiple is applied before normalizing denominator definition, growth, margins, capital intensity, leverage, cycle, and accounting differences. - FAIL if the multiples and relative valuation conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the multiples and relative valuation conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the multiples and relative valuation conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART X - VALUATION | MODULE 048 # Reverse DCF and Expectations Investing > Mission. Infer the growth, margin, capital intensity, and duration assumptions embedded in the stock price. ## Decision output Objective: Infer the growth, margin, capital intensity, and duration assumptions embedded in the stock price. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Start from current enterprise/equity value and a clean capital-structure bridge, then solve for the operating assumptions required to justify price. 1. Vary one major unknown at a time before solving combinations: revenue growth, steady-state margin, ROIC/reinvestment, competitive-advantage period, or terminal economics. 1. Translate implied financial outcomes into units, customers, market share, capacity, price, or other physical business reality. 1. Compare implied assumptions with historical ranges, peer economics, industry capacity, TAM/adoption constraints, and management plans. 1. Build an expectations matrix showing what combinations of growth and margin or ROIC/duration are embedded, without labeling high implied growth automatically expensive or cheap. 1. Frame the variant view as the specific expectation that evidence suggests is too optimistic or too pessimistic and define the data that would resolve it. ## Required evidence and model bridge - Primary-source set: normalized forecasts, capital structure, market data, peer definitions, scenario assumptions. Preserve exact document/version, date, period, and source location for every material factual input used in reverse dcf and expectations investing. - For each key concept - current EV, implied growth, implied margins, duration, reinvestment, market share - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as reverse dcf and expectations investing may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | implied CAGR | (Terminal or target value / current value)^(1/years) - 1 for the operating metric implied by the current market price or valuation. | implied CAGR: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | implied margin | implied margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | implied margin: Recalculate implied margin from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | expectation duration | Number of years the reverse-DCF requires above-normal growth, margins, or ROIC before fading to a mature-state assumption. | expectation duration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Reverse DCF laboratory - Start with current enterprise value and solve for one unknown at a time: revenue CAGR, steady-state margin, reinvestment/ROIC, fade period, or terminal economics. - Translate the solved assumption into physical reality. For example, a revenue path must imply customers, units, capacity, market share, or MW that can be compared with industry constraints. - Do not call the stock cheap because implied growth is high. Determine whether the implied combination of growth, margin, duration, and capital intensity is more or less demanding than evidence supports. ## Worked application > Case: the current price requires 18% growth for a decade and 30% terminal margins. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For reverse dcf and expectations investing, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through current EV, implied growth, implied margins, duration, then identify which link is directly observed and which link remains an assumption. - Calculate implied CAGR, implied margin, expectation duration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for reverse dcf and expectations investing: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: research the assumption that appears hardest for the market to achieve or easiest for the company to exceed. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the reverse dcf and expectations investing conclusion. ## Failure tests - FAIL if current EV cannot be defined and reproduced from the source pack. - FAIL if the market-implied operating path cannot be translated into measurable growth, margin, reinvestment, return, and duration assumptions that can be tested against evidence. - FAIL if the reverse dcf and expectations investing conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the reverse dcf and expectations investing conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the reverse dcf and expectations investing conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART X - VALUATION | MODULE 049 # Sum-of-the-Parts and Conglomerate Analysis > Mission. Value distinct businesses separately and account for central costs, taxes, minority interests, and capital structure. ## Decision output Objective: Value distinct businesses separately and account for central costs, taxes, minority interests, and capital structure. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Identify economically distinct segments/assets that deserve different forecast drivers, capital structures, risk, or valuation frameworks. 1. Build stand-alone operating forecasts and values for each segment using normalized peer/DCF/asset methods appropriate to the segment. 1. Allocate or separately value corporate costs, shared assets, taxes, pensions, minority interests, joint ventures, debt, cash, and non-operating investments. 1. Model stranded costs, separation expenses, tax leakage, dis-synergies, trapped capital, and financing changes for break-up/spin scenarios rather than applying an arbitrary conglomerate discount. 1. Avoid double counting intersegment revenue, shared assets, or centrally held claims and reconcile segment values to consolidated enterprise/equity value. 1. Reverse engineer the market price to determine which segment or central-cost assumptions drive any apparent discount. ## Required evidence and model bridge - Primary-source set: normalized forecasts, capital structure, market data, peer definitions, scenario assumptions. Preserve exact document/version, date, period, and source location for every material factual input used in sum-of-the-parts and conglomerate analysis. - For each key concept - segment economics, segment valuation methods, central costs, tax, debt, minorities - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as sum-of-the-parts and conglomerate analysis may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | segment EV | Segment-specific normalized operating metric multiplied by an appropriate valuation multiple, or segment DCF value, before corporate and balance-sheet adjustments. | segment EV: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | PV of corporate costs | Present value of recurring unallocated corporate cash costs not captured in segment values, tax-affected where appropriate. | PV of corporate costs: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | equity bridge | Enterprise value + non-operating assets - debt - leases/debt-like liabilities - minorities - pension deficits ± other claims = common equity value. | equity bridge: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | ## SOTP laboratory - Forecast and value each economically distinct segment using the framework appropriate to that segment. Allocate shared costs, debt, taxes, pensions, minority interests, and corporate assets explicitly. - Avoid applying a conglomerate discount as an unexplained haircut. Model stranded corporate costs, capital-allocation friction, tax leakage, and separation costs directly where possible. - Reconcile sum of segment enterprise values to consolidated equity value with a transparent bridge. ## Worked application > Case: software and hardware segments are obscured by a consolidated EBITDA multiple. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For sum-of-the-parts and conglomerate analysis, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through segment economics, segment valuation methods, central costs, tax, then identify which link is directly observed and which link remains an assumption. - Calculate segment EV, PV of corporate costs, equity bridge from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for sum-of-the-parts and conglomerate analysis: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: subtract real central and separation costs before claiming hidden value. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the sum-of-the-parts and conglomerate analysis conclusion. ## Failure tests - FAIL if segment economics cannot be defined and reproduced from the source pack. - FAIL if segment values omit shared costs, taxes, ownership, stranded costs, debt-like claims, or interdependencies needed to bridge to common equity. - FAIL if the sum-of-the-parts and conglomerate analysis conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the sum-of-the-parts and conglomerate analysis conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the sum-of-the-parts and conglomerate analysis conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART X - VALUATION | MODULE 050 # Scenario, Sensitivity, and Probability Weighting > Mission. Replace false precision with explicit bull, base, bear, stress, and thesis-break cases. ## Decision output Objective: Replace false precision with explicit bull, base, bear, stress, and thesis-break cases. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define scenarios as coherent operating states with linked demand, price, mix, margin, working capital, capex, financing, share count, and valuation assumptions. 1. Create a base case, plausible upside, plausible downside, severe stress, and explicit thesis-break case when material; distinguish temporary delay from permanent impairment. 1. Use one- and two-way sensitivities only for variables that can vary independently enough to create economically possible combinations. 1. Show scenario values before assigning probabilities. Probability weights are judgments and should include rationale, base rates where available, and an update rule. 1. Calculate expected value only as a decision aid; never let probability weighting hide catastrophic liquidity/dilution outcomes or a wide distribution. 1. Track which scenario drivers actually occurred and use post-mortems to recalibrate future ranges and probabilities. ## Required evidence and model bridge - Primary-source set: normalized forecasts, capital structure, market data, peer definitions, scenario assumptions. Preserve exact document/version, date, period, and source location for every material factual input used in scenario, sensitivity, and probability weighting. - For each key concept - coherent operating cases, correlated drivers, tail case, path dependence, probabilities, sensitivity matrices - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as scenario, sensitivity, and probability weighting may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | expected value | Sum of scenario value × scenario probability across mutually exclusive, collectively exhaustive cases. | expected value: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | downside asymmetry | Expected or stress-case downside magnitude relative to base/upside, including probability, liquidity, dilution, and permanent-impairment effects. | downside asymmetry: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | sensitivity slope | Change in equity value divided by change in the tested assumption over a defined local range; report nonlinear breakpoints where material. | sensitivity slope: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | ## Scenario design laboratory - Define scenarios by coherent operating states, not by changing only valuation multiples. A bear case should connect demand, price, mix, margin, working capital, capex, financing, and share count where those variables interact. - Probability weights are judgments, not facts. Show unweighted cases, weighted value, downside to stress, and the assumptions that would justify changing the weights. - Add two-way sensitivities only for variables that are economically independent enough to vary separately. Avoid grids that combine impossible states. ## Worked application > Case: revenue falls but the first bear case incorrectly leaves margins and working capital unchanged. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For scenario, sensitivity, and probability weighting, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through coherent operating cases, correlated drivers, tail case, path dependence, then identify which link is directly observed and which link remains an assumption. - Calculate expected value, downside asymmetry, sensitivity slope from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for scenario, sensitivity, and probability weighting: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: change linked operating variables together and keep liquidity failure from being averaged away. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the scenario, sensitivity, and probability weighting conclusion. ## Failure tests - FAIL if coherent operating cases cannot be defined and reproduced from the source pack. - FAIL if scenarios are arbitrary percentage haircuts rather than coherent operating states with causal driver interactions and observable evidence that changes probability. - FAIL if the scenario, sensitivity, and probability weighting conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the scenario, sensitivity, and probability weighting conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the scenario, sensitivity, and probability weighting conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XI - CATALYSTS AND EVENTS | MODULE 051 # Earnings Preview and Post-Earnings Procedure > Mission. Define expectations, key debate variables, scenario reactions, and model update rules before the print. ## Decision output Objective: Define expectations, key debate variables, scenario reactions, and model update rules before the print. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Freeze the pre-earnings model, market price, consensus context, management guidance, and your own expectations before the release. 1. Write the three to five variables that matter most, expected ranges, and what result would change the thesis versus create only short-term noise. 1. After release, reconcile the filed/reported statements and KPI definitions before changing the forecast; do not react only to headline EPS or revenue beats. 1. Bridge revenue, margin, working capital, capex, cash, debt, shares, taxes, and guidance to the pre-quarter model and explain each material variance. 1. Compare prepared remarks and Q&A with prior wording for confidence, timing, KPI emphasis, and new qualifiers; management narrative remains a claim until reconciled. 1. Log forecast error, model changes, valuation change, and what the next quarter must show. ## Required evidence and model bridge - Primary-source set: event filings, guidance, consensus context, transaction documents, covenant and liquidity data. Preserve exact document/version, date, period, and source location for every material factual input used in earnings preview and post-earnings procedure. - For each key concept - consensus, guidance, key debate variables, precommitment, variance bridge, transcript changes - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as earnings preview and post-earnings procedure may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | expectation gap | Analyst forecast for a key driver minus the value or consensus-implied expectation, expressed in operating units and valuation impact. | expectation gap: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | variance attribution | Actual minus prior forecast decomposed by driver, price, volume, mix, cost, timing, accounting, and assumption error; contributions sum to total variance. | variance attribution: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | | estimate revision impact | Change in valuation or expected return caused solely by revised operating estimates, holding valuation framework and market inputs constant. | estimate revision impact: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | ## Worked application > Case: a revenue beat comes from low-margin mix while the thesis depends on margin recovery. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For earnings preview and post-earnings procedure, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through consensus, guidance, key debate variables, precommitment, then identify which link is directly observed and which link remains an assumption. - Calculate expectation gap, variance attribution, estimate revision impact from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for earnings preview and post-earnings procedure: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: define what matters before the print so headline beats cannot rewrite the thesis. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the earnings preview and post-earnings procedure conclusion. ## Failure tests - FAIL if consensus cannot be defined and reproduced from the source pack. - FAIL if the analyst has not prewritten decision-relevant expectations and interpretations before the release, allowing the post-event narrative to fit any outcome. - FAIL if the earnings preview and post-earnings procedure conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the earnings preview and post-earnings procedure conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the earnings preview and post-earnings procedure conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XI - CATALYSTS AND EVENTS | MODULE 052 # Guidance Analysis and Estimate Revisions > Mission. Translate management guidance into implied quarterly paths, margins, cash flow, and consensus risk. ## Decision output Objective: Translate management guidance into implied quarterly paths, margins, cash flow, and consensus risk. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Parse guidance into metric, GAAP/non-GAAP definition, range, period, currency/FX assumption, acquisition scope, share count, tax rate, and any explicit operating assumptions. 1. Translate annual guidance into an implied quarterly/half-year path using actual results, seasonality, backlog/orders, margin cadence, working capital, capex, and known events. 1. Separate management midpoint mathematics from your probability distribution; widening ranges, one-sided wording, or changed exclusions can alter risk without changing the midpoint. 1. Build an earnings/FCF bridge for each revised guidance component and calculate the implied exit run rate. 1. Compare guidance accuracy and revision behavior across several years to understand conservatism, visibility, and definition changes. 1. Update estimates only for changed evidence, recording prior/new assumption and valuation sensitivity rather than mechanically matching management guidance or consensus. ## Required evidence and model bridge - Primary-source set: event filings, guidance, consensus context, transaction documents, covenant and liquidity data. Preserve exact document/version, date, period, and source location for every material factual input used in guidance analysis and estimate revisions. - For each key concept - guidance definitions, implied remaining quarters, historical bias, range width, FX/tax/buyback effects, consensus - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as guidance analysis and estimate revisions may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | implied remaining period | Full-period target or guidance minus actual-to-date result, adjusted for seasonality and known discrete items. | implied remaining period: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | guidance bias | Average signed actual-minus-guidance error over multiple periods, supplemented by beat/miss frequency and guidance-width changes. | guidance bias: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | revision breadth | % of tracked analyst estimates or internal forecast lines revised in the same direction over the selected window. | revision breadth: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | ## Worked application > Case: EPS guide rises because tax and share count improve while operations do not. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For guidance analysis and estimate revisions, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through guidance definitions, implied remaining quarters, historical bias, range width, then identify which link is directly observed and which link remains an assumption. - Calculate implied remaining period, guidance bias, revision breadth from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for guidance analysis and estimate revisions: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: decompose every guidance revision into operating and financial components. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the guidance analysis and estimate revisions conclusion. ## Failure tests - FAIL if guidance definitions cannot be defined and reproduced from the source pack. - FAIL if guidance is copied into the model without solving the implied quarterly path, definitions, margins, cash flow, and assumptions required to reach the range. - FAIL if the guidance analysis and estimate revisions conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the guidance analysis and estimate revisions conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the guidance analysis and estimate revisions conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XI - CATALYSTS AND EVENTS | MODULE 053 # M&A and Strategic Transactions > Mission. Analyze purchase price, synergies, financing, accretion, integration, antitrust, and value transfer. ## Decision output Objective: Analyze purchase price, synergies, financing, accretion, integration, antitrust, and value transfer. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Calculate total consideration including equity value, assumed/refinanced debt, options, contingent consideration, transaction fees, retention, integration, and required refinancing. 1. Reconstruct financing mix, pro forma leverage, interest expense, share issuance/dilution, cash use, and covenant/liquidity effects. 1. Model purchase accounting, intangible amortization, step-ups, deferred taxes, goodwill, divestitures, and reported-versus-cash earnings effects. 1. Separate cost synergies, revenue synergies, dis-synergies, integration cost, timing, tax, and capital requirements; assign evidence and realization probability to each. 1. Evaluate strategic fit and return on invested capital/NPV, not merely EPS accretion, because leverage and accounting can create accretion without value creation. 1. Map antitrust/regulatory, financing, shareholder vote, closing-condition, break-fee, and integration risks with dates and scenario values. ## Required evidence and model bridge - Primary-source set: event filings, guidance, consensus context, transaction documents, covenant and liquidity data. Preserve exact document/version, date, period, and source location for every material factual input used in m&a and strategic transactions. - For each key concept - consideration, financing, normalized target, synergies, integration, purchase accounting - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as m&a and strategic transactions may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | purchase multiple | Enterprise purchase price divided by target LTM/NTM revenue, EBITDA, EBIT, or FCF, with debt, cash, earnouts, and leases consistently treated. | purchase multiple: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | post-deal ROIC-WACC spread | ROIC = NOPAT / average invested capital | post-deal ROIC-WACC spread: Recalculate post-deal ROIC-WACC spread from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | | accretion decomposition | Change in pro forma EPS or FCF/share decomposed into operating synergies, financing cost, purchase accounting, tax, share issuance, and lost interest/investment income. | accretion decomposition: Tie opening/closing balances or total change to primary-source financials; verify components sum exactly with no overlap, omission, or sign error. | ## Worked application > Case: EPS accretion is driven by cheap debt even though deal ROIC is below WACC. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For m&a and strategic transactions, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through consideration, financing, normalized target, synergies, then identify which link is directly observed and which link remains an assumption. - Calculate purchase multiple, post-deal ROIC-WACC spread, accretion decomposition from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for m&a and strategic transactions: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: score deals by value creation and post-deal cash returns, not EPS accretion. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the m&a and strategic transactions conclusion. ## Failure tests - FAIL if consideration cannot be defined and reproduced from the source pack. - FAIL if accretion is equated with value creation or if purchase price, standalone value, synergies, financing, integration, tax, and execution risk are not separated. - FAIL if the m&a and strategic transactions conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the m&a and strategic transactions conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the m&a and strategic transactions conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XI - CATALYSTS AND EVENTS | MODULE 054 # Activism, Spin-Offs, and Restructurings > Mission. Evaluate separation economics, stranded costs, incentive resets, and capital structure changes. ## Decision output Objective: Evaluate separation economics, stranded costs, incentive resets, and capital structure changes. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Reconstruct the activist/restructuring thesis into discrete value levers: portfolio change, cost structure, capital allocation, governance, separation, financing, or strategic sale. 1. For a separation, build stand-alone segment forecasts, corporate/stranded costs, dis-synergies, duplicated systems, tax leakage, debt allocation, pension/other liabilities, one-time costs, and required working capital. 1. Model how new management incentives and capital structures change behavior and financial flexibility at each entity. 1. Distinguish accounting restructuring charges from recurring economic cost and test whether savings persist after growth normalizes. 1. Create event path probabilities for proposal, agreement, regulatory/shareholder approval, execution, delay, and failure rather than valuing only the announced end state. 1. Compare the post-transaction sum of values with the status quo after all costs and capital-structure changes. ## Required evidence and model bridge - Primary-source set: event filings, guidance, consensus context, transaction documents, covenant and liquidity data. Preserve exact document/version, date, period, and source location for every material factual input used in activism, spin-offs, and restructurings. - For each key concept - segment standalone economics, stranded costs, tax, separation costs, leverage, incentives - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as activism, spin-offs, and restructurings may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | standalone FCF | Free cash flow of the separated business before separation-specific stranded costs, new public-company costs, financing changes, and one-time transaction costs. | standalone FCF: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | net separation value | SOTP value of separated entities less stranded costs, separation costs, tax leakage, incremental debt, and other value leakage versus status quo. | net separation value: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | leverage allocation | Debt assigned to each separated entity divided by its normalized EBITDA/FCF and compared with required liquidity, covenants, and target credit capacity. | leverage allocation: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Worked application > Case: spin valuation assumes all corporate costs disappear. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For activism, spin-offs, and restructurings, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through segment standalone economics, stranded costs, tax, separation costs, then identify which link is directly observed and which link remains an assumption. - Calculate standalone FCF, net separation value, leverage allocation from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for activism, spin-offs, and restructurings: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: model what actually disappears, transfers, or must be recreated. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the activism, spin-offs, and restructurings conclusion. ## Failure tests - FAIL if segment standalone economics cannot be defined and reproduced from the source pack. - FAIL if proposed value creation ignores stranded costs, separation taxes, transition needs, debt allocation, execution cost, and changes in management incentives. - FAIL if the activism, spin-offs, and restructurings conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the activism, spin-offs, and restructurings conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the activism, spin-offs, and restructurings conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XI - CATALYSTS AND EVENTS | MODULE 055 # Bankruptcy, Distress, and Liquidity Events > Mission. Analyze runway, covenants, collateral, priority, recovery, dilution, and restructuring pathways. ## Decision output Objective: Analyze runway, covenants, collateral, priority, recovery, dilution, and restructuring pathways. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build a near-term cash runway at monthly or 13-week granularity when liquidity is critical, incorporating restricted cash, revolver availability, seasonal working capital, committed capex, interest, maturities, and restructuring cash. 1. Read debt documents for collateral, priority, guarantees, covenants, borrowing-base limits, cross-defaults, baskets, and amendment/waiver mechanics. 1. Stress operating cash flow and working capital before valuation. Identify the date and event that exhausts liquidity under each scenario. 1. Map management financing options by timing and realism: capex cuts, asset sales, dividend suspension, revolver, secured debt, exchange, equity, rescue capital, covenant amendment, or restructuring. 1. Estimate enterprise recovery under multiple normalized operating outcomes and allocate value through legal priority to each debt/preferred/equity claim. 1. For common equity, model dilution and zero-recovery pathways explicitly rather than treating distress as a simple lower multiple. ## Required evidence and model bridge - Primary-source set: event filings, guidance, consensus context, transaction documents, covenant and liquidity data. Preserve exact document/version, date, period, and source location for every material factual input used in bankruptcy, distress, and liquidity events. - For each key concept - security-by-security debt, liens, maturities, covenants, unrestricted cash, revolver - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as bankruptcy, distress, and liquidity events may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | liquidity runway | Unrestricted cash + committed undrawn facilities - required minimum cash, divided by forecast monthly/quarterly cash burn after near-term maturities. | liquidity runway: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | fixed-charge coverage | Cash earnings available for fixed charges divided by cash interest, required lease/rent payments, preferred dividends, and other contractual fixed charges included in the definition. | fixed-charge coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | recovery by claim | Value distributable to each capital-structure class after enterprise-value scenarios and administrative/priority claims, divided by that class's allowed claim. | recovery by claim: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | ## Distress and recovery laboratory - Build a 13-week cash view when liquidity is acute, then a monthly/quarterly runway. Include restricted cash, revolver availability, borrowing-base limits, covenant tests, letters of credit, working-capital seasonality, and restructuring cash costs. - Map legal priority: secured debt, unsecured debt, leases/other claims, preferred equity, and common equity. Estimate enterprise recovery under multiple operating values before allocating to each claim. - For equity, distinguish temporary liquidity pressure from a capital-structure problem that can cause dilution, exchange offers, rescue financing, or zero recovery. ## Worked application > Case: cash appears sufficient but a springing covenant activates earlier than cash runs out. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For bankruptcy, distress, and liquidity events, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through security-by-security debt, liens, maturities, covenants, then identify which link is directly observed and which link remains an assumption. - Calculate liquidity runway, fixed-charge coverage, recovery by claim from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for bankruptcy, distress, and liquidity events: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: model contractual triggers and common-equity survival separately from enterprise viability. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the bankruptcy, distress, and liquidity events conclusion. ## Failure tests - FAIL if security-by-security debt cannot be defined and reproduced from the source pack. - FAIL if the analysis applies a lower valuation multiple before modeling cash runway, covenants, maturities, collateral, claim priority, financing access, and recovery scenarios. - FAIL if the bankruptcy, distress, and liquidity events conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the bankruptcy, distress, and liquidity events conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the bankruptcy, distress, and liquidity events conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK | MODULE 056 # Expert Calls and Channel Checks > Mission. Design compliant interviews that test hypotheses without soliciting material nonpublic information. ## Decision output Objective: Design compliant interviews that test hypotheses without soliciting material nonpublic information. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Write the hypothesis and permitted factual questions before the call. Define what information would change the view and what topics are prohibited or require compliance guidance. 1. Use compliant participants and procedures. Never solicit or use material nonpublic information; stop and escalate if a participant begins providing information that appears restricted. 1. Select respondents across roles, geographies, customer sizes, and competitive positions to reduce sample bias; document recruitment and incentives. 1. Use consistent core questions plus open-ended follow-ups, recording exact dates, respondent context, confidence, and contradictions without overstating representativeness. 1. Triangulate qualitative claims with filings, pricing, product data, competitor evidence, or other independent observations. 1. Convert call evidence into a bounded model input only after defining the population and expected relationship to reported outcomes. ## Required evidence and model bridge - Primary-source set: documented datasets, sampling frame, legal/compliance approval, backtests, source provenance. Preserve exact document/version, date, period, and source location for every material factual input used in expert calls and channel checks. - For each key concept - hypothesis guide, participant screening, MNPI stop rule, sampling, role/incentive, independent corroboration - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as expert calls and channel checks may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | independent source count | Number of genuinely independent evidence sources supporting a material conclusion after removing common-origin/circular citations. | independent source count: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | signal consistency | Share of independent indicators pointing in the same economic direction after standardizing sign, horizon, and expected lead/lag. | signal consistency: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | compliance exception rate | Expert-call/channel-check interactions with documented compliance exceptions divided by total interactions, with severity separately tracked. | compliance exception rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: a distributor begins to discuss an undisclosed quarter-end customer order. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For expert calls and channel checks, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through hypothesis guide, participant screening, MNPI stop rule, sampling, then identify which link is directly observed and which link remains an assumption. - Calculate independent source count, signal consistency, compliance exception rate from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for expert calls and channel checks: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: predefine legal/compliance stop rules and use calls to test, not confirm, the thesis. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the expert calls and channel checks conclusion. ## Failure tests - FAIL if hypothesis guide cannot be defined and reproduced from the source pack. - FAIL if expert evidence is collected without a predefined hypothesis, participant screening, compliance boundary, source-role context, and independent corroboration. - FAIL if the expert calls and channel checks conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the expert calls and channel checks conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the expert calls and channel checks conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK | MODULE 057 # Customer and Supplier Diligence > Mission. Triangulate demand, pricing, product quality, switching behavior, inventory, and competitive share. ## Decision output Objective: Triangulate demand, pricing, product quality, switching behavior, inventory, and competitive share. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Map the customer/supplier population and choose a sample that reflects concentration, geography, size, channel, tenure, and product exposure relevant to the thesis. 1. For customers, test purchase trigger, ROI, satisfaction, alternatives, switching costs, renewal/expansion, pricing, budget priority, and implementation constraints. 1. For suppliers, test order cadence, lead times, capacity, allocation, cancellations, inventory, pricing, payment terms, customer concentration, and visibility into true end demand. 1. Separate anecdotal experience from measurable frequency or magnitude; one large customer can matter financially but still not represent the market. 1. Cross-check claims with company/peer filings, channel inventory, product availability, pricing, and reported working-capital movements. 1. Document what the evidence can and cannot support and avoid extrapolating beyond the sampled population. ## Required evidence and model bridge - Primary-source set: documented datasets, sampling frame, legal/compliance approval, backtests, source provenance. Preserve exact document/version, date, period, and source location for every material factual input used in customer and supplier diligence. - For each key concept - buying process, customer ROI, switching, dual sourcing, discounting, supplier lead times - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as customer and supplier diligence may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | customer dependency | Revenue, gross profit, backlog, or receivables attributable to top customers divided by the relevant total, with renewal/contract timing overlaid. | customer dependency: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | supplier dependency | Spend, input volume, or critical component capacity sourced from top suppliers divided by total requirement, adjusted for qualified alternatives and switching lead time. | supplier dependency: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | switching-cost evidence | Observed retention, migration time/cost, contract penalties, implementation burden, and win/loss behavior synthesized into a documented switching-cost score or quantified range. | switching-cost evidence: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | ## Worked application > Case: customers like the product but increasingly dual-source it. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For customer and supplier diligence, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through buying process, customer ROI, switching, dual sourcing, then identify which link is directly observed and which link remains an assumption. - Calculate customer dependency, supplier dependency, switching-cost evidence from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for customer and supplier diligence: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: behavior and contract economics matter more than sentiment. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the customer and supplier diligence conclusion. ## Failure tests - FAIL if buying process cannot be defined and reproduced from the source pack. - FAIL if a small or biased customer/supplier sample is generalized to the whole market without exposure weighting, segmentation, and competing explanations. - FAIL if the customer and supplier diligence conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the customer and supplier diligence conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the customer and supplier diligence conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK | MODULE 058 # Web, App, Hiring, and Product Data > Mission. Use digital signals as noisy indicators that require baselines, controls, and definition discipline. ## Decision output Objective: Use digital signals as noisy indicators that require baselines, controls, and definition discipline. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define the observable signal and causal hypothesis before collecting clicks, downloads, traffic, rankings, jobs, reviews, product releases, or usage proxies. 1. Preserve raw series, geography/device/channel coverage, sampling rules, seasonality, platform methodology, revisions, and historical breaks. 1. Normalize for marketing campaigns, bots, platform algorithm changes, app-store policy, hiring reposts, product bundles, and other non-economic artifacts. 1. Link the signal to a reported KPI with an expected lead/lag and test historical fit across multiple periods, not only the latest quarter. 1. Use holdout periods or peer controls where possible to reduce overfitting and spurious correlation. 1. Downgrade or discard the signal when definitions/coverage change faster than the economic relationship can be validated. ## Required evidence and model bridge - Primary-source set: documented datasets, sampling frame, legal/compliance approval, backtests, source provenance. Preserve exact document/version, date, period, and source location for every material factual input used in web, app, hiring, and product data. - For each key concept - causal bridge, provider methodology, coverage, seasonality, revisions, platform breaks - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as web, app, hiring, and product data may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | out-of-sample correlation/error | Correlation or forecast error measured only on holdout periods not used to tune the alternative-data signal; report MAE/RMSE and stability by regime. | out-of-sample correlation/error: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | coverage ratio | Observed entities, SKUs, geographies, or transactions represented in the alternative dataset divided by the economically relevant universe. | coverage ratio: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | revision rate | Number of historical data points materially revised after initial publication divided by total observations, with average revision magnitude. | revision rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: job postings jump because replacements and geography mix change. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For web, app, hiring, and product data, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through causal bridge, provider methodology, coverage, seasonality, then identify which link is directly observed and which link remains an assumption. - Calculate out-of-sample correlation/error, coverage ratio, revision rate from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for web, app, hiring, and product data: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: preserve vintages and platform-change metadata so backtests use information available at the time. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the web, app, hiring, and product data conclusion. ## Failure tests - FAIL if causal bridge cannot be defined and reproduced from the source pack. - FAIL if a digital signal changes the model before methodology, coverage, seasonality, confounders, revisions, and historical relationship to the target KPI are validated. - FAIL if the web, app, hiring, and product data conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the web, app, hiring, and product data conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the web, app, hiring, and product data conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK | MODULE 059 # Pricing and Inventory Scraping > Mission. Track availability, discounting, lead times, SKU breadth, and channel inventory without overfitting. ## Decision output Objective: Track availability, discounting, lead times, SKU breadth, and channel inventory without overfitting. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Define permitted data sources, terms/robots/compliance constraints, SKU/location universe, frequency, timestamp, and raw-page retention before scraping. 1. Create stable product identifiers and handle variants, bundles, coupons, loyalty pricing, shipping, taxes, geography, and out-of-stock states consistently. 1. Distinguish list price, transacted/advertised price, discount depth, promotion frequency, availability, lead time, and inventory proxy; do not collapse them into one price series. 1. Monitor site layout/API changes and build validation checks for missing pages, duplicated SKUs, unit changes, bot blocking, and false stockouts. 1. Aggregate by economically meaningful category and weight, then compare with company-reported pricing, volume, mix, inventory, and channel evidence. 1. Backtest the scraped signal before using it to forecast revenue or margin and preserve the raw observations for audit. ## Required evidence and model bridge - Primary-source set: documented datasets, sampling frame, legal/compliance approval, backtests, source provenance. Preserve exact document/version, date, period, and source location for every material factual input used in pricing and inventory scraping. - For each key concept - permission, SKU identifiers, timestamps, seller/channel, matched baskets, stock status - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as pricing and inventory scraping may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | in-stock rate | Observed SKU-location checks showing available inventory divided by valid SKU-location observations. | in-stock rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | matched-SKU price index | Current price of the same SKU/location basket divided by base-period price of that identical basket, weighted by a fixed basket or documented economic weights. | matched-SKU price index: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | promotion intensity | Discounted/promoted observations or promotion dollars divided by total observations or gross sales, with depth and duration separately tracked. | promotion intensity: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: average price rises because low-priced SKUs go out of stock. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For pricing and inventory scraping, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through permission, SKU identifiers, timestamps, seller/channel, then identify which link is directly observed and which link remains an assumption. - Calculate in-stock rate, matched-SKU price index, promotion intensity from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for pricing and inventory scraping: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: separate price from mix and monitor scraping failure as a data-quality event. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the pricing and inventory scraping conclusion. ## Failure tests - FAIL if permission cannot be defined and reproduced from the source pack. - FAIL if scraped price or availability is interpreted without SKU matching, channel coverage, promotions, product mix, data-quality controls, and a baseline. - FAIL if the pricing and inventory scraping conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the pricing and inventory scraping conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the pricing and inventory scraping conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XII - ALTERNATIVE DATA AND CHANNEL WORK | MODULE 060 # Alternative Data Validation > Mission. Backtest data against reported outcomes and measure false positives, revisions, survivorship, and coverage bias. ## Decision output Objective: Backtest data against reported outcomes and measure false positives, revisions, survivorship, and coverage bias. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Write the data-generating process: population, sample, coverage, frequency, revision policy, missingness, survivorship, and how the signal should causally reach a company KPI. 1. Create historical features using only information that would have been available at each date to avoid look-ahead and revision bias. 1. Backtest against reported outcomes over enough periods and regimes to estimate correlation, predictive error, false positives, false negatives, and lead/lag stability. 1. Use holdouts, peer/control groups, and simple baselines before complex models; a signal must beat a naive forecast after transaction/research costs to deserve weight. 1. Track vendor methodology/coverage changes and rerun historical relationships when the dataset changes. 1. Set a retirement rule for signals whose predictive relationship, coverage, or economic mechanism deteriorates. ## Required evidence and model bridge - Primary-source set: documented datasets, sampling frame, legal/compliance approval, backtests, source provenance. Preserve exact document/version, date, period, and source location for every material factual input used in alternative data validation. - For each key concept - target KPI, frozen vintages, baseline, train/validation/holdout, robustness, false positives - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as alternative data validation may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | out-of-sample error | Forecast minus realized outcome on holdout periods, summarized with MAE, MAPE/RMSE as appropriate and compared with a simple benchmark. | out-of-sample error: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | incremental explanatory power | Increase in out-of-sample R-squared, reduction in forecast error, or decision accuracy after adding the alternative signal to the baseline model. | incremental explanatory power: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | false-positive rate | Negative/normal realized outcomes incorrectly flagged as positive/adverse signals divided by all truly negative/normal outcomes. | false-positive rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: revised vendor history looks predictive but real-time vintages were weak. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For alternative data validation, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through target KPI, frozen vintages, baseline, train/validation/holdout, then identify which link is directly observed and which link remains an assumption. - Calculate out-of-sample error, incremental explanatory power, false-positive rate from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for alternative data validation: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: approve a signal for production only after vintage-aware out-of-sample validation. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the alternative data validation conclusion. ## Failure tests - FAIL if target KPI cannot be defined and reproduced from the source pack. - FAIL if a signal is backtested using revised data, look-ahead information, survivor-only samples, or in-sample correlation without false-positive and out-of-sample analysis. - FAIL if the alternative data validation conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the alternative data validation conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the alternative data validation conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING | MODULE 061 # Risk Register and Thesis-Break Conditions > Mission. Maintain explicit operational, financial, regulatory, competitive, valuation, and governance risks. ## Decision output Objective: Maintain explicit operational, financial, regulatory, competitive, valuation, and governance risks. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. For each risk, record failure mode, causal transmission mechanism, leading indicator, lag to financial statements, financial exposure, time horizon, mitigants, source, owner, and review date. 1. Separate operational, competitive, regulatory, accounting, governance, balance-sheet, liquidity, financing, valuation, and event risks so one label does not hide multiple mechanisms. 1. Quantify ranges for impact and identify interactions; risks often compound through demand, margin, working capital, and financing at the same time. 1. Define thesis-break conditions before adverse evidence arrives and distinguish permanent impairment, temporary volatility, and thesis delay. 1. Link every thesis break to a monitoring source and alert threshold that can be observed in time to act analytically. 1. After material events, close, downgrade, escalate, or rewrite each affected risk rather than carrying stale boilerplate forward. ## Required evidence and model bridge - Primary-source set: risk register, stress model, correlation/liquidity inputs, decision journal, forecast-error history. Preserve exact document/version, date, period, and source location for every material factual input used in risk register and thesis-break conditions. - For each key concept - failure mode, transmission mechanism, leading indicator, probability, severity, evidence source - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as risk register and thesis-break conditions may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | risk exposure | Dollar, earnings, cash-flow, or valuation amount exposed to a specified risk before mitigation, multiplied by sensitivity where appropriate. | risk exposure: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | detection lead time | Time between the first reliable leading indicator crossing its threshold and the reported financial or thesis event it is intended to anticipate. | detection lead time: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | thesis-break coverage | % of material thesis assumptions with a predefined measurable break condition, monitoring source, owner, and review cadence. | thesis-break coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: replace generic "competition risk" with measurable retention and price conditions. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For risk register and thesis-break conditions, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through failure mode, transmission mechanism, leading indicator, probability, then identify which link is directly observed and which link remains an assumption. - Calculate risk exposure, detection lead time, thesis-break coverage from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for risk register and thesis-break conditions: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: risk entries must be causal, monitorable, and modelable. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the risk register and thesis-break conditions conclusion. ## Failure tests - FAIL if failure mode cannot be defined and reproduced from the source pack. - FAIL if a risk has no transmission mechanism, leading indicator, financial exposure, monitoring source, and precommitted thesis-break threshold. - FAIL if the risk register and thesis-break conditions conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the risk register and thesis-break conditions conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the risk register and thesis-break conditions conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING | MODULE 062 # Balance Sheet and Liquidity Stress Testing > Mission. Stress cash burn, maturities, covenants, refinancing rates, collateral, and working-capital shocks. ## Decision output Objective: Stress cash burn, maturities, covenants, refinancing rates, collateral, and working-capital shocks. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Start with unrestricted cash and committed availability after borrowing-base/covenant restrictions, then build scheduled debt/lease/other fixed payments by date. 1. Stress revenue, margin, working-capital reversal, capex commitments, restructuring, collateral, interest rates, and refinancing together over the actual risk window. 1. Calculate minimum cash, covenant headroom, fixed-charge/interest coverage, maturity wall, and time to liquidity shortfall under each scenario. 1. Model management responses with timing and feasibility, including capex deferral, asset sales, dividend/buyback suspension, revolver draw, secured financing, equity, waiver, or restructuring. 1. Do not assume refinancing simply because debt is termed out today; specify market access, collateral, pricing, and leverage at the future refinance date. 1. Translate the stress into dilution/recovery and valuation so balance-sheet risk is not represented by an arbitrary multiple haircut. ## Required evidence and model bridge - Primary-source set: risk register, stress model, correlation/liquidity inputs, decision journal, forecast-error history. Preserve exact document/version, date, period, and source location for every material factual input used in balance sheet and liquidity stress testing. - For each key concept - all fixed claims, debt maturities, covenant definitions, restricted cash, revolver, working capital - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as balance sheet and liquidity stress testing may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | minimum liquidity | Lowest cash plus committed available liquidity across the stress horizon after required operating cash, maturities, collateral calls, and covenant-driven restrictions. | minimum liquidity: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | maturity wall | Debt principal and mandatory financing obligations due by period, shown in dollars and as a % of liquidity/normalized FCF. | maturity wall: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | covenant headroom | Distance between forecast covenant metric and covenant limit, expressed in absolute units and % headroom under base and stress cases. | covenant headroom: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Liquidity stress laboratory - Stress operating cash generation, working-capital reversal, capex commitments, debt maturities, interest rates, collateral/borrowing base, covenant headroom, and minimum cash simultaneously. - Show monthly or quarterly headroom until the risk window passes. A company with positive annual FCF can still fail if cash needs occur before receipts or refinancing. - Identify management actions by timing and realism: capex deferral, asset sales, dividend suspension, revolver draw, equity issuance, covenant amendment, restructuring. ## Worked application > Case: reported cash looks safe but minimum operating cash and a maturity create a refinancing gap. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For balance sheet and liquidity stress testing, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through all fixed claims, debt maturities, covenant definitions, restricted cash, then identify which link is directly observed and which link remains an assumption. - Calculate minimum liquidity, maturity wall, covenant headroom from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for balance sheet and liquidity stress testing: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use debt-document definitions and find the first date action becomes necessary. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the balance sheet and liquidity stress testing conclusion. ## Failure tests - FAIL if all fixed claims cannot be defined and reproduced from the source pack. - FAIL if stress testing lowers earnings or valuation without modeling cash timing, working capital, mandatory uses, financing availability, covenants, and minimum operating liquidity. - FAIL if the balance sheet and liquidity stress testing conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the balance sheet and liquidity stress testing conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the balance sheet and liquidity stress testing conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING | MODULE 063 # Position Sizing Inputs for Research Teams > Mission. Translate conviction, downside, liquidity, catalyst path, and correlation into research inputs without confusing analysis with mandate. ## Decision output Objective: Translate conviction, downside, liquidity, catalyst path, and correlation into research inputs without confusing analysis with mandate. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Provide research inputs rather than an unauthorized portfolio decision: expected operating range, valuation distribution, permanent-loss case, catalyst path, liquidity, and confidence in key assumptions. 1. Separate thesis confidence from upside magnitude. A high-upside security can have low evidence quality, binary risk, or severe dilution/illiquidity. 1. Describe downside by mechanism and recovery, including balance-sheet path, not only historical volatility or a percentage stop. 1. Estimate trading liquidity, event gaps, borrow/short mechanics where relevant, and correlation/common-factor exposure that can make several seemingly different positions fail together. 1. State which risks are diversifiable versus thesis-specific and which evidence would warrant changing the research confidence. 1. Provide scenario inputs consistently so the portfolio owner can apply mandate-specific risk budgets independently. ## Required evidence and model bridge - Primary-source set: risk register, stress model, correlation/liquidity inputs, decision journal, forecast-error history. Preserve exact document/version, date, period, and source location for every material factual input used in position sizing inputs for research teams. - For each key concept - return distribution, permanent impairment, liquidity, correlation, factor exposure, catalyst timing - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as position sizing inputs for research teams may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | expected value | Sum of scenario value × scenario probability across mutually exclusive, collectively exhaustive cases. | expected value: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | severe downside | Equity value or loss in the predefined severe but plausible operating/financing scenario, including dilution, refinancing, and claim-priority effects. | severe downside: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | liquidity days | Available unrestricted liquidity divided by average daily cash operating outflow under the relevant stress case. | liquidity days: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Worked application > Case: two ideas have equal upside but radically different downside and liquidity. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For position sizing inputs for research teams, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through return distribution, permanent impairment, liquidity, correlation, then identify which link is directly observed and which link remains an assumption. - Calculate expected value, severe downside, liquidity days from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for position sizing inputs for research teams: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: research supplies transparent distributions and constraints instead of hiding everything in a conviction score. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the position sizing inputs for research teams conclusion. ## Failure tests - FAIL if return distribution cannot be defined and reproduced from the source pack. - FAIL if research communicates conviction without explicit downside distribution, evidence quality, liquidity, duration, correlation drivers, and thesis-break path. - FAIL if the position sizing inputs for research teams conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the position sizing inputs for research teams conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the position sizing inputs for research teams conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING | MODULE 064 # Decision Journaling and Pre-Mortems > Mission. Record the decision, evidence, disconfirming facts, expected path, and what would change the view. ## Decision output Objective: Record the decision, evidence, disconfirming facts, expected path, and what would change the view. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Before the decision outcome is known, record thesis, valuation range, key assumptions, confidence, expected path, catalysts, risks, strongest contrary evidence, and what would change the view. 1. Run a pre-mortem: assume the thesis failed badly and list plausible reasons across demand, competition, execution, accounting, regulation, balance sheet, valuation, and process. 1. Convert the most plausible pre-mortem failures into leading indicators and thesis-break conditions with dates/sources. 1. Record alternatives considered and why they were rejected so later reviewers can distinguish a good process from hindsight reconstruction. 1. Freeze the journal version at each major decision or material thesis update. Do not rewrite prior reasoning after the outcome. 1. Use the journal in post-mortems to compare what was expected with what actually occurred and where uncertainty was misjudged. ## Required evidence and model bridge - Primary-source set: risk register, stress model, correlation/liquidity inputs, decision journal, forecast-error history. Preserve exact document/version, date, period, and source location for every material factual input used in decision journaling and pre-mortems. - For each key concept - dated decision, market expectations, assumptions, probability ranges, expected evidence, pre-mortem - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as decision journaling and pre-mortems may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | precommitment coverage | % of identified material decisions with a dated pre-mortem, base-rate expectation, thesis-break rule, and action trigger documented before the outcome. | precommitment coverage: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | calibration | For forecasts assigned probability p, realized frequency of the event in that probability bucket; compare realized frequency with stated probability. | calibration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | thesis drift count | Number of material thesis assumptions changed after adverse evidence without a contemporaneous documented rationale or explicit thesis reset. | thesis drift count: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | ## Worked application > Case: the stock rises even though a precommitted retention threshold breaks. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For decision journaling and pre-mortems, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through dated decision, market expectations, assumptions, probability ranges, then identify which link is directly observed and which link remains an assumption. - Calculate precommitment coverage, calibration, thesis drift count from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for decision journaling and pre-mortems: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: judge the new evidence against the old standard before price action changes the narrative. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the decision journaling and pre-mortems conclusion. ## Failure tests - FAIL if dated decision cannot be defined and reproduced from the source pack. - FAIL if the prior decision state cannot be reconstructed or if a pre-mortem does not identify distinct failure mechanisms and early warning evidence. - FAIL if the decision journaling and pre-mortems conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the decision journaling and pre-mortems conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the decision journaling and pre-mortems conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIII - RISK, PORTFOLIO CONTEXT, AND DECISION MAKING | MODULE 065 # Post-Mortems and Error Taxonomy > Mission. Separate thesis errors, sizing errors, timing errors, data errors, process errors, and luck. ## Decision output Objective: Separate thesis errors, sizing errors, timing errors, data errors, process errors, and luck. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Compare actual results with the dated forecast and decision journal, separating business outcome, stock outcome, and information available at the time. 1. Attribute error to data/source, accounting/definition, model mechanics, forecast assumption, industry/competitive thesis, valuation, event timing, risk interaction, communication, or process discipline. 1. Separate bad process with a good outcome from good process with a bad outcome and from pure luck. 1. Measure forecast error by driver, not only EPS/revenue, to identify systematic optimism, conservatism, or missed nonlinear relationships. 1. Write the process change required to prevent a repeat: new source, check, scenario, review gate, cadence, or training drill. 1. Feed recurring errors into analyst training, model QA, and research-priority rules rather than merely documenting them. ## Required evidence and model bridge - Primary-source set: risk register, stress model, correlation/liquidity inputs, decision journal, forecast-error history. Preserve exact document/version, date, period, and source location for every material factual input used in post-mortems and error taxonomy. - For each key concept - original record, data errors, accounting errors, model errors, forecast errors, thesis errors - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as post-mortems and error taxonomy may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | error contribution | Absolute or signed forecast/valuation error attributable to a specific error category divided by total error across the decision. | error contribution: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | repeat-error rate | Previously identified error types recurring in later analyses divided by total subsequent analyses or relevant decisions. | repeat-error rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | forecast calibration | Actual outcome relative to the analyst's stated forecast distribution, range, or scenario probabilities; track coverage, bias, and dispersion over time. | forecast calibration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: macro rates hurt the stock while company operations meet assumptions. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For post-mortems and error taxonomy, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through original record, data errors, accounting errors, model errors, then identify which link is directly observed and which link remains an assumption. - Calculate error contribution, repeat-error rate, forecast calibration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for post-mortems and error taxonomy: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: separate bad outcome from bad process and make each lesson change a concrete control. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the post-mortems and error taxonomy conclusion. ## Failure tests - FAIL if original record cannot be defined and reproduced from the source pack. - FAIL if outcome quality is confused with process quality or if errors are not classified into source, model, assumption, timing, behavior, risk, and luck with a corrective action. - FAIL if the post-mortems and error taxonomy conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the post-mortems and error taxonomy conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the post-mortems and error taxonomy conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY | MODULE 066 # Investment Memo Writing > Mission. Write concise, decision-useful research that separates facts, assumptions, variant views, valuation, and risks. ## Decision output Objective: Write concise, decision-useful research that separates facts, assumptions, variant views, valuation, and risks. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Lead with the decision-relevant conclusion, variant perception, valuation range, key evidence, and what changes the view; do not force readers through chronology before the answer. 1. Label material statements as fact/source, management claim, analyst calculation, external estimate, or judgment when ambiguity could mislead. 1. Explain the business and industry in causal terms, then show forecast drivers and why they differ from market-implied expectations. 1. Present valuation with explicit scenarios, EV-to-equity bridge, sensitivities, and terminal/normalization assumptions rather than a target price without mechanics. 1. Give risks equal analytical quality: mechanism, leading indicator, exposure, and thesis-break condition. 1. Maintain a concise decision version and an evidence appendix/model that allows full audit without bloating the main narrative. ## Required evidence and model bridge - Primary-source set: investment memo, evidence pack, model outputs, monitoring dashboard, review and automation logs. Preserve exact document/version, date, period, and source location for every material factual input used in investment memo writing. - For each key concept - decision conclusion, business drivers, facts versus estimates, variant view, valuation bridge, risks - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as investment memo writing may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | claim traceability | % of material factual claims that link to a dated source, exact page/section, and model or memo location. | claim traceability: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | thesis falsifiability | % of major thesis claims paired with a measurable future observation that would materially reduce confidence or invalidate the claim. | thesis falsifiability: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | decision density | Number of decision-relevant conclusions, quantified implications, or action triggers per page/minute, excluding background description. | decision density: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | ## Worked application > Case: replace "huge TAM, best in class" with measurable expectations and evidence. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For investment memo writing, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through decision conclusion, business drivers, facts versus estimates, variant view, then identify which link is directly observed and which link remains an assumption. - Calculate claim traceability, thesis falsifiability, decision density from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for investment memo writing: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: make the bear case strong enough that a skeptic recognizes it. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the investment memo writing conclusion. ## Failure tests - FAIL if decision conclusion cannot be defined and reproduced from the source pack. - FAIL if the reader cannot identify the thesis, variant view, key drivers, valuation range, downside, evidence quality, and what changes the view from the first two pages. - FAIL if the investment memo writing conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the investment memo writing conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the investment memo writing conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY | MODULE 067 # Charts, Tables, and Evidence Design > Mission. Use visuals to reveal relationships rather than decorate the narrative. ## Decision output Objective: Use visuals to reveal relationships rather than decorate the narrative. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Every visual must answer a decision question faster or more accurately than prose. Remove charts that merely decorate the memo. 1. Use consistent units, dates, scale, definitions, source notes, and axis treatment; disclose breaks in series and avoid truncated axes that exaggerate change without clear purpose. 1. Prefer driver bridges, cohort curves, scatterplots with economic interpretation, waterfall/reconciliation charts, and scenario tables over dense dashboards with no hierarchy. 1. Annotate material events and definition changes so correlation is not mistaken for causation. 1. Avoid dual axes unless the relationship truly requires them and labels/scales make interpretation unambiguous. 1. Design tables so reported, normalized, forecast, and scenario data are visually distinct and a reviewer can trace key numbers to sources. ## Required evidence and model bridge - Primary-source set: investment memo, evidence pack, model outputs, monitoring dashboard, review and automation logs. Preserve exact document/version, date, period, and source location for every material factual input used in charts, tables, and evidence design. - For each key concept - question-first visual choice, scales, denominators, annotations, uncertainty, source/date - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as charts, tables, and evidence design may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | source completeness | % of material claims and model inputs with complete source metadata, including document, date, page/section, and retrieval/version information. | source completeness: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | denominator integrity | % of KPI/ratio calculations using a denominator that matches the numerator's scope, period, geography, cohort, and accounting definition. | denominator integrity: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | question-answer speed | Elapsed time from a senior-review/IC question to a sourced, reconciled, decision-useful answer using the maintained research system. | question-answer speed: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: a truncated axis visually exaggerates a 100bp margin change. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For charts, tables, and evidence design, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through question-first visual choice, scales, denominators, annotations, then identify which link is directly observed and which link remains an assumption. - Calculate source completeness, denominator integrity, question-answer speed from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for charts, tables, and evidence design: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use visuals to compress causal evidence, not decorate the memo. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the charts, tables, and evidence design conclusion. ## Failure tests - FAIL if question-first visual choice cannot be defined and reproduced from the source pack. - FAIL if a visual omits units, period, source, definition, or uses scales and formatting that distort the economic relationship it is meant to show. - FAIL if the charts, tables, and evidence design conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the charts, tables, and evidence design conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the charts, tables, and evidence design conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY | MODULE 068 # Investment Committee Defense > Mission. Prepare for adversarial questions, alternative explanations, and explicit uncertainty. ## Decision output Objective: Prepare for adversarial questions, alternative explanations, and explicit uncertainty. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Prepare a one-page steelman of the opposing thesis before the meeting, including the strongest evidence and valuation case against your view. 1. Know the source and definition of every decision-critical number and be prepared to reproduce the bridge rather than defend a memorized output. 1. Identify the two assumptions that drive most valuation dispersion and show operating sensitivities and thesis-break cases in advance. 1. Answer uncertainty directly. Distinguish what is known, estimated, judgmental, and unresolved instead of improvising precision under pressure. 1. When a committee objection is valid, update the analysis rather than defending sunk work. Record open questions, owner, source, and deadline. 1. After the meeting, log which objections changed assumptions and which were resolved by evidence so committee review improves the research process. ## Required evidence and model bridge - Primary-source set: investment memo, evidence pack, model outputs, monitoring dashboard, review and automation logs. Preserve exact document/version, date, period, and source location for every material factual input used in investment committee defense. - For each key concept - highest-sensitivity assumptions, weakest evidence, opposing thesis, source defense, coherent scenarios, open questions - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as investment committee defense may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | open-question value exposure | Sum of absolute valuation sensitivity × plausible uncertainty range for unresolved questions, adjusted for probability when appropriate. | open-question value exposure: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | assumption concentration | Share of total valuation variance or expected-value dispersion attributable to the top one or top three assumptions. | assumption concentration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | bear-case coherence | % of bear-case assumptions that are causally linked and internally consistent across revenue, margins, working capital, capex, financing, and valuation. | bear-case coherence: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Worked application > Case: committee challenges a 35% terminal margin supported only by management targets. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For investment committee defense, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through highest-sensitivity assumptions, weakest evidence, opposing thesis, source defense, then identify which link is directly observed and which link remains an assumption. - Calculate open-question value exposure, assumption concentration, bear-case coherence from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for investment committee defense: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: answer directly, show the bridge, and change the view when evidence defeats the thesis. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the investment committee defense conclusion. ## Failure tests - FAIL if highest-sensitivity assumptions cannot be defined and reproduced from the source pack. - FAIL if the analyst cannot identify the highest-sensitivity weakest-evidence assumptions, strongest contrary evidence, and the model impact of changing them. - FAIL if the investment committee defense conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the investment committee defense conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the investment committee defense conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY | MODULE 069 # Analyst Training and Deliberate Practice > Mission. Build a 12-month progression from filings and accounting to modeling, industry expertise, and judgment. ## Decision output Objective: Build a 12-month progression from filings and accounting to modeling, industry expertise, and judgment. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Sequence training from source retrieval and accounting to historical reconstruction, driver modeling, valuation, industry specialization, decision process, and communication. 1. Use deliberate drills with objectively checkable outputs: statement tie-outs, footnote extraction, revenue bridges, working-capital builds, DCF reconstruction, reverse DCF, and memo defense. 1. Maintain an error log by category and require the analyst to write the process control that prevents recurrence. 1. Train on the same company across several quarters so the analyst experiences estimate formation, surprise, model updates, and post-mortem rather than only static case studies. 1. Add sector specialization after core competence, including native KPIs, cycle indicators, accounting traps, and value-chain economics. 1. Advance responsibility only when the analyst can reproduce work, explain uncertainty, defend an opposing case, and update beliefs when evidence changes. ## Required evidence and model bridge - Primary-source set: investment memo, evidence pack, model outputs, monitoring dashboard, review and automation logs. Preserve exact document/version, date, period, and source location for every material factual input used in analyst training and deliberate practice. - For each key concept - skill matrix, progressive cases, closed-book drills, mentor review, red teaming, error taxonomy - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as analyst training and deliberate practice may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | mastery rubric | Weighted competency score across sourcing, accounting, modeling, valuation, industry work, risk, communication, speed, and error control, using predefined pass standards. | mastery rubric: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. | | repeat-error rate | Previously identified error types recurring in later analyses divided by total subsequent analyses or relevant decisions. | repeat-error rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | probability calibration | Realized event frequency within each stated probability bucket compared with the bucket probability; measure Brier score or absolute calibration error. | probability calibration: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Worked application > Case: trainee must reconstruct filings and models before making a stock call. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For analyst training and deliberate practice, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through skill matrix, progressive cases, closed-book drills, mentor review, then identify which link is directly observed and which link remains an assumption. - Calculate mastery rubric, repeat-error rate, probability calibration from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for analyst training and deliberate practice: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: advance responsibility after demonstrated mastery, not time served. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the analyst training and deliberate practice conclusion. ## Failure tests - FAIL if skill matrix cannot be defined and reproduced from the source pack. - FAIL if advancement is based on reading or tenure rather than reproducible source work, model builds, oral defense, timed drills, and documented error reduction. - FAIL if the analyst training and deliberate practice conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the analyst training and deliberate practice conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the analyst training and deliberate practice conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. ## PART XIV - RESEARCH COMMUNICATION AND MASTERY | MODULE 070 # Research System Automation and AI Assistance > Mission. Use automation and AI for retrieval, extraction, QA, and scenario generation while keeping human verification and accountability. ## Decision output Objective: Use automation and AI for retrieval, extraction, QA, and scenario generation while keeping human verification and accountability. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule. ## Explicit operating procedure 1. Build an immutable source-ingestion layer with document identifier, URL/accession, timestamp, version/hash, parsed text/tables, and exact source spans. 1. Use structured extraction schemas carrying entity, metric, value, unit, period, definition, source location, and confidence; route discrepancies and low-confidence fields to review. 1. Use retrieval that prioritizes approved primary sources and preserves provenance. Retrieved document instructions are untrusted data, not executable commands. 1. Use deterministic code or spreadsheet logic for arithmetic, reconciliations, valuation, and file writes whenever possible; use language models for classification, contradiction search, extraction assistance, hypothesis generation, and drafting under checks. 1. Run automated numeric and definition QA: statement totals, XBRL versus filing tables, segment reconciliation, signs/units, current versus prior definitions, stale-source detection, and citation support. 1. Maintain a golden evaluation set and track extraction accuracy, citation accuracy, numeric reconciliation, hallucination rate, false-positive red flags, regression stability, analyst correction rate, and net automation yield. 1. Require human approval before material assumption changes, published conclusions, compliance-sensitive use of channel data, or external actions; version prompts, models, schemas, code, and reviewer decisions so prior outputs are reproducible. ## Required evidence and model bridge - Primary-source set: investment memo, evidence pack, model outputs, monitoring dashboard, review and automation logs. Preserve exact document/version, date, period, and source location for every material factual input used in research system automation and ai assistance. - For each key concept - source ingestion, immutable raw store, structured extraction, RAG, provenance, deterministic math - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence. - Map only economically relevant findings into the model or decision record. Process-control modules such as research system automation and ai assistance may have no direct valuation line; in that case document the downstream error or governance risk the control prevents. ## Metrics and calculation controls | Metric / concept | Construction | Required validation | | --- | --- | --- | | citation accuracy | % of sampled citations that directly support the adjacent factual claim with the correct document, date, page/section, and interpretation. | citation accuracy: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | numeric reconciliation | % of sampled extracted or modeled figures that reproduce the primary-source value after unit, sign, scale, and period normalization. | numeric reconciliation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | hallucination rate | Unsupported or materially incorrect AI-generated factual claims divided by all sampled factual claims before human correction. | hallucination rate: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | automation yield | automation yield = annualized economic output / current market value or invested base; match numerator and denominator. | automation yield: Recalculate automation yield from cited inputs; reconcile definition, period, units, signs, and source version; investigate and document any variance before use. | ## AI research system architecture - Create an immutable raw-source layer. Store filing accession/URL, publication timestamp, document hash/version, extracted text, tables, and source spans before any LLM transformation. - Use structured schemas for extraction. Each extracted value should carry company, period, metric, value, unit, source document, page/section, exact supporting span, and confidence. - Use deterministic code for arithmetic, reconciliations, valuation, and spreadsheet writes whenever possible. The LLM can propose mappings or explanations, but calculations should be reproducible. - Build retrieval with source whitelists and provenance. Primary-source passages should outrank summaries. The answer layer must cite the exact evidence used, not a nearby document. - Run automated checks: totals versus components, XBRL versus filing table, current versus prior filing definition, sign/unit consistency, and balance-sheet/segment reconciliation. Route failures to human review. - Maintain an evaluation set containing known filings and expected outputs. Track extraction accuracy, citation support, numeric reconciliation pass rate, hallucination rate, false-positive red flags, and analyst correction rate over time. - Defend against prompt injection and untrusted content. Treat instructions inside retrieved documents/web pages as data, not executable instructions. Separate retrieval, transformation, and action permissions. - Human approval gates are mandatory before changing valuation assumptions, publishing a research conclusion, using channel evidence with compliance implications, or taking any external action. - Version prompts, models, schemas, source sets, and outputs. A research result is not auditable if a reviewer cannot reconstruct which model and evidence produced it. ## Worked application > Case: LLM extracts 10-Q numbers that are checked against XBRL and source spans before model write. - Reconstruct the relevant reported fact from primary evidence before interpreting the case. For research system automation and ai assistance, show the raw components rather than only the resulting ratio or narrative. - Build the causal chain through source ingestion, immutable raw store, structured extraction, RAG, then identify which link is directly observed and which link remains an assumption. - Calculate citation accuracy, numeric reconciliation, hallucination rate, automation yield from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation. - Translate the difference between cases into the variable that matters for research system automation and ai assistance: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them. - Expert consistency test: use AI for retrieval, extraction, contradiction hunting, and drafting while humans retain factual and decision accountability. - Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the research system automation and ai assistance conclusion. ## Failure tests - FAIL if source ingestion cannot be defined and reproduced from the source pack. - FAIL if an AI-produced material fact, calculation, model change, or conclusion cannot be traced to versioned sources, deterministic checks, evaluated workflows, and an accountable human approval. - FAIL if the research system automation and ai assistance conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version. - FAIL if evidence materially inconsistent with the research system automation and ai assistance conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis. ## Completion test A senior reviewer must be able to reproduce the research system automation and ai assistance conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open. # PART XV - SECTOR PLAYBOOKS: COMMON PROTOCOL - For every sector, build five annual years and at least eight quarters of the listed KPIs when available, extending through a prior cycle where relevant. - Record issuer definition, normalized definition, source/date, unit, and definition changes for every KPI before comparing companies. - Decompose demand, supply, price, cost, capital, and cycle. Identify the binding constraint and marginal price setter. - Create a price-volume-mix-capacity bridge and reconcile operating evidence to statements, working capital, capex, financing, and cash. - Benchmark direct peers only after normalizing geography, product mix, capital structure, accounting, and KPI definitions. - Use at least two sector-appropriate valuation methods plus reverse expectations. Do not average incompatible methods to hide disagreement. - Build a sector-specific bear/stress case through operating drivers first, then liquidity, dilution, and valuation. - Maintain a sector dashboard, historical KPI bridge, peer table, driver forecast, scenario matrix, valuation, and risk register. ## PART XV - SECTOR PLAYBOOKS | MODULE 071 # Software and SaaS Analyst Playbook > Mission. Build a sector-specific research system for Software and SaaS that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Decompose recurring revenue into beginning cohort, churn, contraction, expansion, new logos, price, and usage. Reconcile RPO/cRPO to billing terms and revenue recognition. Test whether FCF depends on durable gross profit or temporary working-capital and SBC effects. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | ARR or subscription revenue | Annualized recurring revenue from active subscription contracts at period end; exclude one-time services and clearly separate usage-based or non-recurring revenue. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | net revenue retention | NRR = beginning-cohort recurring revenue after churn, contraction, and expansion / beginning-cohort recurring revenue | | gross retention | GRR = beginning-cohort recurring revenue retained before expansion / beginning-cohort recurring revenue | | RPO/cRPO | Remaining performance obligations are contracted revenue not yet recognized; cRPO is the portion expected to be recognized within the next 12 months. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | CAC payback | CAC payback months = customer acquisition cost / monthly gross profit from new customer | | sales efficiency | Incremental gross profit or ARR generated per dollar of sales and marketing spend; use a consistent lag between spend and bookings/ARR creation. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | gross margin | Gross margin = gross profit / revenue | | FCF margin | FCF margin = normalized free cash flow / revenue | ## Sector-specific accounting and comparability traps - Capitalized commissions: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Stock compensation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Multi-year contract timing: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Seat versus usage pricing: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Restructuring add-backs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/revenue: enterprise value divided by normalized revenue; use only with an explicit gross-margin, operating-margin, growth, and capital-intensity bridge. - EV/gross profit: enterprise value divided by normalized gross profit; useful when revenue recognition/pass-through differs, but still requires opex and capital-intensity normalization. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - Rule of 40 framing: revenue growth plus a consistently defined FCF or operating margin; use as a diagnostic, not a valuation substitute, and reconcile SBC/capitalization policies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Software and SaaS, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress ARR or subscription revenue and net revenue retention together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test capitalized commissions. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model recurring revenue through beginning ARR, new ARR, gross churn, contraction, expansion, price, usage, and FX. Convert bookings and RPO to revenue only after modeling contract duration and billing terms. ### Leading-indicator dashboard Track renewal cohorts, net retention, sales capacity, cloud-consumption trends, customer optimization, deferred revenue, RPO duration, and seat-versus-usage mix. ### Primary-source map SEC 10-K/10-Q/8-K and XBRL; issuer ARR/RPO/NRR definitions and earnings materials; contract/pricing documentation; peer filings; customer procurement and cloud-platform disclosures where publicly available. ### Accounting normalization test Capitalized commissions, SBC, restructuring, acquisitions, usage revenue timing, and multi-year prepayments can distort FCF and growth comparisons. ### Valuation implementation Use DCF, EV/FCF, and revenue or gross-profit multiples only with explicit mature margin and reinvestment assumptions. Reverse the current multiple into retention, growth duration, and terminal margin. ### Worked numerical mini-case > Illustrative cohort case. Beginning ARR $500m, gross churn 7%, contraction 3%, expansion 18%, new ARR $95m. Ending ARR = 500 x (1 - .07 - .03 + .18) + 95 = $635m, or 27% growth. NRR on the opening cohort is 108%. Do not forecast 27% revenue growth mechanically. Convert ARR to revenue using contract start dates, billing cadence and usage mix, then test whether CAC payback and FCF improve or deteriorate. ### Monitoring and falsification cadence Thesis breaks often emerge through retention, competitive displacement, rising customer acquisition cost, or a shift from scarce software to commoditized functionality. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Software and SaaS work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 072 # Semiconductors Analyst Playbook > Mission. Build a sector-specific research system for Semiconductors that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model units, wafer starts, die size/yield, node mix, ASP, utilization, inventory, and customer/channel inventory. Separate design-win timing from revenue conversion and stress utilization-driven gross margin. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | wafer starts | Number of wafers entering fabrication during the period; reconcile owned-fab, foundry, node, and wafer-size mix before comparing companies. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | units | Physical units shipped, sold, or produced for the defined product scope and period; reconcile returns, channel inventory, and product mix. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | ASP | Revenue attributable to the relevant product family divided by units sold/shipped, adjusted for rebates, mix, and channel treatment. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | inventory days | Inventory days = average inventory / COGS x days in period | | book-to-bill | Book-to-bill = bookings / recognized revenue for the same definition and period | | design wins | Customer platforms/programs formally awarded or qualified during the period; track expected lifetime revenue, production start, and conversion to shipments. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | node mix | Share of wafer starts, revenue, or units produced on each process node; use the same denominator across periods and peers. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Inventory reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Customer concentration: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Channel inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capital intensity: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Government incentives: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Through-cycle P/E: equity value divided by normalized cycle-average EPS, with peak/trough margins, working capital, credit, and capital spending normalized. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - Replacement-cost context: estimate current cost to recreate productive assets/capacity, adjust for age, technology, location, permits, and time-to-build, then compare EV with replacement value. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Semiconductors, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress wafer starts and units together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test inventory reserves. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model units, wafer starts, die size, yield, ASP, node mix, utilization, packaging/test, and inventory across end markets. Separate fabless economics from foundry and IDM capital intensity. ### Leading-indicator dashboard Track book-to-bill where meaningful, lead times, distributor inventory, foundry utilization, equipment orders, memory pricing, design wins, product qualification, and customer capex. ### Primary-source map SEC filings; foundry and equipment-vendor filings; Semiconductor Industry Association and WSTS industry data; U.S. Commerce/BIS rules; customer capex and inventory disclosures. ### Accounting normalization test Channel inventory, customer concentration, purchase commitments, capitalized manufacturing cost, government incentives, and rapid obsolescence can distort cycle signals. ### Valuation implementation Use mid-cycle earnings, DCF, EV/EBIT, and FCF yield with normalized utilization and capex. Peak-cycle low multiples can be value traps. ### Worked numerical mini-case > Illustrative utilization case. A fab has 100k wafer starts/month, 82% utilization, $2,400 revenue per utilized wafer and 48% gross margin. If utilization falls to 68% with fixed conversion cost unchanged, model both the 17% shipment decline and the margin de-absorption rather than cutting revenue alone. The investment conclusion should separate cyclical utilization recovery from structural node/share loss. ### Monitoring and falsification cadence Key breaks include architecture displacement, node or packaging disadvantage, customer insourcing, excess capacity, and product cycles that fail to convert design wins into volume. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Semiconductors work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 073 # AI Accelerators and Compute Analyst Playbook > Mission. Build a sector-specific research system for AI Accelerators and Compute that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Trace accelerator units, ASP, HBM content, packaging capacity, rack power, networking content, lead times, customer capex, and utilization. Distinguish supply-constrained shipments from sustainable end demand and test concentration risk. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | accelerator shipments | AI accelerator units or systems shipped and recognized in revenue during the period, separated by product generation and form factor where disclosed. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | ASP | Revenue attributable to the relevant product family divided by units sold/shipped, adjusted for rebates, mix, and channel treatment. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | HBM content | HBM memory capacity or dollar content attached per accelerator/system multiplied by accelerator shipments; distinguish generations and supplier mix. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | rack density | IT/accelerator power load or compute capacity per rack, typically kW or accelerators per rack, using deployed rather than theoretical maximum configuration. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | power per rack | Average or design electrical demand per deployed rack in kW, including the defined IT load and clearly stating whether cooling/overhead is excluded. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | lead time | Elapsed time from order acceptance to shipment/installation for the relevant product or capacity, reported as median/range and compared with normal cycle levels. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | cloud capex | Capital expenditures by hyperscale/cloud customers attributable to compute, networking, data-center buildings, and power infrastructure; normalize definition across companies. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Supply allocation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Customer concentration: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Rapid obsolescence: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized software: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warranty and returns: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/sales with margin bridge: enterprise value divided by revenue plus an explicit path from gross margin to EBIT/FCF, including SBC, capex, working capital, and terminal margin assumptions. - EV/EBIT: enterprise value divided by normalized operating profit after depreciation; useful where depreciation is economically meaningful and capital intensity differs. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. - Reverse DCF: solve for the growth, margin, reinvestment, ROIC, and duration assumptions required for the current market price, then compare those expectations with evidence. ## Sector diligence questions - What is the most important leading indicator for AI Accelerators and Compute, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress accelerator shipments and ASP together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test supply allocation. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model accelerator units, ASP, HBM content, networking, rack density, power per rack, software attach, cloud utilization, and customer concentration. Constrain shipments by packaging, memory, foundry, power, and deployment capacity. ### Leading-indicator dashboard Track hyperscaler capex, backlog, lead time, CoWoS or advanced-packaging capacity, HBM supply, rack deliveries, power procurement, model efficiency, utilization, and custom silicon adoption. ### Primary-source map SEC filings; hyperscaler capex disclosures; foundry and advanced-packaging supplier filings; memory-vendor filings; U.S. Commerce/BIS export-control releases; public power/interconnection disclosures. ### Accounting normalization test Supply allocation, customer prepayments, concentrated demand, rapid product transitions, warranty, capitalized software, and channel inventory can make short-term revenue nonrepeatable. ### Valuation implementation Reverse DCF should solve for unit growth, ASP/content, mature margin, and duration. Test value under custom silicon, efficiency gains, and lower accelerator intensity per workload. ### Worked numerical mini-case > Illustrative supply-constrained case. Demand supports 1.20m accelerators at $28k ASP, but packaging/HBM capacity limits shipments to 0.90m. Revenue is constrained to $25.2bn before networking/software attach. A 15% capacity expansion raises the ceiling to 1.035m, not to unconstrained demand. Value the bottleneck duration explicitly and stress custom silicon, utilization and compute-efficiency improvements. ### Monitoring and falsification cadence Thesis breaks include demand concentration reversing, model-efficiency reducing compute intensity faster than workload growth, customer vertical integration, or supply scarcity rents normalizing sooner than expected. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The AI Accelerators and Compute work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 074 # Cloud and Data Centers Analyst Playbook > Mission. Build a sector-specific research system for Cloud and Data Centers that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model commissioned MW, booked MW, utilization, power availability, PUE, lease pricing, construction cost, time to energization, and financing. Power interconnection and equipment lead times are often the binding growth constraints. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | MW commissioned | Data-center critical IT megawatts placed in service and available for customer load during the period. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | MW under construction | Critical IT megawatts in active construction with committed capital and defined expected delivery dates; separate owned and partner capacity. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | booked MW | Critical IT megawatts covered by signed customer commitments or leases, whether operating or under construction; disclose commencement timing. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | power cost | Electricity expense divided by kWh consumed, or contracted all-in $/MWh including delivery/hedge effects where relevant. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | revenue per MW | Annualized data-center revenue divided by average commissioned or occupied critical IT MW, using a consistent occupancy basis. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | PUE | Total facility energy consumed divided by IT equipment energy consumed; lower values indicate less non-IT overhead. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | ## Sector-specific accounting and comparability traps - Lease accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized interest: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Construction-in-progress: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Joint ventures: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Power commitments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - EV/MW: enterprise value divided by owned/contracted operating MW, adjusted for development stage, technology, duration, capacity factor, PPAs, debt, and project economics. - AFFO: start with FFO and deduct recurring capital expenditures and economically recurring adjustments; use a clearly defined issuer-independent AFFO before applying a multiple. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Cloud and Data Centers, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress MW commissioned and MW under construction together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test lease accounting. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model commissioned MW, booked MW, utilization, revenue per MW or rack, power cost, PUE, land, interconnection, construction cost, and lease duration. Separate powered shell, wholesale, colocation, and cloud service economics. ### Leading-indicator dashboard Track utility interconnection queues, transformer/switchgear lead times, land and power contracts, preleasing, customer capex, construction starts, financing spreads, and regional power prices. ### Primary-source map SEC filings and REIT/developer supplements; EIA and utility tariffs; ISO/RTO interconnection queues; local permitting/land records; hyperscaler capex and lease disclosures. ### Accounting normalization test Backlog can include long-dated options; construction in progress, capitalized interest, leases, and development JV accounting can flatter near-term operating metrics. ### Valuation implementation Use project-level DCF/NAV plus corporate valuation. Cap rates and EV/EBITDA need normalization for development pipeline and capital intensity. ### Worked numerical mini-case > Illustrative MW economics. A 100 MW campus is 80% leased at $185/kW-month. Annualized gross revenue = 100,000 kW x 80% x $185 x 12 = $177.6m. If delivered project cost is $10m/MW, test unlevered yield against financing cost and required return. Do not value booked MW as operating MW. Model energization, lease commencement, tenant improvements and power availability. ### Monitoring and falsification cadence Thesis breaks include power unavailable on schedule, customer concentration, overbuilding in a region, financing cost exceeding project returns, or technology reducing space/power demand per workload. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Cloud and Data Centers work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 075 # Industrial Machinery Analyst Playbook > Mission. Build a sector-specific research system for Industrial Machinery that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Bridge orders, backlog, cancellations, shipments, price, mix, utilization, service revenue, labor/material costs, and working capital. Map dealer inventory and end-market cycle separately from reported backlog. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | orders | Gross value of customer purchase orders or bookings received during the period, adjusted for cancellations and scope changes where disclosed. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | book-to-bill | Book-to-bill = bookings / recognized revenue for the same definition and period | | price-cost | Change in realized selling price minus change in relevant input and conversion cost on a comparable-unit basis; express in dollars and margin points. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | service mix | Service revenue or gross profit divided by total revenue or gross profit, with recurring aftermarket separated from project/install services. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | capacity utilization | Actual production/output divided by practical productive capacity for the period after downtime, maintenance, and yield constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | inventory turns | Annualized COGS divided by average inventory; for retailers also monitor weeks of supply and aged inventory. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | incremental margin | incremental margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | ## Sector-specific accounting and comparability traps - Percentage-of-completion: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warranty reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Restructuring: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Cyclical inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - Mid-cycle P/E: current equity value divided by estimated mid-cycle diluted EPS using normalized volumes, prices, margins, credit, taxes, and share count. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Industrial Machinery, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress orders and backlog together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test percentage-of-completion. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model units, installed base, price, mix, aftermarket, utilization, backlog conversion, dealer inventory, labor hours, and material cost. Distinguish original equipment from higher-margin service. ### Leading-indicator dashboard Track PMI/capex indicators, dealer inventories, lead times, book-to-bill, rental utilization, construction/manufacturing activity, freight, commodity inputs, and order cancellations. ### Primary-source map SEC filings; U.S. Census durable-goods and construction/manufacturing data; Federal Reserve industrial production; ISM surveys; dealer/rental-company filings and backlog disclosures. ### Accounting normalization test Percent-completion, backlog quality, dealer financing, restructuring, pension, and working-capital swings can obscure normalized earnings. ### Valuation implementation Use mid-cycle EBIT/FCF, DCF, and SOTP where aftermarket differs materially. Normalize margin for utilization and price-cost cycle. ### Worked numerical mini-case > Illustrative price-volume-mix case. Prior revenue $2.0bn. Units decline 8%, price rises 5%, mix adds 2%, and aftermarket grows 6%. Build equipment and aftermarket separately so price-cost and installed-base resilience are visible. A headline flat-revenue quarter can hide an OEM downturn if aftermarket and price are doing all the work. ### Monitoring and falsification cadence Breaks include dealer destocking, capacity overbuild, aftermarket disruption, aggressive price-cost assumptions, or structurally lower end-market capital intensity. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Industrial Machinery work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 076 # Aerospace and Defense Analyst Playbook > Mission. Build a sector-specific research system for Aerospace and Defense that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model backlog quality, funded versus unfunded demand, production rates, learning curves, customer advances, milestone payments, contract mix, loss provisions, and supplier bottlenecks. Cash timing can diverge materially from accounting profit. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | bookings | Value of customer commitments accepted during the period under the company's booking policy, net of cancellations where disclosed. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | funded backlog | Backlog supported by appropriated/authorized customer funding divided by total backlog; particularly relevant for government/defense programs. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | delivery rates | Units, aircraft, systems, or program milestones delivered per month/quarter relative to planned schedule and contractual commitments. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | aftermarket mix | Aftermarket/service revenue or gross profit divided by total revenue or gross profit for the relevant installed-base business. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | program margin | program margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | | cash conversion | Cash conversion cycle = DSO + inventory days - DPO | | R&D | Research and development expense, plus material capitalized development when applicable, divided by revenue or analyzed by absolute spend and program mix. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | ## Sector-specific accounting and comparability traps - Contract accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Loss reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Customer advances: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Program charges: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Aerospace and Defense, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress bookings and backlog together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test contract accounting. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model program units, shipsets, content per platform, backlog, production rates, aftermarket flight hours, contract type, cost curves, and government budgets. ### Leading-indicator dashboard Track OEM build rates, supplier deliveries, engine removals, flight hours, defense appropriations, contract awards, program milestones, and quality/regulatory actions. ### Primary-source map SEC filings; FAA certification and airworthiness material; U.S. DoD budget and contract awards; prime/OEM production-rate disclosures; Bureau of Transportation Statistics and flight-hour data. ### Accounting normalization test Program accounting, loss reserves, customer advances, pension, cost-to-complete estimates, and supplier concessions require deep footnote work. ### Valuation implementation Use DCF, EV/EBIT, FCF yield, and SOTP. Value long-cycle defense backlog differently from commercial aftermarket and new-platform ramps. ### Worked numerical mini-case > Illustrative program case. A supplier has $1.5m content per aircraft at 40 monthly shipsets. A planned ramp to 50 implies $180m additional annual revenue before scrap, learning curve and supplier constraints. If actual deliveries cap at 44, the model must use the constrained rate. For fixed-price programs, pair volume upside with cost-to-complete and loss-reserve sensitivity. ### Monitoring and falsification cadence Breaks include certification/quality failure, supplier bottlenecks, program cancellation, fixed-price cost overrun, or production-rate assumptions that exceed supply-chain capability. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Aerospace and Defense work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 077 # Airlines Analyst Playbook > Mission. Build a sector-specific research system for Airlines that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Build capacity, load factor, yield, PRASM, CASM ex fuel, fuel, labor, aircraft ownership, maintenance, and liquidity. Stress demand and fare simultaneously while respecting largely fixed near-term capacity and debt obligations. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | ASM | PRASM = passenger revenue / available seat miles | | RPM | Revenue passenger miles: paying passenger miles flown, equal to revenue passengers multiplied by distance traveled. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | load factor | Load factor = revenue passenger miles / available seat miles | | yield | FCF yield = normalized free cash flow to equity / current equity value | | PRASM | PRASM = passenger revenue / available seat miles | | CASM ex fuel | CASM = operating expense / available seat miles; specify fuel and special-item exclusions | | fuel price | Average economic fuel cost per gallon/liter after hedge settlements and taxes, reconciled to reported fuel expense and consumption. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | liquidity | Unrestricted cash and short-term investments plus committed undrawn credit capacity less unavailable/restricted amounts and near-term mandatory uses. Validation: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | ## Sector-specific accounting and comparability traps - Aircraft leases: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Loyalty economics: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Maintenance reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Hedges: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Deferred revenue: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDAR: lease-adjusted enterprise value divided by EBITDAR; capitalize or otherwise normalize rent consistently and test fleet/asset replacement requirements. - Through-cycle FCF: estimate normalized free cash flow across a full cycle, including mid-cycle prices/margins, working capital, maintenance capex, cash taxes, and financing needs. - Asset value: estimate market or DCF value of identifiable assets/project interests, subtract asset-level and corporate liabilities, and apply ownership, tax, and liquidity adjustments. - Liquidity-adjusted equity value: base enterprise/equity value reduced for expected cash burn, refinancing cost, dilution, covenant stress, and probability of distressed financing before thesis realization. ## Sector diligence questions - What is the most important leading indicator for Airlines, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress ASM and RPM together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test aircraft leases. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model ASM, RPM, load factor, yield, ancillary revenue, fuel, labor, fleet, utilization, maintenance, and capacity by region. ### Leading-indicator dashboard Track booking curves, fare data, TSA/passenger data, corporate travel, competitor capacity, fuel, aircraft deliveries, maintenance events, and credit-card remuneration. ### Primary-source map SEC filings; U.S. DOT Form 41 and T-100 data; TSA throughput; FAA fleet/operational data; EIA jet-fuel prices; airport and competitor capacity schedules. ### Accounting normalization test Sale-leasebacks, loyalty-program economics, maintenance capitalization, pension, and deferred ticket revenue can complicate cash and leverage. ### Valuation implementation Use normalized EV/EBITDAR or EV/EBIT, FCF through cycle, and asset/liquidity analysis. Peak travel margins should not be capitalized perpetually. ### Worked numerical mini-case > Illustrative unit-revenue case. ASM grows 6%, load factor falls from 84% to 82%, and yield rises 3%. RPM grows only about 3.5% before mix. Combine PRASM with CASM ex-fuel and fuel per gallon to determine whether added capacity creates or destroys EBIT. A capacity plan is not bullish if fare dilution and labor/fuel costs overwhelm unit growth. ### Monitoring and falsification cadence Breaks include capacity oversupply, labor inflation, aircraft constraints, fuel spikes without pricing, balance-sheet stress, or loyalty economics weakening. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Airlines work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 078 # Automotive OEMs Analyst Playbook > Mission. Build a sector-specific research system for Automotive OEMs that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model units by region/model, transaction price, incentives, mix, warranty, manufacturing utilization, battery/material cost, dealer inventory, captive finance, capex, and launch cadence. Separate reported ASP from mix-driven price changes. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | unit deliveries | Vehicles or other finished units delivered to end customers and accepted for revenue recognition during the period. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | ASP | Revenue attributable to the relevant product family divided by units sold/shipped, adjusted for rebates, mix, and channel treatment. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | incentives | Manufacturer/customer incentives, rebates, financing subsidies, and discounts per unit or as a % of gross selling price. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | mix | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | days supply | Channel or dealer inventory units divided by average daily retail sales over a recent normalized period. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | warranty | Warranty expense or provision divided by product sales, supported by claims frequency, cost per claim, and reserve roll-forward. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | finance penetration | Financed or leased unit sales using captive/partner financing divided by total eligible retail unit sales. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | capacity utilization | Actual production/output divided by practical productive capacity for the period after downtime, maintenance, and yield constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Dealer inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warranty reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Finance subsidiaries: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Lease residuals: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized development: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBIT: enterprise value divided by normalized operating profit after depreciation; useful where depreciation is economically meaningful and capital intensity differs. - Normalized P/E: equity value divided by through-cycle diluted EPS after normalizing credit, reserves, margins, taxes, unusual gains/losses, and share count. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Automotive OEMs, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress unit deliveries and ASP together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test dealer inventory. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model wholesale units, retail sell-through, ASP, incentives, mix, dealer inventory, warranty, financial-services contribution, plant utilization, and capex. ### Leading-indicator dashboard Track registrations, dealer days supply, incentives, used-car values, order banks, production schedules, battery/material costs, fleet regulation, and finance delinquencies. ### Primary-source map SEC filings; NHTSA recalls and safety data; EPA fuel-economy/emissions rules; BEA motor-vehicle data; state registration data where public; captive-finance disclosures. ### Accounting normalization test Captive finance, pension, warranty reserves, lease residuals, incentives, inventory financing, and restructuring can move earnings materially. ### Valuation implementation Use mid-cycle earnings, DCF, SOTP for captive finance, and asset/liquidity stress. Treat peak pricing and low incentives as cyclical until proven structural. ### Worked numerical mini-case > Illustrative incentive case. Wholesale units rise 4% but average incentive per vehicle rises from $2,000 to $3,200 on a $45,000 ASP. The extra $1,200 incentive reduces revenue/margin by roughly $1.2bn per 1m vehicles before mix and cost actions. Separate true demand from channel stuffing, fleet mix and captive-finance support. ### Monitoring and falsification cadence Breaks include price war, residual-value collapse, platform transition failure, excess capacity, warranty/recall burden, or financing losses. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Automotive OEMs work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 079 # EV and Battery Manufacturers Analyst Playbook > Mission. Build a sector-specific research system for EV and Battery Manufacturers that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model capacity, qualified capacity, utilization, yield, cell chemistry, energy density, cost per kWh, scrap, customer qualification, subsidies, capex, and cash runway. Distinguish nameplate capacity from economically sellable qualified output. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | GWh shipped | Battery cell or pack energy content shipped during the period, in gigawatt-hours, based on nameplate energy per unit × units shipped. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | yield | FCF yield = normalized free cash flow to equity / current equity value | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | cell cost/kWh | Total cell manufacturing cost, materials + conversion + yield/scrap + allocated factory overhead, divided by sellable kWh produced. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | ASP/kWh | Battery revenue for the defined cell/pack product divided by kWh shipped, after rebates and mix effects. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | energy density | Usable energy capacity divided by cell/pack mass or volume, typically Wh/kg or Wh/L, measured on a consistent test basis. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | scrap | Production input or output rejected/scrapped divided by total material input or gross production; report by process stage if available. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | customer awards | Signed sourcing awards/design nominations with defined program, volume, timing, and customer; distinguish non-binding indications from awarded business. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | ## Sector-specific accounting and comparability traps - Inventory valuation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Government credits: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warranty: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized startup costs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Take-or-pay contracts: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/GWh capacity: enterprise value divided by effective sellable battery capacity, but only after adjusting utilization, yield, chemistry, customer qualification, capex, and margin per kWh. - EV/sales: enterprise value divided by normalized revenue; pair with explicit gross-margin, operating-cost, reinvestment, and capital-intensity assumptions. - Gross-profit bridge: value the business from normalized gross profit after reconciling revenue mix/pass-through effects, then deduct required opex, SBC, capex, working capital, and taxes. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for EV and Battery Manufacturers, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress GWh shipped and yield together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test inventory valuation. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model vehicle or cell units, kWh, ASP, chemistry mix, yield, utilization, material cost per kWh, warranty, credits, and capex by plant. ### Leading-indicator dashboard Track registrations, order lead times, cell pricing, lithium/nickel costs, plant ramp, yield, incentives, charging infrastructure, and competitor price changes. ### Primary-source map SEC filings; DOE and Argonne battery/EV publications; EPA vehicle data; EIA electricity/charging data; USGS mineral statistics; plant permits and incentive agreements. ### Accounting normalization test Government incentives, capitalized development, supplier prepayments, warranty, inventory, and rapid technology obsolescence can distort profitability. ### Valuation implementation Use long-horizon DCF with explicit financing/dilution, unit economics, and plant-level returns. Revenue multiples require a credible mature margin and capital-intensity bridge. ### Worked numerical mini-case > Illustrative cell-cost case. A 75 kWh pack uses cells costing $92/kWh, or $6,900 per vehicle. A decline to $78/kWh saves $1,050 per vehicle before pack overhead. At 500k vehicles the gross cost opportunity is $525m. Do not capitalize the full commodity benefit if price cuts, lower utilization or warranty costs absorb it. ### Monitoring and falsification cadence Breaks include demand elasticity, slower yield ramp, chemistry displacement, subsidy change, capital shortfall, or price declines outrunning cost reductions. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The EV and Battery Manufacturers work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 080 # BESS Electrical Balance-of-System Analyst Playbook > Mission. Build a sector-specific research system for BESS Electrical Balance-of-System that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model storage deployments, power-conversion content, switchgear/transformer availability, interconnection timing, gross margin by project, service attach, warranty reserves, working capital, and backlog conversion. Stress project delays and component bottlenecks. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | GW/GWh projects | Power rating in GW and energy capacity in GWh of storage projects contracted, under construction, or operating; always pair MW/GW with duration. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | PCS MW shipped | Power-conversion-system AC megawatts shipped or recognized in revenue during the period, separated by inverter/PCS class where relevant. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | book-to-bill | Book-to-bill = bookings / recognized revenue for the same definition and period | | gross margin | Gross margin = gross profit / revenue | | service attach | Projects or installed base covered by paid long-term service/software agreements divided by eligible installed projects/base. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | warranty rate | Warranty claims expense or reserve accrual divided by relevant product revenue or installed capacity, with cohort age separately monitored. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | grid-forming mix | Grid-forming-enabled PCS or project MW divided by total PCS/project MW shipped or awarded. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Project timing: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Revenue recognition: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warranty reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Customer concentration: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Working-capital swings: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - Normalized EV/sales: enterprise value divided by revenue normalized for commodity pass-through, project timing, acquisitions, and unusual mix, with margins/capital intensity explicitly bridged. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for BESS Electrical Balance-of-System, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress GW/GWh projects and PCS MW shipped together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test project timing. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model MW/MWh deployed, inverter or PCS content, switchgear/transformer content, engineering/service attach, project timing, price per kW, warranty, and backlog conversion. ### Leading-indicator dashboard Track interconnection queues, battery project awards, transformer and switchgear lead times, PCS orders, grid-code changes, utility procurement, fire/safety standards, and storage economics. ### Primary-source map SEC filings; FERC and ISO/RTO interconnection/market rules; EIA storage deployment data; utility procurement dockets; UL/NFPA/IEEE standards where applicable; EPC and inverter supplier disclosures. ### Accounting normalization test Project milestones, customer advances, warranties, long lead procurement, supplier concentration, and percentage-of-completion can distort revenue and cash timing. ### Valuation implementation Use DCF, EV/EBIT, and backlog-adjusted scenarios with explicit normalized margins after supply scarcity. Separate equipment from recurring service/software value. ### Worked numerical mini-case > Illustrative project economics. A 200 MW / 800 MWh project carries $95/kW of inverter/EBoS content, implying $19m addressable revenue. At 28% gross margin, project gross profit is $5.3m before service. Delay the COD six months and model revenue timing, working capital and liquidated-damages exposure. The thesis should distinguish equipment content growth from commoditization, interconnection delays and warranty/service obligations. ### Monitoring and falsification cadence Breaks include vertical integration by battery suppliers, standardization commoditizing PCS, transformer bottleneck shifting value elsewhere, project delays, or warranty failures. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The BESS Electrical Balance-of-System work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 081 # Grid Equipment and Electrification Analyst Playbook > Mission. Build a sector-specific research system for Grid Equipment and Electrification that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Track orders, backlog, book-to-bill, factory capacity, lead times, copper/electrical steel, price-cost lag, utility capex cycles, service mix, and capacity expansion. Test whether backlog reflects genuine demand or extended delivery times. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | orders | Gross value of customer purchase orders or bookings received during the period, adjusted for cancellations and scope changes where disclosed. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | lead times | Time from order to shipment/commissioning for key equipment or projects, measured by product category and compared with normal and peak-cycle levels. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | capacity additions | Incremental nameplate/effective production or infrastructure capacity entering service during the period, net of retirements. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | price-cost | Change in realized selling price minus change in relevant input and conversion cost on a comparable-unit basis; express in dollars and margin points. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | service revenue | Revenue from maintenance, monitoring, software, spare parts, and lifecycle services; report as % of total and per installed-unit/MW base. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | factory utilization | Actual factory output hours/units divided by practical available production capacity after downtime and yield losses. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | utility capex | Utility capital expenditures for generation, transmission, distribution, and other rate-base assets, reconciled to cash flow and regulatory plans. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | ## Sector-specific accounting and comparability traps - Long-cycle contracts: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Project reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capacity capex: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Grid Equipment and Electrification, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress orders and backlog together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test long-cycle contracts. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model units or capacity for transformers, switchgear, breakers, conductors, protection, power electronics, and service. Tie demand to utility capex, load growth, interconnection, replacement, and data-center/industrial projects. ### Leading-indicator dashboard Track utility rate-base plans, interconnection queues, lead times, factory expansions, copper/electrical steel prices, order backlog, cancellations, and manufacturing capacity. ### Primary-source map SEC filings; DOE Grid Deployment Office; FERC and ISO/RTO planning dockets; utility integrated-resource and transmission plans; transformer/switchgear supplier filings; Census construction data. ### Accounting normalization test Long-cycle backlog, price-escalation clauses, advance payments, pension, project accounting, and capacity-expansion capex require normalization. ### Valuation implementation Use DCF, EV/EBIT, and FCF with margin normalization as scarcity eases. Evaluate replacement-cost and capacity economics for new entrants. ### Worked numerical mini-case > Illustrative backlog conversion. Opening backlog is $4.0bn, new orders $3.2bn and revenue $2.6bn. Ending backlog = $4.6bn before cancellations/FX. If lead times fall while book-to-bill drops below 1.0, do not treat backlog growth as proof of accelerating demand. Model transformer/switchgear capacity, price realization and working-capital needs explicitly. ### Monitoring and falsification cadence Breaks include capacity response collapsing price, regulatory delays, utility affordability constraints, project cancellations, or a change in grid architecture reducing equipment content. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Grid Equipment and Electrification work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 082 # Electric Utilities Analyst Playbook > Mission. Build a sector-specific research system for Electric Utilities that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model rate base, allowed ROE, equity layer, capex, depreciation, regulatory lag, load growth, fuel/purchased power recovery, storm costs, financing, and customer affordability. Value creation depends on earning allowed returns on prudent investment. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | rate base | Net regulatory investment on which the utility is permitted to earn a return, including eligible plant and working-capital components less accumulated depreciation and deferred items. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | allowed ROE | Regulator-authorized return on common equity embedded in customer rates for the relevant jurisdiction and rate case. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | load growth | load growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | capex | Maintenance capex is the spending required to preserve current earning power, estimated from asset replacement and operating evidence rather than management labels alone | | FERC/state mix | Share of earnings, rate base, or revenue governed by FERC versus state commissions, measured on a consistent economic base. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | earned ROE | Regulated net income attributable to common equity divided by average common equity supporting the regulated business. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | customer count | Average or period-end customer accounts served, segmented by class and adjusted for acquisitions/dispositions. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | debt cost | Interest expense, net of capitalized interest where appropriate, divided by average interest-bearing debt, or weighted coupon/yield on debt outstanding. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | ## Sector-specific accounting and comparability traps - Regulatory assets: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Storm costs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Decommissioning: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized afudc: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - P/B versus ROE: price-to-book or price-to-tangible-book interpreted against sustainable ROE/ROTCE, growth, payout, risk, and cost of equity rather than as a standalone multiple. - Dividend yield: annualized sustainable common dividend divided by share price; test payout coverage, regulatory/capital constraints, cyclicality, and reinvestment needs. - Rate-base DCF: project rate-base growth, allowed and earned returns, capital structure, depreciation, taxes, customer bill impact, and financing needs under regulatory lag. ## Sector diligence questions - What is the most important leading indicator for Electric Utilities, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress rate base and allowed ROE together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test regulatory assets. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model rate base, allowed ROE, load growth, generation mix, fuel, purchased power, capex, depreciation, financing, and customer rates by jurisdiction. ### Leading-indicator dashboard Track rate cases, regulatory orders, load forecasts, weather, data-center connections, fuel spreads, capex plans, credit metrics, and political affordability pressure. ### Primary-source map SEC and FERC filings; state public-utility commission rate cases; integrated-resource plans; EIA generation/fuel data; ISO/RTO market data; debt and rating-agency disclosures. ### Accounting normalization test Regulatory assets/liabilities, securitization, storm costs, pension, decommissioning, and capitalized AFUDC can make GAAP earnings differ from cash. ### Valuation implementation Use regulated-asset and dividend/FCFE frameworks, P/E relative to growth and allowed returns, and credit-sensitive DCF. Capital needs and dilution are central. ### Worked numerical mini-case > Illustrative rate-base case. Beginning rate base $10bn, capex $1.5bn, depreciation $0.7bn and retirements $0.1bn imply ending rate base near $10.7bn before adjustments. At a 9.5% allowed ROE and 52% equity layer, estimate incremental earnings only after regulatory lag. Growth capex creates value only if prudently included in rate base and earned above the financing cost. ### Monitoring and falsification cadence Breaks include adverse regulatory outcomes, cost overruns, wildfire/nuclear liabilities, financing pressure, load forecasts that fail to materialize, or customer affordability backlash. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Electric Utilities work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 083 # Renewable Developers Analyst Playbook > Mission. Build a sector-specific research system for Renewable Developers that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model MW pipeline by stage, interconnection, permits, equipment, PPA pricing, capacity factors, tax credits, financing, construction cost, COD timing, asset recycling, and counterparty risk. Probability-weight pipeline instead of treating all announced MW equally. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | MW operating | Nameplate or net renewable generation capacity in commercial operation, with ownership percentage and technology clearly stated. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | MW backlog | Renewable/project megawatts under signed contracts or advanced development expected to enter construction/operation, net of cancellations. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | capacity factor | Actual electricity generated divided by maximum possible generation at nameplate capacity over the same period. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | PPA price | Contracted electricity revenue per MWh under power purchase agreements, including escalators and market/REC components as defined. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | project IRR | Discount rate that sets project-level unlevered or levered cash-flow NPV to zero; state tax credits, leverage, terminal value, and ownership assumptions. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | tax-credit value | Expected cash or present value of transferable/usable tax credits attributable to a project, net of discount, fees, and tax capacity constraints. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | interconnection queue | Project MW with active grid-interconnection requests by stage; track MW advanced to study/agreement/energization rather than gross queue alone. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | capex/MW | Total project capital expenditure divided by installed or commissioned MW, with storage duration, technology, and owner-supplied equipment consistently treated. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | ## Sector-specific accounting and comparability traps - Tax equity: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Project finance: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Development gains: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Impairments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Contract liabilities: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. - Project NPV: discount project-level after-tax cash flows using project-specific construction, operating, financing, tax-credit, terminal/decommissioning, and delay assumptions. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Renewable Developers, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress MW operating and MW backlog together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test tax equity. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model project MW, capacity factor, PPA price, merchant exposure, construction cost, tax credits, financing, curtailment, degradation, and operating expense. ### Leading-indicator dashboard Track interconnection, queue progress, equipment prices, interest rates, tax-credit transfer pricing, PPAs, permitting, transmission, and project-sale markets. ### Primary-source map SEC filings; FERC/ISO interconnection queues; EIA project and generation data; utility/commission procurement dockets; tax-credit guidance; project permits and offtake disclosures. ### Accounting normalization test Tax equity, development gains, project sales, nonrecourse debt, capitalized interest, and unconsolidated JVs complicate reported earnings. ### Valuation implementation Use project NAV/DCF and corporate SOTP. Stress discount rates, merchant tails, curtailment, construction cost, and financing availability. ### Worked numerical mini-case > Illustrative project case. A 200 MW project at $1.3m/MW costs $260m. At 35% capacity factor and $45/MWh realized price, gross annual energy revenue is about $27.6m before credits, congestion, curtailment and O&M. Calculate project IRR/NPV with tax credits, financing and degradation rather than applying a corporate EBITDA multiple to pipeline MW. ### Monitoring and falsification cadence Breaks include interconnection failure, cost inflation, financing spread, PPA repricing, policy change, or lower capacity factor/greater curtailment. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Renewable Developers work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 084 # Oil and Gas E&P Analyst Playbook > Mission. Build a sector-specific research system for Oil and Gas E&P that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model production by commodity, decline curves, realized price, basis/hedges, LOE, gathering, maintenance versus growth capex, inventory depth, breakevens, royalty/tax, and balance sheet. Use strip and multiple commodity scenarios rather than one price deck. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | production | Oil, gas, NGL, mineral, or other physical output produced during the period in standardized units, net or gross according to stated ownership convention. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | oil/gas/NGL mix | Percentage of total production or revenue represented by oil, natural gas, and NGLs using consistent energy or volume units. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | realized price | Revenue for the commodity/product divided by sales volume after quality/location differentials and before or after hedges as explicitly stated. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | hedges | Volume, price, tenor, and fair-value/cash-settlement profile of derivative positions relative to forecast production or commodity exposure. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | LOE | Lease operating expense divided by production volume, typically $/boe, excluding or separately stating production taxes and transportation. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capex | Maintenance capex is the spending required to preserve current earning power, estimated from asset replacement and operating evidence rather than management labels alone | | inventory locations | Count and working-interest share of economic drilling locations meeting defined return/cost thresholds at stated commodity prices. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | decline rate | Percentage reduction in production from an existing well/cohort over a defined period, separated into initial and terminal decline assumptions. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Successful-efforts/full-cost: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Dd&a: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Asset retirement obligations: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Hedge accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Reserve revisions: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDAX: enterprise value divided by EBITDA before exploration expense; normalize commodity prices, hedges, reserve replacement, decline, and sustaining drilling capital. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - NAV: mark identifiable assets and liabilities to economic value, subtract debt and other claims, and divide residual value by diluted shares; document commodity/discount-rate assumptions. - PDP/PUD valuation: DCF proved developed producing reserves separately from undeveloped inventory using decline curves, type curves, commodity prices, costs, taxes, timing, and risk haircuts. ## Sector diligence questions - What is the most important leading indicator for Oil and Gas E&P, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress production and oil/gas/NGL mix together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test successful-efforts/full-cost. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model production by commodity, decline curves, realized price, differentials, hedges, lifting cost, royalties, drilling/completion cost, inventory depth, and maintenance capital. ### Leading-indicator dashboard Track rigs, frac spreads, permits, basin takeaway, storage, commodity curves, service costs, well productivity, decline, and operator capex. ### Primary-source map SEC filings and reserve disclosures; EIA production/inventory/price data; state oil-and-gas regulators; Baker Hughes rig counts; midstream constraints; royalty/acreage records where public. ### Accounting normalization test Reserve revisions, successful-efforts/full-cost accounting, impairments, derivative marks, asset retirement obligations, and acquisition adjustments matter. ### Valuation implementation Use NAV by acreage/project, mid-cycle FCF, EV/EBITDA only with maintenance-capex context, and commodity sensitivities. ### Worked numerical mini-case > Illustrative well economics. A well costs $9m and produces 650 mboe over its economic life. At $52/boe realized net price and $17/boe cash operating/tax cost, undiscounted field margin is roughly $22.75m before decline timing and corporate costs. Run price, EUR, basis differential and service-cost sensitivities and reconcile reserve replacement to sustaining capital. ### Monitoring and falsification cadence Breaks include lower resource productivity, cost inflation, basis widening, regulatory limits, weak balance sheet, or inventory exhaustion. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Oil and Gas E&P work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 085 # Midstream Energy Analyst Playbook > Mission. Build a sector-specific research system for Midstream Energy that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model throughput, contract type, MVCs, tariff escalators, commodity-sensitive exposure, maintenance/growth capex, leverage, distribution coverage, project backlog, and counterparty quality. Separate EBITDA stability from refinancing and volume risk. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | throughput | Physical volume processed, transported, refined, or handled through the asset during the period in standardized units. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | capacity | Practical or nameplate maximum output/throughput available over the period, adjusted for maintenance and operating constraints when relevant. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | contract type | Share of revenue/EBITDA governed by take-or-pay, fee-based, commodity-sensitive, cost-plus, fixed-price, or other contract structures. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | MVC coverage | Actual or forecast customer volume divided by minimum volume commitment; also quantify deficiency payments when volume falls short. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | tariff | Contracted or regulated fee per unit of throughput/transportation, including escalators and fuel/power surcharges as applicable. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | distribution coverage | Distributable cash flow divided by cash distributions to common unitholders/shareholders for the same period. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | leverage | Net leverage = net debt / normalized EBITDA, with leases and other debt-like items treated consistently | | project backlog | Value/capacity of sanctioned or contracted projects not yet recognized in revenue/EBITDA, with expected in-service dates and remaining capex. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | ## Sector-specific accounting and comparability traps - Joint ventures: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Non-controlling interests: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Maintenance capex: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Commodity sensitivity: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Contract liabilities: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. - Distribution yield: annualized cash distribution divided by equity/unit price; test distributable-cash-flow coverage, leverage, capex, contract durability, and sponsor incentives. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Midstream Energy, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress throughput and capacity together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test joint ventures. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model volumes, contracted capacity, tariff, commodity sensitivity, minimum commitments, expansion capex, maintenance capex, and counterparty credit. ### Leading-indicator dashboard Track producer activity, basin differentials, contract renewals, pipeline utilization, project approvals, regulatory rulings, and customer leverage. ### Primary-source map SEC filings; FERC pipeline tariffs and dockets; EIA production/flow data; shipper/producer filings; contract and minimum-volume disclosures; debt/covenant documents. ### Accounting normalization test MLP or partnership structures, noncontrolling interests, equity-method JVs, maintenance-capex definitions, and distributable cash flow adjustments require reconciliation. ### Valuation implementation Use DCF/distributable cash flow, EV/EBITDA with contract quality, and project returns. Include debt and distribution coverage. ### Worked numerical mini-case > Illustrative contract case. A pipeline transports 1.0 Bcf/day at $0.65/Mcf, or about $237m annual gross revenue at full contracted volume. If 75% is minimum-volume protected, stress the uncovered 25% first and evaluate counterparty credit. Separate commodity exposure from volume, contract duration, inflation escalators and recontracting risk. ### Monitoring and falsification cadence Breaks include contract roll-off at lower rates, producer distress, regulatory blockage, overbuild, or maintenance needs higher than reported. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Midstream Energy work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 086 # Refiners Analyst Playbook > Mission. Build a sector-specific research system for Refiners that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model throughput, utilization, product yields, crack spreads, crude differentials, RINs/carbon costs, turnaround expense, working capital, and sustaining capex. Normalize through-cycle margins and recognize inventory accounting effects. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | throughput | Physical volume processed, transported, refined, or handled through the asset during the period in standardized units. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | crack spread | Market value of refined products produced from a barrel of crude minus crude feedstock cost, using a stated product slate such as 3-2-1. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capture rate | Realized refining margin divided by benchmark crack spread for the same region and period. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | RIN cost | Net Renewable Identification Number compliance cost divided by refined gallons/barrels or reported as total expense net of internally generated credits. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | turnaround expense | Maintenance turnaround cash/expense incurred during planned refinery outages, shown by facility and period and separated from normal maintenance. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | inventory | Inventory days = average inventory / COGS x days in period | | renewable diesel economics | Realized product price + credits/incentives - feedstock - conversion - logistics and variable costs per gallon/barrel. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Lifo/fifo: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Inventory gains: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Hedges: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Turnaround accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Environmental credits: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Mid-cycle EV/EBITDA: enterprise value divided by through-cycle EBITDA after normalizing commodity prices, utilization, spreads, temporary outages, and one-time costs. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - Replacement value: estimate the cost and time to recreate the asset network/capacity today, net of obsolescence and required upgrades, then compare with enterprise value. ## Sector diligence questions - What is the most important leading indicator for Refiners, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress throughput and utilization together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test LIFO/FIFO. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model throughput, utilization, crude differentials, crack spreads, product yields, RIN/renewable obligations, turnaround, and working capital. ### Leading-indicator dashboard Track crack spreads, inventories, utilization, outages, crude differentials, product demand, refinery closures/additions, and regulatory credit prices. ### Primary-source map SEC filings; EIA refinery utilization, inventories, crack-spread inputs and product demand; EPA renewable-fuel rules; regional product flows; turnaround disclosures. ### Accounting normalization test Inventory accounting, turnaround capitalization, environmental obligations, and working-capital swings create large earnings/cash timing differences. ### Valuation implementation Use mid-cycle earnings/FCF and asset replacement economics. Peak cracks should not be capitalized as permanent. ### Worked numerical mini-case > Illustrative margin case. Throughput is 200 kbpd and capture-adjusted refining margin is $12/bbl. Annual gross refining margin is roughly $876m before opex and turnarounds. A $3/bbl margin compression reduces annual gross margin about $219m. Use regional crack spreads, capture rate, utilization and RIN/turnaround assumptions rather than spot headline cracks alone. ### Monitoring and falsification cadence Breaks include structural demand decline, new low-cost capacity, feedstock disadvantage, environmental capex, or prolonged utilization weakness. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Refiners work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 087 # Chemicals Analyst Playbook > Mission. Build a sector-specific research system for Chemicals that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model volume, price, mix, feedstocks, energy, utilization, capacity additions, inventory, regional spreads, turnarounds, and environmental liabilities. Separate price-cost timing from true structural margin improvement. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | volume | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | price | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | mix | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | feedstock cost | Weighted-average cost of biomass/oil/other feedstock consumed per gallon/barrel/ton of finished product. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | spreads | Selling price or product benchmark minus key raw-material/input benchmark on a matched unit and period basis. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capacity additions | Incremental nameplate/effective production or infrastructure capacity entering service during the period, net of retirements. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | regional mix | Revenue, volume, gross profit, or capacity by geography divided by consolidated total using a consistent denominator. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Environmental liabilities: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Jv accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Restructuring: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Mid-cycle EV/EBITDA: enterprise value divided by through-cycle EBITDA after normalizing commodity prices, utilization, spreads, temporary outages, and one-time costs. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Chemicals, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress volume and price together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test inventory. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model volume, price, mix, feedstock, energy, utilization, capacity additions, maintenance outages, and product-specific spreads. ### Leading-indicator dashboard Track operating rates, inventories, feedstock spreads, China/global capacity, freight, end-market production, and plant outages. ### Primary-source map SEC filings; EIA and commodity-feedstock data; Federal Reserve industrial production; Census end-market data; peer capacity additions; environmental/regulatory permits and disclosures. ### Accounting normalization test Inventory, pension, environmental liabilities, JV accounting, restructuring, and maintenance turnaround timing require normalization. ### Valuation implementation Use mid-cycle EV/EBITDA, DCF, FCF yield, and replacement-cost context. Normalize for cycle and feedstock advantage. ### Worked numerical mini-case > Illustrative spread case. A plant sells 1.0m tonnes at $1,100/tonne with variable feedstock/energy of $700/tonne, creating $400m contribution before fixed costs. A $120/tonne spread compression cuts contribution by $120m at constant volume. Overlay utilization, new capacity, contract lags and inventory before calling the move structural. ### Monitoring and falsification cadence Breaks include sustained global overcapacity, feedstock disadvantage, regulation, substitution, or structurally lower demand in key end markets. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Chemicals work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 088 # Mining and Metals Analyst Playbook > Mission. Build a sector-specific research system for Mining and Metals that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model production, grade, recovery, realized price, by-product credits, cash cost/AISC, strip ratio, sustaining/growth capex, reserve life, permits, royalties, and closure liabilities. Build mine-level NAV where project economics differ materially. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | production | Oil, gas, NGL, mineral, or other physical output produced during the period in standardized units, net or gross according to stated ownership convention. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | grade | Average ore/metal concentration in mined or processed material, such as g/t or %, measured on a consistent reserve/production basis. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | recovery | Payable metal/mineral output recovered divided by contained metal/mineral in processed ore/feed. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | realized price | Revenue for the commodity/product divided by sales volume after quality/location differentials and before or after hedges as explicitly stated. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | cash cost | Direct mine/site operating cash costs net of applicable by-product credits divided by payable production units, using the issuer's consistent definition. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | AISC | All-in sustaining cost: cash operating cost plus sustaining capex and sustaining corporate/exploration costs, net of defined by-product credits, divided by payable production. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | reserve life | Proven and probable reserves divided by normalized annual production, adjusted for mine sequencing and recoveries where appropriate. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | capex | Maintenance capex is the spending required to preserve current earning power, estimated from asset replacement and operating evidence rather than management labels alone | ## Sector-specific accounting and comparability traps - Stripping costs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Asset retirement obligations: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Reserve assumptions: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Impairment: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Streaming agreements: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - NAV: mark identifiable assets and liabilities to economic value, subtract debt and other claims, and divide residual value by diluted shares; document commodity/discount-rate assumptions. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/NAV: common equity market value divided by independently estimated net asset value; stress commodity prices, reserve quality, development capex, discounts, taxes, and ownership. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. ## Sector diligence questions - What is the most important leading indicator for Mining and Metals, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress production and grade together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test stripping costs. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model production, grade, recovery, realized price, treatment charges, cash cost, sustaining capex, growth capex, reserves, and mine life. ### Leading-indicator dashboard Track benchmark prices, inventories, treatment charges, production guidance, grade, permitting, labor, power, freight, and new project supply. ### Primary-source map SEC filings and technical/reserve reports; USGS mineral statistics; mine regulator/permit data; benchmark commodity prices; smelter/refiner disclosures; jurisdiction royalty/tax rules. ### Accounting normalization test Stripping, reserve estimates, rehabilitation liabilities, JVs, royalties, impairment, and exploration capitalization can distort comparisons. ### Valuation implementation Use project NAV, commodity sensitivities, mid-cycle FCF, and asset quality. Corporate overhead and future development capital matter. ### Worked numerical mini-case > Illustrative mine case. Production 500koz, realized gold $2,200/oz and AISC $1,350/oz imply $425m pre-tax mine margin before sustaining/growth distinctions and corporate costs. A $200/oz price change moves mine margin about $100m. Reconcile grade, recovery, strip ratio, reserve life and sustaining capital. Do not value ounces without extraction economics. ### Monitoring and falsification cadence Breaks include reserve/grade disappointment, project overrun, jurisdiction change, lower commodity incentive price, or capital needs exceeding balance-sheet capacity. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Mining and Metals work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 089 # Banks Analyst Playbook > Mission. Build a sector-specific research system for Banks that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model loans, deposits, deposit beta/mix, asset yields, NIM, fees, expenses, charge-offs, reserves, CET1, AOCI, securities duration, and liquidity. Run parallel credit, rate, deposit-flight, and capital stresses. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | NIM | Annualized net interest income divided by average interest-earning assets. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | loan growth | loan growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | deposit beta | Cumulative change in deposit rate divided by cumulative change in the reference policy/market rate over the same tightening or easing cycle. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | deposit mix | Noninterest-bearing, interest-bearing checking, savings, money-market, time deposits, and brokered deposits divided by total deposits. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | charge-offs | Net charge-offs divided by average loans for the period, annualized when reporting quarterly. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | NPLs | Nonperforming loans divided by total loans, with 90+ days past due and nonaccrual definitions reconciled to bank disclosures. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | CET1 | Common Equity Tier 1 regulatory capital divided by risk-weighted assets under the applicable regulatory framework. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | tangible book value | Common shareholders' equity less goodwill and identifiable intangible assets, divided by common shares for TBV/share. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Cecl allowance: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Afs/htm marks: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Nonaccrual loans: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Securities duration: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Off-balance-sheet commitments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/TBV versus ROTCE: price-to-tangible-book interpreted against sustainable return on tangible common equity, growth, payout, credit cost, and cost of equity. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - Dividend discount: present value of sustainable common dividends or distributable capital, constrained by regulatory capital, growth funding, payout policy, and cost of equity. - Excess-capital analysis: value deployable capital above operating/regulatory requirements separately from the earnings franchise, net of tax and deployment constraints. ## Sector diligence questions - What is the most important leading indicator for Banks, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress NIM and loan growth together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test CECL allowance. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model average loans/securities, yields, deposits, deposit beta, funding mix, NIM, fees, expenses, charge-offs, provisions, capital, and buybacks. ### Leading-indicator dashboard Track deposit flows and pricing, loan growth, credit delinquencies, charge-offs, securities marks, yield curve, funding markets, capital ratios, and regulatory actions. ### Primary-source map SEC filings and call reports; FDIC BankFind/Quarterly Banking Profile; Federal Reserve Y-9C and H.8 data; FFIEC data; yield-curve/rate data; regulatory capital and stress-test disclosures. ### Accounting normalization test CECL/reserve assumptions, AOCI, held-to-maturity marks, nonaccruals, loan modifications, and capital treatment are central. ### Valuation implementation Use P/TBV, P/E, residual income, and excess-capital approaches tied to normalized ROE and cost of equity. ### Worked numerical mini-case > Illustrative NIM and credit case. Average earning assets $50bn and NIM 3.20% imply $1.60bn annual net interest income. A 20 bp NIM decline costs about $100m pre-tax. Separately, a 40 bp increase in net charge-offs on $35bn loans costs about $140m. Bridge NII, fees, provisions, expenses and capital together, then value sustainable ROTCE relative to cost of equity and tangible book. ### Monitoring and falsification cadence Breaks include deposit franchise weakening, credit losses above reserve, capital shortfall, funding stress, regulatory constraint, or asset-liability mismatch. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Banks work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 090 # Property and Casualty Insurance Analyst Playbook > Mission. Build a sector-specific research system for Property and Casualty Insurance that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model premiums, rate, exposure units, retention, loss ratio, catastrophe load, reserve development, expense ratio, reinsurance, investment income, and capital. Separate accident-year profitability from reserve releases. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | premium growth | premium growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | rate | Cap rate = stabilized NOI / property value | | retention | NRR = beginning-cohort recurring revenue after churn, contraction, and expansion / beginning-cohort recurring revenue | | loss ratio | Loss ratio = incurred losses and loss-adjustment expense / earned premium | | expense ratio | Insurance underwriting expenses divided by net premiums earned; for funds/asset products, operating expenses divided by average assets as context requires. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | combined ratio | Combined ratio = loss ratio + expense ratio | | reserve development | Change in prior accident-year loss reserves recognized in the current period, expressed in dollars and as a % of opening reserves/premiums. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | investment yield | investment yield = annualized economic output / current market value or invested base; match numerator and denominator. | ## Sector-specific accounting and comparability traps - Loss reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Catastrophe estimates: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Reinsurance: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Dac: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Fair-value portfolio marks: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/B versus ROE: price-to-book or price-to-tangible-book interpreted against sustainable ROE/ROTCE, growth, payout, risk, and cost of equity rather than as a standalone multiple. - Normalized P/E: equity value divided by through-cycle diluted EPS after normalizing credit, reserves, margins, taxes, unusual gains/losses, and share count. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Property and Casualty Insurance, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress premium growth and rate together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test loss reserves. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model premiums, exposure units, rate, retention, loss ratio, expense ratio, catastrophe load, reserve development, reinsurance, float, and investment income. ### Leading-indicator dashboard Track rate filings, claims inflation, catastrophe events, renewal retention, reinsurance pricing, reserve development, and bond yields. ### Primary-source map SEC filings; statutory statements and NAIC data; state insurance-department rate filings; catastrophe-model/event data from public agencies; reinsurance disclosures; bond portfolio and rate data. ### Accounting normalization test Loss reserves, reinsurance recoverables, catastrophe accounting, prior-year development, and investment marks dominate quality. ### Valuation implementation Use P/B or P/TBV relative to normalized ROE, underwriting-cycle earnings, and excess capital. ### Worked numerical mini-case > Illustrative combined-ratio case. Net earned premium is $5.0bn. A 94% combined ratio produces $300m underwriting profit. If catastrophe and reserve development push the ratio to 100%, underwriting profit falls to zero before investment income. Separate accident-year loss ratio from prior-year reserve releases, then evaluate rate versus loss-cost trend and capital/reinsurance capacity. ### Monitoring and falsification cadence Breaks include adverse reserve development, social inflation outrunning pricing, reinsurance unavailable, catastrophe concentration, or capital erosion. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Property and Casualty Insurance work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 091 # Life Insurance Analyst Playbook > Mission. Build a sector-specific research system for Life Insurance that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model account values, spreads, mortality/morbidity, lapses, guarantees, hedging, RBC/capital, investment portfolio, new business economics, and reinsurance. Stress asset-liability duration and policyholder behavior together. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | sales | New policy premiums, deposits, annuity sales, or product sales under the issuer's definition; reconcile gross versus net and funded versus submitted amounts. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | spread income | Investment yield earned on spread-based assets minus credited rate/funding cost, multiplied by average spread-based liabilities/assets as appropriate. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | policyholder behavior | Observed lapse, withdrawal, premium-payment, utilization, or conversion behavior by product/cohort versus priced assumptions. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | surrenders | Policy/contract value surrendered during the period divided by average account value or policies in force. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | RBC | Risk-based capital ratio: total adjusted capital divided by company action level RBC, or the jurisdiction-specific equivalent. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | investment yield | investment yield = annualized economic output / current market value or invested base; match numerator and denominator. | | hedging | Net derivative notional, sensitivities, collateral, and realized/unrealized hedge results relative to the underlying insurance/market risk exposure. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capital return | Common dividends plus net share repurchases divided by net income/FCF, or absolute dollars returned, with regulatory capital constraints shown separately. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Actuarial assumptions: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Dac: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Separate accounts: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Derivatives: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Statutory capital: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/B: common equity value divided by common book value; interpret with asset quality, mark-to-market exposure, sustainable ROE, growth, and cost of equity. - Normalized P/E: equity value divided by through-cycle diluted EPS after normalizing credit, reserves, margins, taxes, unusual gains/losses, and share count. - Embedded-value context: value in-force insurance business as adjusted net worth plus present value of future distributable profits, with lapse, mortality, spread, capital, and discount assumptions. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Life Insurance, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress sales and spread income together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test actuarial assumptions. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model policies or account value, premiums, spreads, lapses, mortality/morbidity, hedging, statutory capital, and investment yield. ### Leading-indicator dashboard Track lapse rates, credit spreads, mortality, annuity sales, new-money yields, hedge effectiveness, regulatory capital, and reinsurance. ### Primary-source map SEC filings; statutory/NAIC statements; state insurance regulation; Federal Reserve rate data; mortality/morbidity public datasets; separate-account and hedging disclosures. ### Accounting normalization test Actuarial assumptions, DAC, market-risk benefits, reinsurance, statutory versus GAAP capital, and investment impairments require specialist treatment. ### Valuation implementation Use book-value/ROE, distributable capital, embedded-value concepts where appropriate, and earnings normalized for assumption updates. ### Worked numerical mini-case > Illustrative spread case. A $40bn general-account portfolio earning 4.8% against credited/benefit cost of 3.6% has a 120 bp gross spread, or $480m before expenses/hedging. A 30 bp spread compression costs about $120m. Model lapse behavior, duration mismatch, statutory capital and variable-annuity guarantees rather than relying on GAAP EPS alone. ### Monitoring and falsification cadence Breaks include lapse shock, asset-liability mismatch, reserve strengthening, hedge failure, credit losses, or capital/regulatory constraint. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Life Insurance work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 092 # Asset Managers and Brokers Analyst Playbook > Mission. Build a sector-specific research system for Asset Managers and Brokers that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Bridge AUM by market performance, net flows, acquisitions, and FX; then model fee rate, mix, performance fees, compensation, client concentration, capital return, and seed/principal investments. Separate market beta from organic franchise growth. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | AUM | AUM bridge = beginning AUM + market performance + net flows + acquisitions/divestitures + FX/other | | net flows | Client contributions/subscriptions minus withdrawals/redemptions over the period, excluding market appreciation and acquisition effects. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | fee rate | Asset-management or advisory fee revenue divided by average fee-bearing AUM/assets, annualized and adjusted for performance fees when separately disclosed. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | market beta | Change in AUM or revenue attributable to market appreciation/depreciation, estimated from asset-class exposures and market returns rather than net flows. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | performance fees | Performance/incentive fee revenue earned under fund mandates, shown as dollars and as a % of relevant AUM and total revenue. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | advisor count | Average or period-end productive advisors/representatives under a consistent definition, separating employees from independent/contracted channels. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | client assets | Market value of client assets under management, administration, or advisement using the issuer's stated scope; separate fee-bearing assets. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capital return | Common dividends plus net share repurchases divided by net income/FCF, or absolute dollars returned, with regulatory capital constraints shown separately. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Seed investments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Consolidation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Compensation accruals: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Fair-value marks: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Principal investments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - % of AUM: enterprise/equity value divided by AUM, interpreted through fee rate, product mix, net flows, margins, performance fees, capital intensity, and persistence. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Asset Managers and Brokers, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress AUM and net flows together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test seed investments. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model AUM by asset class, market beta, net flows, fee rate, performance fees, advisor/broker activity, compensation, and capital return. ### Leading-indicator dashboard Track fund performance, net flows, market levels, fee compression, advisor recruiting, trading/underwriting volumes, and client cash balances. ### Primary-source map SEC filings; Form ADV where relevant; fund filings; Federal Reserve/market data; exchange and FINRA statistics; custody/clearing disclosures; peer AUM and flow reports. ### Accounting normalization test Seed investments, consolidation, fair-value marks, principal investments, performance-fee timing, and compensation accruals affect comparability. ### Valuation implementation Use P/E, FCF, SOTP, and percent-of-AUM context. Separate market appreciation from organic flows and fee-rate mix. ### Worked numerical mini-case > Illustrative fee-rate case. Average AUM $200bn at a 42 bp management fee produces $840m base fees. A 10% market decline with zero net flows reduces annualized fees roughly $84m before performance fees and mix. Decompose AUM change into market, flows, FX and acquisitions, then model compensation and operating leverage. ### Monitoring and falsification cadence Breaks include persistent outflows, fee compression, poor performance, advisor attrition, regulatory changes, or capital trapped in low-return businesses. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Asset Managers and Brokers work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 093 # Payments and Fintech Analyst Playbook > Mission. Build a sector-specific research system for Payments and Fintech that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model payment volume, transactions, take rate, interchange/network fees, value-added services, loss/fraud, incentives, customer acquisition, funding cost, and regulatory constraints. Reconcile gross versus net revenue presentation. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | TPV | Total payment volume processed through the platform during the period, net or gross of refunds according to the disclosed definition. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | transactions | Count of successfully processed payment/commerce transactions during the period, with duplicated/failed activity excluded. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | take rate | Net revenue attributable to transaction volume divided by TPV/GMV or other monetized volume, after pass-through items as defined. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | active accounts | Unique accounts meeting the issuer's activity threshold during the defined trailing period; reconcile definition changes and duplicates. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | loss rate | Credit or fraud losses divided by relevant payment volume, receivables, or originated balance over the same cohort/period. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | funding cost | Interest and financing expense divided by average interest-bearing funding balances, including securitization/warehouse costs when applicable. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | gross profit | Revenue less directly attributable cost of revenue/COGS under the company's accounting definition; reconcile major pass-through and depreciation classifications. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | CAC | CAC payback months = customer acquisition cost / monthly gross profit from new customer | ## Sector-specific accounting and comparability traps - Gross versus net revenue: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Credit reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Held-for-sale loans: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Securitization: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Stock compensation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/gross profit: enterprise value divided by normalized gross profit; useful when revenue recognition/pass-through differs, but still requires opex and capital-intensity normalization. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Payments and Fintech, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress TPV and transactions together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test gross versus net revenue. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model payment volume, transactions, active users/merchants, take rate, value-added services, credit losses if applicable, incentives, and processing cost. ### Leading-indicator dashboard Track consumer spend, cross-border travel, merchant adds, payment volume, interchange/regulation, fraud losses, funding costs, and network tokenization. ### Primary-source map SEC filings; Federal Reserve payments data; CFPB releases; network operating statistics; bank/merchant disclosures; credit-performance data; state/federal licensing and enforcement records. ### Accounting normalization test Gross-versus-net revenue, customer incentives, pass-through network fees, credit receivables, securitization, and SBC can distort growth/margins. ### Valuation implementation Use DCF, EV/FCF, and growth/margin frameworks with mature take rate and capital needs. Separate networks, processors, lenders, and software-like models. ### Worked numerical mini-case > Illustrative take-rate case. TPV is $150bn and net revenue take rate 1.20%, implying $1.8bn revenue. A 10 bp take-rate decline costs $150m even if TPV is unchanged. Separate volume growth from mix, incentives, interchange/network costs, credit losses and fraud. High TPV growth can destroy value if unit contribution deteriorates. ### Monitoring and falsification cadence Breaks include take-rate compression, regulation, merchant/customer concentration, fraud/credit losses, platform disintermediation, or customer acquisition economics deteriorating. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Payments and Fintech work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 094 # REITs Analyst Playbook > Mission. Build a sector-specific research system for REITs that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model property-level occupancy, rent, mark-to-market, same-store NOI, lease expirations, concessions, capex/TI/LC, development pipeline, debt maturities, NAV, and AFFO. Separate accounting FFO from recurring economic cash available to equity. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | occupancy | Occupied leasable area/units divided by total available leasable area/units, typically average or period-end as stated. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | same-store NOI | Property revenue minus property operating expenses for the constant-property pool, excluding acquisitions, dispositions, and development per issuer definition. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | rent spreads | New or renewal rent per square foot/unit minus prior expiring rent, divided by prior expiring rent, for comparable leases. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | lease expirations | Annual base rent or square footage scheduled to expire by period divided by total annual base rent or occupied area. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | AFFO | AFFO = FFO adjusted for recurring capital needs and other economically recurring items | | cap rate | Cap rate = stabilized NOI / property value | | development yield | development yield = annualized economic output / current market value or invested base; match numerator and denominator. | | net debt/EBITDA | Net debt = interest-bearing debt + debt-like obligations - excess cash - non-operating liquid investments | ## Sector-specific accounting and comparability traps - Straight-line rent: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Tenant allowances: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Impairment: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Joint ventures: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Secured versus unsecured debt: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/AFFO: equity value per share divided by normalized AFFO/share after recurring capex and non-cash/one-time adjustments are consistently treated. - NAV: mark identifiable assets and liabilities to economic value, subtract debt and other claims, and divide residual value by diluted shares; document commodity/discount-rate assumptions. - Implied cap rate: property NOI divided by enterprise value attributable to operating real estate, adjusting debt, development, JV interests, and non-income assets. - Dividend yield: annualized sustainable common dividend divided by share price; test payout coverage, regulatory/capital constraints, cyclicality, and reinvestment needs. ## Sector diligence questions - What is the most important leading indicator for REITs, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress occupancy and same-store NOI together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test straight-line rent. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model same-store NOI, occupancy, rent spread, lease maturity, tenant credit, development, acquisitions/dispositions, recurring capex, debt, and shares. ### Leading-indicator dashboard Track leasing activity, market rents, vacancy, cap rates, transaction volumes, financing spreads, construction supply, and tenant health. ### Primary-source map SEC filings and supplemental packages; property-level leasing disclosures; Census construction data; local assessor/permit records; broker market data where available; debt maturity and secured-financing documents. ### Accounting normalization test FFO/AFFO definitions, straight-line rent, tenant improvements, leasing commissions, unconsolidated JVs, and development capitalization need reconciliation. ### Valuation implementation Use NAV, implied cap rate, AFFO/FCF, and property-level DCF with leverage. ### Worked numerical mini-case > Illustrative NOI/NAV case. Same-store NOI is $400m and cap rate 5.5%, implying about $7.27bn gross property value before development, debt and other claims. A 50 bp cap-rate expansion lowers that value to about $6.67bn. Reconcile AFFO for recurring capex, leasing costs and straight-line rent, and model lease rollover rather than relying on headline FFO. ### Monitoring and falsification cadence Breaks include tenant distress, refinancing at uneconomic rates, oversupply, development overruns, cap-rate expansion, or recurring capex above AFFO assumptions. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The REITs work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 095 # Homebuilders Analyst Playbook > Mission. Build a sector-specific research system for Homebuilders that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model communities, absorption, orders, cancellations, backlog, closings, ASP, incentives, lot costs, land pipeline, cycle time, gross margin, and inventory turns. Stress price and pace together because fixed land positions create operating leverage. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | orders | Gross value of customer purchase orders or bookings received during the period, adjusted for cancellations and scope changes where disclosed. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | closings | Completed home/property transactions recognized during the period under the company's revenue-recognition policy. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | backlog | Contracted or awarded revenue not yet recognized, using the issuer's disclosed backlog definition and separating cancellable/undedicated amounts when possible. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | ASP | Revenue attributable to the relevant product family divided by units sold/shipped, adjusted for rebates, mix, and channel treatment. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | gross margin | Gross margin = gross profit / revenue | | cancellations | Cancelled home orders divided by gross new orders during the period, preferably on a unit basis and supplemented by dollar value. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | community count | Average or period-end active selling communities available to take orders during the period. Validation: Reperform the count from the defined population, inspect every material exception, and confirm the denominator/universe did not change between periods. | | land lots | Owned and controlled homebuilding lots by stage, with optioned versus owned lots and years of supply separately shown. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Land options: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Inventory impairments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Interest capitalization: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Mortgage jvs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Incentives: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Mid-cycle P/E: current equity value divided by estimated mid-cycle diluted EPS using normalized volumes, prices, margins, credit, taxes, and share count. - P/B: common equity value divided by common book value; interpret with asset quality, mark-to-market exposure, sustainable ROE, growth, and cost of equity. - Land-adjusted NAV: mark owned/controlled land and housing inventory to economic value, subtract development costs, debt, taxes, and corporate claims, then derive equity value. - FCF valuation: capitalize or discount normalized free cash flow after maintenance capex, working capital, cash taxes, SBC/dilution, and required reinvestment are explicitly modeled. ## Sector diligence questions - What is the most important leading indicator for Homebuilders, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress orders and closings together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test land options. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model orders, closings, backlog, ASP, incentives, gross margin, lots, land spend, cycle time, cancellations, and mortgage-rate affordability. ### Leading-indicator dashboard Track new-home sales, permits, starts, mortgage rates, resale inventory, incentives, community count, order pace, land prices, and cancellations. ### Primary-source map SEC filings; Census housing starts/permits/new-home sales; Freddie Mac mortgage rates; local land/permit data; lumber/material inputs; backlog, cancellation and community-count disclosures. ### Accounting normalization test Land impairments, optioned versus owned lots, mortgage operations, incentives, and capitalized interest affect cycle comparisons. ### Valuation implementation Use normalized P/E, P/B, FCF, land value, and through-cycle ROE. ### Worked numerical mini-case > Illustrative backlog case. Backlog 8,000 homes at $480k average value equals $3.84bn gross backlog. If cancellation rises from 12% to 20%, about $307m more backlog is at risk before replacement orders. Model communities, absorptions, incentives, land basis and mortgage-rate affordability rather than treating backlog as guaranteed revenue. ### Monitoring and falsification cadence Breaks include affordability shock, land overcommitment, cancellations, resale inventory normalization, or margin compression from incentives. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Homebuilders work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 096 # Restaurants Analyst Playbook > Mission. Build a sector-specific research system for Restaurants that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model units, same-store sales from traffic and ticket, mix, food/labor/occupancy cost, franchise economics, unit openings/closures, development cost, payback, and digital/loyalty. Avoid mistaking price-led comps for healthy traffic. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | same-store sales | Sales growth for locations operating in both comparison periods under a consistent eligibility definition, decomposed into traffic and ticket where possible. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | traffic | Customer visits or transactions at comparable locations, usually expressed as year-over-year % change using a constant-store base. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | ticket | Average sales per transaction/order, equal to comparable sales divided by comparable transaction count. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | unit growth | unit growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | restaurant margin | restaurant margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | | labor | Labor expense divided by sales, or labor hours/cost per unit, with wage rate and productivity separated. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | food cost | Food and beverage input cost divided by restaurant sales, adjusted for commodity inflation, waste, and menu mix. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | franchise mix | Franchised units, system sales, or franchise revenue divided by total system units/sales/revenue as defined. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Lease accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pre-opening expense: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Impairments: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Franchise accounting: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Loyalty deferrals: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - Unit-economics DCF: forecast customers/units by cohort, ARPU/ticket, retention, contribution margin, acquisition cost, payback, fixed costs, reinvestment, and mature cohort economics. ## Sector diligence questions - What is the most important leading indicator for Restaurants, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress same-store sales and traffic together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test lease accounting. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model units, same-store sales from traffic and ticket, restaurant margin, labor, food, occupancy, new-store ramp, closures, franchise mix, and capex. ### Leading-indicator dashboard Track traffic, menu pricing, promotions, commodity inputs, wage inflation, unit openings, franchisee health, digital mix, and consumer spending. ### Primary-source map SEC filings; Bureau of Labor Statistics wage/CPI data; Census retail/food-service sales; public menu pricing; location openings/closures; franchise disclosures where public. ### Accounting normalization test Franchise versus company-store mix, gift cards, leases, closure costs, and preopening expense change comparability. ### Valuation implementation Use unit-level DCF, EV/EBITDA/FCF, and growth-adjusted frameworks tied to new-unit returns and mature store economics. ### Worked numerical mini-case > Illustrative comp-sales case. Traffic declines 4% while menu price rises 6% and mix adds 1%, producing roughly 3% comparable sales before interaction. If restaurant-level labor/food inflation totals 5%, positive comps may still produce margin compression. Separate company-owned and franchised economics, new-unit maturation, closures and cannibalization. ### Monitoring and falsification cadence Breaks include traffic elasticity to price, unit cannibalization, labor inflation, weak franchisee economics, or declining new-store cash-on-cash returns. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Restaurants work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 097 # Retail Analyst Playbook > Mission. Build a sector-specific research system for Retail that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model stores, traffic, conversion, units per transaction, AUR, e-commerce, gross margin, markdowns, shrink, inventory turns, occupancy, labor, and working capital. Track inventory composition and promotional intensity by channel. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | comps | Comparable-store sales growth for the constant-store base; reconcile traffic, ticket, calendar, FX, and closure impacts. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | traffic | Customer visits or transactions at comparable locations, usually expressed as year-over-year % change using a constant-store base. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | ticket | Average sales per transaction/order, equal to comparable sales divided by comparable transaction count. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | units | Physical units shipped, sold, or produced for the defined product scope and period; reconcile returns, channel inventory, and product mix. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | gross margin | Gross margin = gross profit / revenue | | inventory turns | Annualized COGS divided by average inventory; for retailers also monitor weeks of supply and aged inventory. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | markdowns | Gross margin reduction from price markdowns divided by gross sales or inventory retail value, with clearance and promotional markdowns separated. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | e-commerce mix | Digital/e-commerce sales divided by total retail sales using a consistent fulfillment/revenue-recognition definition. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | ## Sector-specific accounting and comparability traps - Inventory reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Vendor allowances: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Leases: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Gift cards: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Loyalty: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Retail, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress comps and traffic together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test inventory reserves. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model stores, square footage, traffic, conversion, units per transaction, ASP, e-commerce, gross margin, shrink, inventory turns, rent, labor, and capex. ### Leading-indicator dashboard Track foot traffic, card spend, promotions, inventory, markdowns, freight, vendor terms, store openings/closures, and consumer credit. ### Primary-source map SEC filings; Census retail sales; company inventory and traffic disclosures; public pricing/promotions; import/shipping data where lawful and reliable; lease/credit disclosures. ### Accounting normalization test Lease obligations, vendor allowances, gift cards, inventory reserves, private-label credit, and supplier finance can distort cash/margins. ### Valuation implementation Use mid-cycle FCF, EV/EBIT, and SOTP for credit/e-commerce where relevant. Inventory and lease intensity matter. ### Worked numerical mini-case > Illustrative inventory case. Sales $10bn, gross margin 35%, and inventory rises 18% while sales rise 3%. If markdowns reduce gross margin 150 bp, gross profit falls $150m before fixed-cost leverage. Track traffic, conversion, units per transaction, AUR, weeks of supply and lease obligations. Inventory quality matters more than inventory growth alone. ### Monitoring and falsification cadence Breaks include sustained traffic loss, markdown cycle, inventory obsolescence, lease burden, e-commerce economics, or vendor tightening. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Retail work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 098 # Consumer Packaged Goods Analyst Playbook > Mission. Build a sector-specific research system for Consumer Packaged Goods that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model volume, price, mix, distribution, market share, commodities, productivity, advertising, innovation, retailer inventory, and elasticities. Separate nominal pricing from real unit growth and distribution gains. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | organic sales | Reported sales growth excluding acquisitions/divestitures and usually FX, using the company's disclosed constant-currency/organic methodology. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | volume | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | price | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | mix | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | gross margin | Gross margin = gross profit / revenue | | market share | Company sales/units divided by total category/market sales/units for the same geography, channel, and period. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | promotion | Promotional discount dollars, promoted volume, or promotion weeks divided by total sales/volume/time, with depth and frequency separated. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | advertising | Advertising and marketing spend divided by sales, supplemented by incremental sales/gross profit per dollar where measurable. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Trade spend: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Inventory: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Restructuring: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Brand impairment: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Consumer Packaged Goods, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress organic sales and volume together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test trade spend. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model volume, price, mix, distribution, market share, commodities, advertising, trade spend, and working capital by category/geography. ### Leading-indicator dashboard Track scanner data, retailer inventory, price gaps, promotion, commodity costs, share, distribution points, consumer confidence, and private-label penetration. ### Primary-source map SEC filings; BLS/CPI and commodity inputs; public scanner/retailer data when available; Census consumption categories; retailer filings; brand/distribution disclosures. ### Accounting normalization test Trade promotions, pension, restructuring, brand intangibles, FX, and acquisition accounting can flatter organic comparisons. ### Valuation implementation Use DCF, P/E, EV/EBIT, and FCF yield with durable brand/share and reinvestment assumptions. ### Worked numerical mini-case > Illustrative price-volume case. Volume falls 6%, price rises 8% and mix adds 1%. Nominal sales grow about 3%, but unit economics and share may be deteriorating. If commodity relief adds 150 bp margin, separate that from brand pricing power. Test elasticity, private-label share, distribution and advertising support before treating pricing as durable. ### Monitoring and falsification cadence Breaks include volume elasticity, private-label share gain, retailer bargaining power, brand underinvestment, or price increases masking unit decline. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Consumer Packaged Goods work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 099 # Pharmaceuticals Analyst Playbook > Mission. Build a sector-specific research system for Pharmaceuticals that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Build product-level patients, diagnosis, penetration, price/gross-to-net, adherence, exclusivity/patent, pipeline probability, R&D, milestones, royalties, and geographic mix. Explicitly model LOE erosion and pipeline replacement needs. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | product sales | Revenue from commercialized products recognized during the period, net of gross-to-net deductions under the relevant accounting policy. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | TRx/NBRx | Total prescriptions and new-brand prescriptions over the period from a consistent prescription dataset, normalized for days and channel coverage. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | price | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | gross-to-net | Gross product sales less rebates, chargebacks, discounts, returns, and other deductions divided by gross product sales. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | patent life | Time remaining until key composition-of-matter, use, or regulatory exclusivity expires, adjusted for jurisdiction and expected litigation/extension outcomes. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | pipeline milestones | Dated clinical, regulatory, launch, or commercial events for pipeline assets, linked to development stage, probability, and value impact. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | R&D | Research and development expense, plus material capitalized development when applicable, divided by revenue or analyzed by absolute spend and program mix. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | probability of success | Probability that an asset advances from its current development stage to approval/commercial success, using stage-specific base rates adjusted for asset evidence. Validation: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | ## Sector-specific accounting and comparability traps - Milestones: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Acquired ipr&d: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Contingent consideration: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Gross-to-net reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Litigation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Risk-adjusted NPV: probability-weight each asset/project cash flow by technical/regulatory/commercial success, discount by timing/risk, and subtract remaining development/funding costs. - P/E ex pipeline: value the commercial earnings base using normalized EPS, then add separately risk-adjusted pipeline value and subtract associated development funding needs. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Pharmaceuticals, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress product sales and TRx/NBRx together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test milestones. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model prescriptions/patients, price, gross-to-net, indication, geography, patent/exclusivity, R&D, milestones, royalties, and launch curves. ### Leading-indicator dashboard Track prescriptions, formulary access, trial readouts, competitor data, payer coverage, patent litigation, manufacturing, and regulatory milestones. ### Primary-source map SEC filings; FDA labels, approvals, Complete Response Letters and trial databases; CMS pricing/reimbursement releases; patent/Orange Book or Purple Book data; prescription data where lawfully available. ### Accounting normalization test Gross-to-net reserves, acquired IPR&D, collaboration revenue, contingent consideration, milestone accounting, and patent lives are critical. ### Valuation implementation Use product-level DCF/SOTP with probability, patent cliffs, and replacement pipeline. Mature P/E alone can hide concentration. ### Worked numerical mini-case > Illustrative product case. 100k eligible patients x 20% penetration x $40k net annual price = $800m peak revenue before persistence, ramp and gross-to-net changes. A patent loss five years earlier can remove multiple years of high-margin cash flow. Build product-level DCF with indication-specific patients, probability, launch timing and exclusivity rather than a companywide sales multiple. ### Monitoring and falsification cadence Breaks include clinical failure, reimbursement restriction, safety signal, patent loss, competitor superiority, or pipeline unable to replace cliffs. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Pharmaceuticals work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 100 # Biotechnology Analyst Playbook > Mission. Build a sector-specific research system for Biotechnology that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Build probability-adjusted program value by indication, phase, addressable patients, price, probability of technical/regulatory success, timeline, dilution, partnership economics, and cash runway. Use scenario trees, not a single success probability. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | cash runway | Unrestricted cash and marketable securities divided by forecast monthly/quarterly net cash burn after committed financings and milestone payments. Validation: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | trial enrollment | Patients enrolled or randomized divided by target enrollment and elapsed enrollment time; track sites activated and enrollment rate per site. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | endpoints | DPO = average accounts payable / COGS x days in period | | probability of success | Probability that an asset advances from its current development stage to approval/commercial success, using stage-specific base rates adjusted for asset evidence. Validation: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. | | cash burn | Cash operating and investing outflows less recurring cash inflows over the period, excluding financing inflows; use a normalized forward burn rate. Validation: Tie cash, debt, facilities, maturities, and fixed charges to balance-sheet/footnote data; stress availability restrictions, refinancing assumptions, and downside cash generation. | | dilution | Net dilution = ending diluted share count / beginning diluted share count - 1, adjusted for major capital actions | | partner economics | Company share of economics from partnered assets, including upfronts, milestones, cost sharing, royalties, profit splits, and commercialization obligations. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | launch readiness | Weighted completion of manufacturing, regulatory, payer, distribution, commercial, and field-force milestones required for launch by the planned date. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - R&d expense: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Milestone recognition: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Warrants: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Convertibles: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Going concern: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - Risk-adjusted NPV: probability-weight each asset/project cash flow by technical/regulatory/commercial success, discount by timing/risk, and subtract remaining development/funding costs. - Cash plus pipeline value: unrestricted net cash plus probability-weighted NPV of pipeline assets minus remaining R&D, corporate burn, milestone obligations, and dilution/funding needs. - Scenario valuation: assign internally consistent operating and financing cases, value each case independently, and probability-weight only after defining mutually exclusive outcomes and thesis-break states. ## Sector diligence questions - What is the most important leading indicator for Biotechnology, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress cash runway and trial enrollment together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test R&D expense. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model patient population, diagnosis, eligible share, penetration, price, persistence, trial probability, launch timing, cash burn, and financing. ### Leading-indicator dashboard Track enrollment, trial timelines, biomarker data, regulatory interactions, competitor readouts, manufacturing readiness, KOL adoption, and cash runway. ### Primary-source map SEC filings; ClinicalTrials.gov; FDA meeting/approval documents and labels; peer scientific publications; patent records; financing and cash-runway disclosures. ### Accounting normalization test R&D expense, acquired IPR&D, milestone payments, collaboration accounting, warrant/convertible dilution, and going-concern risk dominate. ### Valuation implementation Use probability-adjusted product DCF/SOTP and explicit financing dilution. Avoid revenue multiples without clinical and capital path. ### Worked numerical mini-case > Illustrative rNPV case. A program has $2.0bn peak sales potential, 60% operating margin, 35% probability of approval and four years to launch. Probability-adjust cash flows and subtract expected R&D plus future financing dilution; do not multiply peak sales by probability. Cash runway and next catalyst timing are part of intrinsic value because financing can change ownership before clinical success is known. ### Monitoring and falsification cadence Breaks include trial failure, safety, regulatory delay, financing shortfall, competing therapy superiority, or commercial adoption below the required patient share. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Biotechnology work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 101 # Medical Devices Analyst Playbook > Mission. Build a sector-specific research system for Medical Devices that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model procedure volumes, installed base, utilization, ASP, consumables/service attach, placements, reimbursement, salesforce productivity, manufacturing yield, and R&D. Separate capital placements from recurring high-margin pull-through. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | procedure volume | Number of procedures performed using the relevant therapy/device during the period, adjusted for reporting coverage and seasonality. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | installed base | Number of active installed systems/devices available for clinical/commercial use at period end, net of removals and inactive units. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | utilization | Actual productive output or occupied capacity divided by practical available capacity after planned downtime, yield loss, and maintenance constraints. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | ASP | Revenue attributable to the relevant product family divided by units sold/shipped, adjusted for rebates, mix, and channel treatment. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | consumables | Disposable/recurring product revenue or units per installed system/procedure, with utilization and pricing separately identified. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | service | Service/maintenance revenue per installed system or as a % of total revenue, with contract attach and renewal rates where available. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | gross margin | Gross margin = gross profit / revenue | | R&D | Research and development expense, plus material capitalized development when applicable, divided by revenue or analyzed by absolute spend and program mix. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | ## Sector-specific accounting and comparability traps - Warranty: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Inventory obsolescence: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized software: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Acquisition intangibles: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Regulatory reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Medical Devices, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress procedure volume and installed base together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test warranty. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model procedures, installed base, units per procedure, ASP, disposable pull-through, service, surgeon/hospital adoption, and gross margin. ### Leading-indicator dashboard Track procedure volumes, hospital capex, utilization, trial/approval milestones, reimbursement, sales-force productivity, and competitor launches. ### Primary-source map SEC filings; FDA 510(k), PMA, recall and MAUDE databases; CMS reimbursement; procedure-volume sources; hospital-capex disclosures; peer product launches. ### Accounting normalization test Consignment inventory, warranty, capitalized development, acquisition amortization, and distributor inventory can affect comparisons. ### Valuation implementation Use DCF, EV/EBIT, and growth/ROIC frameworks. Separate installed-base capital from recurring consumables/service economics. ### Worked numerical mini-case > Illustrative installed-base case. 5,000 installed systems generate 120 procedures each per year and $450 disposable revenue per procedure, implying $270m recurring disposable revenue before service. A 10% utilization increase adds $27m without new installs. Separate installed-base placements from utilization and disposable pull-through, then stress recall/reimbursement and sales-force productivity. ### Monitoring and falsification cadence Breaks include safety/recall, reimbursement, slower utilization, surgeon switching, hospital budget pressure, or installed base failing to generate expected pull-through. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Medical Devices work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 102 # Managed Care Analyst Playbook > Mission. Build a sector-specific research system for Managed Care that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model members by product, premium yield, MLR, risk adjustment, utilization, acuity, provider rates, SG&A, Stars/quality, Medicaid/Medicare rate notices, and capital. Separate temporary utilization noise from structural medical-cost trend. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | membership | Average or period-end covered lives/members under the health plan, segmented by product and funding type. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | premium yield | premium yield = annualized economic output / current market value or invested base; match numerator and denominator. | | medical cost trend | Year-over-year change in medical cost per member/procedure after utilization, unit cost, mix, and acuity effects. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | MLR | Medical claims expense plus defined quality/improvement items divided by premium revenue under the applicable accounting/regulatory definition. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | risk adjustment | Net risk-adjustment receivable/payable or revenue effect divided by premium revenue/member months, reconciled to program methodology. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | SG&A | Selling, general, and administrative expense divided by revenue/premiums, with acquisition, commission, and restructuring items separately identified. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | star ratings | CMS Medicare Advantage Star Ratings by contract weighted by membership/revenue, with the share of members in 4+ star plans highlighted. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capital | Statutory/regulatory capital and surplus or risk-based capital available above required minimums, reconciled to parent liquidity and dividend restrictions. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | ## Sector-specific accounting and comparability traps - Ibnr reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Risk adjustment: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Premium deficiency: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Regulatory capital: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Acquisition intangibles: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Managed Care, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress membership and premium yield together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test IBNR reserves. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model members by product, premiums, medical cost trend, MLR, risk adjustment, pharmacy, administrative cost, stars/quality, and capital. ### Leading-indicator dashboard Track enrollment, utilization, provider rates, drug trend, government rate notices, risk adjustment, star ratings, and policy changes. ### Primary-source map SEC filings; CMS rate notices, enrollment and star-rating data; state Medicaid procurement/rate materials; HHS/CMS policy releases; provider and pharmacy disclosures. ### Accounting normalization test Medical claims reserves, risk adjustment, pharmacy rebates, acquisitions, statutory capital, and government program timing are key. ### Valuation implementation Use P/E/FCF and DCF with normalized margin by product and capital needs. Growth without adequate pricing can destroy value. ### Worked numerical mini-case > Illustrative MLR case. Premium revenue $20bn at 84% MLR implies $3.2bn gross margin before SG&A. A 150 bp MLR increase consumes $300m pretax. If pricing resets with a lag, liquidity and capital absorb the mismatch first. Model enrollment, rate, utilization, provider cost, risk adjustment and stars by product. Growth at inadequate pricing is negative value. ### Monitoring and falsification cadence Breaks include medical-cost trend outrunning pricing, reimbursement cuts, quality-rating decline, risk-adjustment change, or regulatory constraints. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Managed Care work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 103 # Telecom Analyst Playbook > Mission. Build a sector-specific research system for Telecom that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model subscribers, adds/churn, ARPU, device economics, network capex, spectrum, tower/fiber commitments, convergence, competition, and leverage. Distinguish accounting EBITDA from cash after spectrum and network investment. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | subscribers | Paying customer/subscriber accounts at period end or average for the period under a consistent active-status definition. Validation: Tie physical/operating units to company disclosures or source-system data; reconcile beginning/ending populations where applicable and test scope, ownership, and period consistency. | | ARPU | Average revenue per user/subscriber: relevant service revenue divided by average users/subscribers and period units, such as month or quarter. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | churn | Subscribers/customers lost during the period divided by opening or average subscriber base under a clearly stated gross/net churn convention. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | net adds | Gross subscriber additions minus disconnects/churn during the period, reconciled to beginning and ending subscriber counts. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | fiber passings | Locations serviceable by the fiber network, distinguished from connected subscribers and homes/businesses merely planned for build. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capex | Maintenance capex is the spending required to preserve current earning power, estimated from asset replacement and operating evidence rather than management labels alone | | spectrum | MHz-pop or licensed spectrum holdings by band/geography, adjusted for ownership and usable bandwidth; pair with subscribers/traffic for capacity context. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | service margin | service margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | ## Sector-specific accounting and comparability traps - Spectrum licenses: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Leases: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Device financing: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized labor: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - Sum-of-the-parts: value each economically distinct segment with its appropriate framework, then subtract corporate costs and all non-common claims before deriving equity value. ## Sector diligence questions - What is the most important leading indicator for Telecom, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress subscribers and ARPU together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test spectrum licenses. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model subscribers, gross adds, churn, ARPU, device economics, network usage, spectrum, capex, tower/fiber lease, and promotions. ### Leading-indicator dashboard Track porting, subscriber adds, pricing/promotions, handset cycles, spectrum auctions, capex, network quality, fixed-wireless/fiber adds, and competition. ### Primary-source map SEC filings; FCC subscriber, spectrum and broadband data; auction records; company network/capex disclosures; public porting/coverage data where available; tower/fiber counterparties. ### Accounting normalization test Device financing, leases, spectrum capitalization, pension, tower transactions, and customer acquisition costs complicate FCF. ### Valuation implementation Use DCF, EV/EBITDA with capex/leasing normalization, and sum-of-parts where infrastructure assets differ. ### Worked numerical mini-case > Illustrative subscriber case. 50m subscribers at $55 monthly ARPU produce $33bn annual service revenue. A 50 bp increase in monthly churn materially raises gross-add requirements and acquisition cost even if period-end subscribers look stable. Bridge net adds, ARPU, device subsidies, spectrum, network capex and lease obligations before using EV/EBITDA. ### Monitoring and falsification cadence Breaks include price war, churn spike, network disadvantage, capex intensity, spectrum cost, or leverage limiting investment. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Telecom work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 104 # Internet Platforms and Marketplaces Analyst Playbook > Mission. Build a sector-specific research system for Internet Platforms and Marketplaces that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model users/buyers/sellers, engagement, transactions/GMV, take rate, ad load/yield, acquisition cost, cohort retention, trust/safety expense, network effects, and regulatory risk. Separate monetization gains from deteriorating ecosystem health. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | MAU/DAU | Monthly or daily active users meeting the platform's activity definition; use DAU/MAU to assess frequency where definitions are stable. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | engagement | Time spent, sessions, content interactions, transactions, or other core activity per active user over a defined period. Validation: Verify start/end timestamps or periods from source records, use a consistent calendar/business-day convention, and test outliers rather than averaging them away. | | GMV | Gross merchandise value transacted through the platform before merchant payouts, returns, taxes, or pass-through items according to the stated definition. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | take rate | Net revenue attributable to transaction volume divided by TPV/GMV or other monetized volume, after pass-through items as defined. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | ad load | Advertising impressions or ads served divided by content units, sessions, or time, using a stable denominator that reflects user exposure. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | ARPU | Average revenue per user/subscriber: relevant service revenue divided by average users/subscribers and period units, such as month or quarter. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | seller/buyer growth | seller/buyer growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | contribution margin | contribution margin = relevant profit or cash-flow numerator / relevant revenue base, using a consistent definition. | ## Sector-specific accounting and comparability traps - Gross versus net: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Stock compensation: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Content costs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized software: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Regulatory contingencies: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/revenue: enterprise value divided by normalized revenue; use only with an explicit gross-margin, operating-margin, growth, and capital-intensity bridge. - EV/gross profit: enterprise value divided by normalized gross profit; useful when revenue recognition/pass-through differs, but still requires opex and capital-intensity normalization. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Internet Platforms and Marketplaces, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress MAU/DAU and engagement together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test gross versus net. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model users/buyers/sellers, engagement, transactions or GMV, take rate, ad load/pricing, fulfillment, payment, and content costs. ### Leading-indicator dashboard Track traffic/app engagement, conversion, merchant/seller adds, ad pricing, e-commerce spend, regulatory changes, and competitive acquisition costs. ### Primary-source map SEC filings; platform-reported users/GMV/ad metrics; public app/web statistics with methodology controls; regulator/antitrust filings; merchant/advertiser disclosures; payment/fulfillment data. ### Accounting normalization test Gross-versus-net revenue, traffic acquisition costs, SBC, content capitalization, payments/credit, and international FX can distort margins. ### Valuation implementation Use DCF, EV/FCF, and segment SOTP. Reverse price into user growth, monetization, take rate, and mature margin. ### Worked numerical mini-case > Illustrative marketplace case. $100bn GMV at 12% take rate produces $12bn gross revenue before payment/fulfillment subsidies. A 100 bp take-rate decline costs $1bn unless GMV or ancillary monetization offsets it. Model users, frequency, spend/order, take rate and variable fulfillment/payment economics, then test multi-homing and disintermediation. ### Monitoring and falsification cadence Breaks include platform disintermediation, regulation, user engagement decline, take-rate pressure, rising acquisition cost, or network effects weakening through multi-homing. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Internet Platforms and Marketplaces work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 105 # Cybersecurity Analyst Playbook > Mission. Build a sector-specific research system for Cybersecurity that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model ARR, new ARR, NRR, platform/module adoption, billings/RPO, gross margin, sales productivity, CAC payback, partner channel, SBC, and FCF. Test whether consolidation improves customer economics or simply bundles discounting. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | ARR | Annualized recurring run-rate from active contracts at period end, usually monthly recurring revenue × 12 plus other contractually recurring components; exclude one-time services. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | NRR | NRR = ending recurring revenue from beginning cohort / beginning cohort recurring revenue | | RPO | Contracted revenue not yet recognized under ASC 606/IFRS 15-style disclosure scope; separate current RPO from amounts expected beyond 12 months. Validation: Reconcile beginning balance + additions - revenue/shipments - cancellations/adjustments to ending balance where data allow; verify cancellation rights, timing, and definition changes. | | billings | Revenue plus period-over-period increase in deferred revenue/contract liabilities, adjusted for acquisitions, FX, and contract assets when needed for comparability. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | platform mix | Revenue, bookings, workload, or customer usage attributable to the strategic platform/product family divided by the corresponding total. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | seat/usage growth | seat/usage growth = current period / comparable prior period - 1; decompose organic, price, volume, mix, FX, and M&A where material. | | SBC | Stock-based compensation expense recognized for the period, reconciled to equity awards, cash-flow add-back, and diluted share count. Validation: Tie the dollar measure to filed statements/footnotes; reconcile classification adjustments, one-time items, acquisitions/FX, and period consistency before using it analytically. | | FCF margin | FCF margin = normalized free cash flow / revenue | ## Sector-specific accounting and comparability traps - Contract timing: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Deferred revenue: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Sbc: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Capitalized commissions: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - M&a add-backs: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/revenue: enterprise value divided by normalized revenue; use only with an explicit gross-margin, operating-margin, growth, and capital-intensity bridge. - EV/FCF: enterprise value divided by unlevered free cash flow; ensure interest is excluded from FCF and debt-like claims are included in enterprise value. - Rule of 40: revenue growth plus a consistently defined profitability metric, normally FCF or operating margin; use to frame quality/growth tradeoffs, not as a direct valuation formula. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Cybersecurity, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress ARR and NRR together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test contract timing. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model ARR, new ARR, retention, platform/module adoption, billings, RPO, seats/endpoints/usage, gross margin, sales productivity, and SBC. ### Leading-indicator dashboard Track breach environment, IT budgets, platform consolidation, sales hiring, renewal commentary, cloud workloads, channel activity, and competitor displacement. ### Primary-source map SEC filings; issuer ARR/RPO/retention definitions; CISA vulnerability and incident publications; federal IT/security spending; channel/partner disclosures; customer and cloud-platform evidence. ### Accounting normalization test Contract timing, deferred revenue, capitalized commissions, SBC, M&A add-backs, and usage pricing changes can mask economics. ### Valuation implementation Use DCF, EV/FCF, and revenue/gross-profit multiples with mature margin and retention. Reverse DCF should test platform consolidation and duration. ### Worked numerical mini-case > Illustrative ARR case. Beginning ARR $1.0bn, gross churn 5%, contraction 2%, expansion 15%, and new ARR $220m. Ending ARR = 1.0 x 1.08 + .22 = $1.30bn. Opening-cohort NRR is 108%. Tie platform/module adoption to retention and sales efficiency, then subtract SBC dilution and capitalized commissions when judging FCF quality. ### Monitoring and falsification cadence Breaks include retention decline, sales efficiency deterioration, platform commoditization, cloud-provider bundling, product failure/breach, or excessive dilution. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Cybersecurity work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. ## PART XV - SECTOR PLAYBOOKS | MODULE 106 # Railroads and Logistics Analyst Playbook > Mission. Build a sector-specific research system for Railroads and Logistics that converts operating data into financial outcomes, highlights the accounting areas most likely to distort comparability, and selects valuation methods that reflect the sector's economics. ## Economic engine and binding constraints Model volume by commodity, revenue per unit, fuel surcharge, velocity, dwell, train length, labor productivity, network capacity, capex, service reliability, and pricing. Stress service degradation because near-term cost cuts can damage long-term network economics. ## Primary KPI stack | KPI | Construction / analyst control | | --- | --- | | volume | Revenue growth bridge = volume effect + price effect + mix effect + FX/acquisition effects, using a consistent base | | revenue per unit | Revenue attributable to the defined transportation/service unit divided by units/loads/cars/shipments handled in the same period. Validation: Recalculate price/cost from underlying dollars and physical units; test mix, rebates, FX, timing, and unit-definition effects; reconcile to reported revenue or expense. | | fuel surcharge | Fuel-surcharge revenue divided by applicable traffic/revenue units, reconciled to index formulas and fuel-cost changes. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | operating ratio | Railroad operating expenses divided by operating revenue; lower ratios indicate higher operating margin, subject to classification consistency. Validation: Recalculate from same-scope numerator and denominator; confirm period, units, cohort/geography, and issuer definition; reconcile material differences to filings or operating data. | | velocity | Average train/car movement speed over the network, typically miles per hour, using the railroad's published operating definition. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | dwell | Average hours railcars spend in yards/terminals between arrival and departure under a consistent operational definition. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | labor productivity | Output units such as carloads, train miles, ton-miles, or revenue divided by labor hours or employee count; choose the denominator that matches the operating bottleneck. Validation: Recalculate independently from cited source data; verify definition, period, units, scope, signs, and any reconciliation to reported financial or operating totals. | | capex | Maintenance capex is the spending required to preserve current earning power, estimated from asset replacement and operating evidence rather than management labels alone | ## Sector-specific accounting and comparability traps - Fuel hedges: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Pension: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Asset lives: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Leases: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. - Claims reserves: reconcile issuer treatment with peer treatment and quantify the effect on reported growth, margin, cash flow, capital, or valuation before comparing outputs. ## Valuation frameworks - EV/EBITDA: enterprise value divided by normalized EBITDA; adjust leases, pensions, minorities, recurring restructuring and capital intensity before peer comparison. - P/E: common equity value per share divided by normalized diluted EPS; normalize taxes, one-time items, dilution, cyclicality, and non-operating income. - FCF yield: normalized levered free cash flow divided by equity value; reconcile SBC, working capital, maintenance capex, taxes, and cycle before comparing companies. - DCF: forecast FCFF from operating drivers, discount at a capital-structure-consistent WACC, model terminal growth/ROIC coherently, and bridge enterprise value to common equity. ## Sector diligence questions - What is the most important leading indicator for Railroads and Logistics, and how many months does it lead reported revenue or cash flow? ## Sector stress and falsification - Stress volume and revenue per unit together in the direction most likely to break the equity story; flow the result through working capital, capex, liquidity, financing, dilution, and valuation. - Explicitly test fuel hedges. Determine whether it can make the reported sector comparison look better or worse without equivalent economic change. ## 99-point standalone execution extension ### Model architecture and forecast chain Model volume by commodity/customer, revenue per unit, fuel surcharge, operating ratio, velocity, dwell, labor, equipment utilization, capex, and network capacity. ### Leading-indicator dashboard Track carloads, intermodal volumes, industrial production, port activity, trucking rates, fuel, service metrics, labor agreements, and capex. ### Primary-source map SEC filings; Surface Transportation Board weekly rail/service data; AAR traffic statistics; BTS freight data; port statistics; trucking/fuel indicators; labor and capex disclosures. ### Accounting normalization test Fuel surcharge timing, pension, casualty/environmental reserves, asset lives, and network-capex classification can distort margins/cash. ### Valuation implementation Use DCF, EV/EBIT, FCF yield, and replacement/network value context. Normalize service and volume through cycle. ### Worked numerical mini-case > Illustrative operating-ratio case. Revenue $12bn at 60% operating ratio implies $4.8bn operating income. A 200 bp deterioration to 62% reduces operating income by $240m before volume or price changes. Bridge carloads, revenue per unit, fuel surcharge, velocity, dwell, labor and capex. Service deterioration can create future volume loss even before revenue falls. ### Monitoring and falsification cadence Breaks include service deterioration, labor inflation, modal share loss, regulatory constraints, network bottlenecks, or capex required to sustain service above modeled levels. At every quarterly update, rebuild the driver bridge from operating units to revenue, margin, cash flow and valuation; compare leading indicators with the prior forecast; record definition changes; and precommit the threshold that would trigger a thesis reset rather than a cosmetic estimate change. ## Sector exit standard The Railroads and Logistics work is complete only when the analyst can explain the business in its native operating units, reproduce the KPI history, identify the binding growth constraint and marginal price setter, normalize sector-specific accounting, quantify a coherent adverse case, and translate the current market price into the operating expectations that must be met or exceeded. # APPENDIX A - ZERO-TO-EXPERT FOUNDATION BOOT CAMP > Use this appendix before the operating modules if the analyst does not already possess the baseline accounting, modeling, valuation, and market-structure knowledge assumed by an institutional research desk. ## Accounting mechanics - Understand accrual versus cash accounting; double-entry logic; revenue/expense recognition; assets, liabilities, and equity; current versus noncurrent classification; depreciation/amortization; deferred taxes; leases; stock compensation; and consolidation. - Rebuild one real company from filed statements into a three-statement spreadsheet and prove that cash and equity roll forward. - For every income-statement line, identify the balance-sheet account or cash-flow mechanism that completes the accounting. ## Financial statement reading - Read the face statements first, then accounting policies, segment note, revenue note, debt, tax, stock compensation, commitments/contingencies, acquisitions, fair value, related parties, and auditor report. - Compare the current filing with the prior filing and highlight changed definitions, qualifiers, segment presentation, and material new disclosures. - Never use MD&A narrative as a substitute for the underlying footnote or table when the source exists. ## Spreadsheet modeling - Separate inputs, formulas, outputs, and checks. Use consistent signs, units, dates, scenario switches, and source comments. - Build formulas left-to-right with no hidden constants. Avoid excessive OFFSET/INDIRECT-style opacity, unexplained circularity, and plugs. - Create control totals and error flags before adding valuation. A model that does not reconcile cannot produce a defensible valuation. ## Corporate finance math - Master compounding, present value, annuities/perpetuities, cost of debt/equity, enterprise versus equity value, dilution, NPV, IRR, ROIC, reinvestment, and terminal-value identities. - Know why growth creates value only when incremental returns exceed the opportunity cost of capital after considering risk and reinvestment. - Understand that multiples are compressed DCF statements. Growth, margin, capital intensity, duration, and risk determine justified multiples. ## Statistics and evidence - Distinguish descriptive statistics, causal inference, prediction, and narrative. Understand sampling error, survivorship bias, look-ahead bias, base rates, regression to the mean, confounding, and multiple comparisons. - Use confidence ranges and sensitivity analysis when data cannot justify precise probabilities. - Never convert a noisy alternative-data correlation into a causal forecast without out-of-sample validation and an economic mechanism. ## Market mechanics - Understand shares outstanding, float, short interest, options/convertibles, primary versus secondary issuance, index effects, liquidity, borrow, spreads, and corporate actions. - Know the difference between business value creation and stock-price movement. The research process estimates economics and expectations; market timing remains uncertain. - Timestamp every market price, share count, debt balance, FX rate, and yield used in valuation. ## Foundation proficiency test - Without notes, explain how a credit sale affects all three statements at sale, collection, bad-debt recognition, and write-off. - Build a five-year historical model from a 10-K/10-Q set and tie revenue, operating income, cash, debt, and diluted shares exactly. - Explain why EBITDA can rise while intrinsic value falls. - Derive a DCF from NOPAT and reinvestment, then reconcile terminal growth to terminal reinvestment and ROIC. - Identify three cases in which operating cash flow can improve without underlying economics improving. - Write a one-page investment view that separates facts, estimates, judgments, catalysts, risks, and falsification conditions. # APPENDIX B - INSTITUTIONAL MODEL BUILD STANDARD | Tab / layer | Required contents | Control | | --- | --- | --- | | 00 Cover / Control | Company, ticker, valuation date, model version, scenario, analyst, last filing | No output can be published without current valuation date and source status | | 01 Sources | URL/accession, date, filing period, table/section, definition notes | Every material historical input maps to a source | | 02 Historical IS | Reported and normalized income statement | Tie to filed statements and amendments | | 03 Historical BS | Reported and normalized balance sheet | Assets = liabilities + equity | | 04 Historical CF | Reported and normalized cash flow statement | Beginning cash + movements = ending cash | | 05 Segments/KPIs | Segment and operating-driver history | Segments reconcile to consolidated or difference explained | | 06 Assumptions | All forecast assumptions and scenarios | No hidden constants in forecast formulas | | 07 Revenue driver | Units/customers/capacity/price/mix build | Historical driver model reconciles to revenue | | 08 Margin/cost | Variable/fixed/step cost architecture | Incremental margin economically plausible | | 09 Fixed assets | PP&E/capex/depreciation/intangibles | Beginning + capex/acquisition/FX - depreciation/disposal = ending | | 10 Working capital | AR, inventory, AP, deferred revenue, contract balances | Driver days/turns calculated consistently | | 11 Debt/interest | Instrument, maturity, rate, covenants, revolver, lease/debt-like claims | Interest links to average balance and rate | | 12 Taxes | Book tax, cash tax, NOLs, deferreds, discrete items | Cash tax forecast reconciles to economics | | 13 Shares/dilution | Basic shares, awards, options, converts, issuance, repurchases | Diluted share roll-forward ties | | 14 Forecast statements | Integrated IS/BS/CF | BS balance and cash roll-forward pass in every scenario | | 15 Scenarios | Base, bull, bear, stress, thesis break | Operating drivers change coherently | | 16 Valuation | DCF, multiples, reverse DCF, SOTP if needed | EV-equity bridge and dilution complete | | 17 Dashboard | Decision variables, sensitivities, catalysts, risks | No decorative KPI without a decision use | | 18 Checks | Reconciliation, sign, units, dates, scenario, circularity | All critical checks must be green before distribution | ## Formula engineering rules - Use one formula pattern across a row when economics are the same. Break the pattern only with a written reason. - Never hard-code a forecast output inside a formula. All assumptions belong in the assumptions layer or a clearly labeled scenario table. - Use named ranges sparingly and transparently. A reviewer should be able to trace a formula without a hidden dependency graph. - Use average balance for ratios that compare a period flow with a balance stock when seasonality or growth makes period-end balances misleading. - Document intentional circularity and convergence settings. Prefer algebraic solutions when practical. - Use explicit error flags for #N/A, #DIV/0!, broken source links, missing periods, stale market prices, and scenario inconsistencies. - Protect reported history from accidental overwrite. Normalize in a separate layer with a reconciliation back to reported data. # APPENDIX C - FORMULA AND CONCEPT ENCYCLOPEDIA | Metric / concept | Formula / construction | Interpretation / expert caution | | --- | --- | --- | | NOPAT | Adjusted operating profit x (1 - normalized cash tax rate) | After-tax operating profit independent of financing; normalize unusual operating items and tax appropriately. | | Invested capital | Operating assets - non-interest-bearing operating liabilities, or financing approach reconciled to same economics | Use average invested capital for ROIC and treat goodwill consistently with the question being answered. | | ROIC | NOPAT / average invested capital | Compare with opportunity cost and reinvestment runway. High legacy ROIC does not prove high incremental ROIC. | | Incremental ROIC | Change in NOPAT / change in invested capital over a meaningful interval | Measures economics of new investment. Noisy over short periods and distorted by M&A/accounting. | | Reinvestment rate | Net investment in operating assets / NOPAT | Connects growth to capital required. Include working capital and other operating investment, not just capex. | | Sustainable growth | Reinvestment rate x return on incremental invested capital | Useful terminal identity. Growth without reinvestment is inconsistent when returns are finite. | | FCFF | NOPAT + D&A and other noncash operating charges - capex - change in NWC - other operating investment | Cash available to all capital providers before financing flows. | | FCFE | Net income + noncash charges - capex - change in NWC + net borrowing | Cash available to common equity after debt financing; less robust when leverage changes materially. | | Enterprise value | Equity value + debt + preferred + minority interest - excess cash - non-operating investments | Add/subtract claims/assets consistently with the operating metric being valued. | | Net debt | Debt + debt-like obligations - excess cash - non-operating liquid investments | Restricted/operating cash may not be fully subtractable. | | WACC | E/(D+E) x Ke + D/(D+E) x Kd x (1-T) | Use market-value weights and risk inputs consistent with cash-flow currency and risk. | | Terminal value, perpetuity | FCFF in next period / (WACC - g) | Requires WACC > g and terminal economics consistent with reinvestment and returns. | | Midyear PV factor | 1 / (1 + discount rate)^(t - 0.5) for annual flows occurring through year | Adjust exponent for actual valuation date and cash-flow timing. | | FCF yield | Normalized FCFE / equity value | Useful equity cash-return lens; weak if current FCF is cyclically abnormal. | | EV/EBITDA | Enterprise value / normalized EBITDA | Useful where D&A comparability helps, but ignores capex and working-capital intensity. | | EV/EBIT | Enterprise value / normalized EBIT | Better than EBITDA when depreciation approximates economic asset consumption. | | P/E | Diluted equity value per share / normalized diluted EPS | Affected by leverage, tax, non-operating items, and buyback financing. | | PEG | P/E / growth rate | Heuristic only. Ignores duration, reinvestment, capital intensity, and risk. | | Gross margin | Gross profit / revenue | Interpret with mix, capitalization, pass-through revenue, and principal-agent presentation. | | Incremental margin | Change in profit / change in revenue over an economically comparable interval | Reveals operating leverage, but can be distorted by price-cost timing and mix. | | DSO | Average receivables / revenue x days | Collection/billing behavior. Segment/customer mix can change the normal level. | | Inventory days | Average inventory / COGS x days | Inventory intensity. Rising days with weakening demand can signal channel or obsolescence risk. | | DPO | Average payables / COGS or purchases x days | Supplier financing. A sharp increase may flatter operating cash flow. | | Cash conversion cycle | DSO + inventory days - DPO | Working-capital timing. Not comparable across structurally different business models without normalization. | | Accrual proxy | Net income - CFO, scaled consistently | Screen for earnings/cash divergence, not proof of manipulation. | | NRR | Beginning-cohort recurring revenue after churn, contraction, and expansion / beginning-cohort recurring revenue | Do not mix logo and revenue retention or inconsistent cohort definitions. | | GRR | Beginning-cohort revenue retained before expansion / beginning-cohort revenue | Measures defensive retention without expansion masking churn. | | CAC payback | Customer acquisition cost / monthly gross profit from acquired cohort | Use cohort gross profit and real acquisition cost. Companywide S&M shortcuts are approximate. | | LTV | Present value of expected cohort gross profit less servicing/retention costs | Use observed retention curve and finite economics rather than a perpetual churn shortcut. | | Net leverage | Net debt / normalized EBITDA or other appropriate cash earnings | Test covenant definition separately from analyst economic definition. | | Interest coverage | Normalized EBIT/EBITDA/cash earnings / cash interest | Match numerator to fixed obligations and include debt-like claims when material. | | Dilution rate | Ending diluted shares / beginning diluted shares - 1, adjusted for major issuance/repurchases | Evaluate together with SBC and repurchase cash cost. | | Book-to-bill | Bookings / revenue for matched definitions and period | Leading signal only if cancellations, backlog conversion, and timing are understood. | | Probability-weighted value | Sum of scenario value x scenario probability | Probabilities are judgments. Always show unweighted cases and thesis-break case. | # APPENDIX D - FORENSIC ACCOUNTING RED-FLAG ENCYCLOPEDIA - CFO persistently trails net income without a clear structural working-capital or business-model explanation. - Receivables, contract assets, or unbilled revenue grow materially faster than comparable revenue. - Inventory grows faster than sell-through while management describes demand as strong. - Large recurring one-time charges are repeatedly excluded from adjusted earnings. - Share count rises despite substantial repurchase spending; SBC and repurchases should be analyzed together. - Segment definitions or KPI definitions change near weak operating periods without a clean historical recast. - Management abandons a previously emphasized KPI without reconciling the old series. - Auditor changes, delayed filings, restatements, material weaknesses, or critical audit matters increase. - Capitalized costs rise faster than the activity they purport to support. - Reserve releases or favorable estimate changes support profit during operational weakness. - Tax-rate changes drive EPS without a durable cash-tax mechanism. - Acquisitions repeatedly reset the cost base and obscure organic growth or recurring restructuring. - Customer concentration rises while disclosure specificity falls. - Gross margin improves while cash conversion and working capital deteriorate without a causal explanation. - Bookings/backlog grow because delivery slips or contract terms change rather than demand improves. - Debt or debt-like obligations grow faster than normalized cash earnings and covenant/refinancing headroom shrinks. - Supplier finance, factoring, securitization, or customer advances materially change CFO presentation. - Related-party transactions become more frequent, larger, or less transparent. - Insider sales accelerate around promotional guidance, but interpret sales in context of plans, grants, taxes, and ownership. - Guidance ranges widen while the midpoint is emphasized, obscuring uncertainty. - Adjusted FCF excludes recurring capex-like cash uses, restructuring cash, or acquisition integration costs. - Warranty, returns, rebates, gross-to-net reserves, or loss reserves decline despite rising risk indicators. - Management claims pricing power while unit volume, mix, incentives, or customer retention deteriorate. - Fair-value inputs move favorably while observable market inputs move in the opposite direction. - Useful lives, depreciation, capitalization thresholds, or reserve methods change and materially lift earnings. - Debt documents contain restrictions or collateral requirements inconsistent with simplified liquidity language. - Non-GAAP metric labels resemble GAAP while calculation or prominence could mislead. - Cash flow improvement depends on stretching payables or drawing down inventory below sustainable operating levels. - Organic growth calculations silently change acquisition cutoff, FX convention, or discontinued product scope. - Large asset sales, insurance proceeds, litigation settlements, or gains recur near target-setting periods. ## Red-flag protocol - A flag is not a conclusion. Record it as resolved, open-immaterial, open-material, or thesis-relevant. - Seek the most benign plausible explanation and the most adverse plausible explanation, then identify evidence that distinguishes them. - Quantify the financial and valuation impact of a reasonable normalization. - Escalate when incentive, opportunity, accounting discretion, weak controls, and unreconciled cash/balance-sheet evidence appear together. # APPENDIX E - PRIMARY-SOURCE AND REGULATORY VERIFICATION SYSTEM > Rules, reporting forms, accounting standards, exchange requirements, and regulator interpretations change. This manual therefore specifies source families and a refresh process rather than freezing every rule as permanent. | Source family | Official access point | Use | | --- | --- | --- | | SEC EDGAR company filings | https://www.sec.gov/edgar/search/ | 10-K, 10-Q, 8-K, proxy, ownership, exhibits, amendments | | SEC EDGAR APIs | https://www.sec.gov/search-filings/edgar-application-programming-interfaces | Submissions metadata and XBRL company facts for retrieval/automation | | SEC Regulation S-K interpretations | https://www.sec.gov/rules-regulations/staff-guidance/corporation-finance-interpretations/regulation-s-k | Disclosure interpretations; verify current date/status | | SEC Non-GAAP C&DIs | https://www.sec.gov/rules-regulations/staff-guidance/corporation-finance-interpretations/non-gaap-financial-measures | Non-GAAP presentation, labeling, recurring items, reconciliation and prominence guidance | | FASB Accounting Standards Codification | https://asc.fasb.org/ | Authoritative nongovernmental U.S. GAAP source | | PCAOB standards | https://pcaobus.org/oversight/standards/auditing-standards | Auditor responsibilities, reporting, estimates, fraud, going concern and other audit standards | | IFRS issued standards | https://www.ifrs.org/issued-standards/ | Official IFRS standards and supporting materials | | Nasdaq rules | https://listingcenter.nasdaq.com/rulebook/nasdaq/rules | Listing and continued-listing rules | | NYSE regulation | https://www.nyse.com/regulation | NYSE rules and regulatory resources | ## Quarterly regulatory-change monitor - Record date checked, source, rule/proposal/guidance status, effective date, transition provisions, affected issuers, and analyst owner. - Separate proposed rules, final rules, staff interpretations, accounting standards, exchange rules, and enforcement actions because they carry different authority and timing. - Identify which filing alerts, model tabs, data schemas, screening logic, or research cadence must change if a new requirement becomes effective. - Never rely on an old research memo for a current rule when the primary regulator or standard setter can be checked directly. # APPENDIX F - AI AND AUTOMATION CONTROL FRAMEWORK | Layer | Required design | Failure test | | --- | --- | --- | | Source acquisition | Whitelisted primary-source retrieval with URL/accession, timestamp, document version/hash | Can the analyst reproduce the exact document used? | | Parsing | Preserve tables, sections, units, periods, footnotes, and raw text | Did parsing alter signs, columns, units, or table alignment? | | Structured extraction | Schema with entity, metric, value, unit, period, source span, confidence | Can every extracted value be traced to exact supporting text/table? | | Retrieval | Primary-source-weighted retrieval with metadata filters and provenance | Did a secondary summary displace the primary evidence? | | LLM reasoning | Use for classification, contradiction search, hypothesis generation, drafting | Is any calculation or claim unsupported by retrieved evidence? | | Deterministic compute | Code/spreadsheet for arithmetic, reconciliation, valuation, and checks | Can the calculation be reproduced without the language model? | | Validation | XBRL/table cross-checks, totals, signs, units, historical definitions, peer definitions | Are discrepancies routed to human review rather than silently averaged? | | Security | Treat retrieved content as untrusted data; isolate action permissions | Can a document instruction alter system behavior or trigger an external action? | | Human approval | Required for assumption changes, publication, compliance-sensitive channel work, external actions | Is a named reviewer accountable for the final decision? | | Evaluation | Golden dataset, regression tests, citation accuracy, numeric accuracy, hallucination rate, analyst corrections | Did a model/prompt/schema change degrade known tasks? | | Versioning | Version model, prompt, schema, code, source set, output, and reviewer | Can a past conclusion be reconstructed after systems change? | ## Minimum AI evaluation suite - Numeric extraction: exact match and tolerance-based accuracy by table type, sign, unit, and period. - Citation support: percentage of material claims whose cited span directly supports the claim. - Reconciliation: percentage of extracted statements/segments that tie within stated tolerance. - Definition drift detection: known historical KPI-definition changes should be caught, dated, and surfaced. - Hallucination: unsupported factual claims per research output, with a target approaching zero for publishable material claims. - False-positive red flags: accounting or risk flags that disappear after primary-source reconciliation. - Analyst correction rate: percentage of AI-generated values, classifications, or conclusions requiring human correction. - Automation yield: time saved after correction and review, not gross machine output volume. - Regression stability: rerun a fixed benchmark set after every model, prompt, parser, or schema change. # APPENDIX G - FULL INITIATION WORKFLOW AND DELIVERABLE PACK | Stage | Required output | | --- | --- | | 1. Decision framing | One-sentence decision question; time horizon; key valuation variables; stop conditions. | | 2. Source archive | Latest 10-K/10-Q/8-K/proxy, debt documents, material exhibits, presentations, calls, ownership forms, relevant regulator/industry sources. | | 3. Historical model | Five years and eight quarters where available; statements, segments, KPIs, share count, debt, taxes, working capital, capex. | | 4. Accounting memo | Revenue policy, capitalization, reserves, SBC, leases, taxes, pensions, M&A, fair value, non-GAAP, auditor issues. | | 5. Business model map | Customer, purchase decision, value proposition, pricing, distribution, cost structure, unit economics, capital needs, bottlenecks. | | 6. Industry map | Value chain, profit pools, market size, competitors, substitutes, technology, regulation, cycle, supply constraints. | | 7. Management/governance | Capital-allocation chronology, guidance accuracy, incentives, ownership, board, related parties, disclosure candor. | | 8. Forecast model | Driver-based revenue, margin architecture, working capital, capex, taxes, debt, interest, shares, integrated statements. | | 9. Valuation | DCF, relevant multiples, reverse DCF, SOTP if needed, EV-equity bridge, dilution, sensitivity and scenarios. | | 10. Variant view | Market-implied expectations, evidence-based disagreement, catalyst/validation path, strongest disconfirming evidence. | | 11. Risk and stress | Operational, competitive, regulatory, accounting, liquidity, financing, governance, valuation, thesis-break cases. | | 12. Investment memo | Decision-useful conclusion; evidence; assumptions; valuation; catalysts; risks; what changes the view. | | 13. Monitoring system | Dated KPI triggers, filing alerts, event calendar, estimate errors, thesis breaks, model-change log, post-mortem schedule. | ## Execution gates and handoff standard - Gate 1, source completeness: no historical model work begins until filing periods, amendments, segment definitions, debt documents, share-count instruments, and material exhibits are version-controlled in the source archive. - Gate 2, accounting integrity: historical statements, segment/KPI bridges, cash roll-forward, debt, taxes, dilution, leases, pensions, and acquisition effects must reconcile before forecast assumptions are approved. - Gate 3, forecast causality: every material forecast line must trace to an economic driver or explicit accounting schedule. Unexplained plugs, hard-coded balance-sheet items, and circularity masked by copied values fail the build. - Gate 4, valuation independence: at least one intrinsic method and one expectations/market-relative method must be completed independently before a target value is discussed. - Gate 5, adversarial review: a reviewer who did not build the model must reproduce the top three value drivers, strongest disconfirming fact, largest downside mechanism, and EV-to-equity bridge from the archived workpapers. - Gate 6, monitoring handoff: each thesis variable receives a source, cadence, expected range, warning threshold, thesis-break threshold, owner, and required model action. - File-control standard: maintain dated folders for Sources, Historical Model, Forecast Model, Valuation, Memo, Monitoring, and Archive. Preserve the version used for every published conclusion so future analysts can recreate the decision state. - Final handoff package: one-page conclusion, full memo, integrated model, source log, assumption register, model-change log, risk register, monitoring dashboard, IC question responses, and unresolved-items list. ## Initiation completion test The initiation is complete only when a second analyst can start with the archived source pack, reproduce historicals and valuation within immaterial tolerance, explain every major assumption, and identify exactly what future evidence would trigger a model update or thesis break. # APPENDIX H - INVESTMENT MEMO TEMPLATE ## Decision-relevant conclusion State the analytical conclusion in three to five sentences. Separate business quality, expectations, valuation, and timing. ## Variant perception What does the current price appear to require? Which evidence supports a materially different operating outcome? ## Evidence table List each material claim, evidence type, primary source, date, confidence, and unresolved contradiction. ## Business economics Customers, value proposition, pricing, unit economics, cost structure, capital intensity, reinvestment runway, bottlenecks. ## Industry and competition Market size, value chain, share, substitutes, marginal price setter, cycle, disruption, regulation. ## Accounting and forensics Material policies, normalizations, non-GAAP reconstruction, working capital, reserves, auditor/control issues, dilution. ## Forecast Key operating drivers and why each assumption is reasonable. Show historical-to-forecast bridge and error ranges. ## Valuation Base/bull/bear/stress values, methods, reverse expectations, EV-equity bridge, terminal assumptions, sensitivities. ## Catalysts and validation Dated events or datapoints that should validate or falsify the thesis. Do not require a catalyst if value realization is long-duration, but define monitoring evidence. ## Risks and thesis breaks Transmission mechanism, leading indicator, financial exposure, time horizon, mitigant, and explicit break condition. ## What changes the view State the three most important future facts that would force a material model or thesis update. ## Required memo metadata - Company, ticker, date/time, analyst, reviewer, market price, diluted share count, enterprise value, base-case value range, stated time horizon, and source cut-off time. - Label every material statement as reported fact, management claim, analyst calculation, external estimate, or judgment. Link each material factual claim to the source log. - State the decision question and the one variable that would most change the answer if the analyst is wrong. ## One-page front sheet - Conclusion: three to five sentences, no chronology. State business quality, market-implied expectations, variant evidence, valuation range, and the principal reason the view could fail. - Key numbers: historical revenue/FCF/ROIC, next-two-year driver forecast, base/bull/bear value, liquidity, dilution, and the three most decision-relevant KPIs. - Variant table: market-implied assumption, analyst assumption, evidence, valuation sensitivity, validation date. - Risk table: mechanism, earliest indicator, financial exposure, probability/range, thesis-break threshold. ## Mandatory exhibits - Historical driver bridge and segment/KPI definition history. - Integrated forecast summary with revenue, margins, FCF, balance-sheet/liquidity and share-count bridge. - Valuation summary with DCF assumptions, relative normalization, reverse DCF, scenario matrix and EV-to-equity bridge. - Capital allocation chronology and management forecast-accuracy record where decision-relevant. - Disconfirming-evidence page containing the best bear argument, unresolved contradictions, and evidence that would force a thesis reset. ## Memo writing discipline - Use numbers to answer a decision question, not to decorate prose. Every chart must make a relationship or divergence visible faster than text. - Avoid adjectives such as strong, weak, attractive, conservative, or aggressive unless the memo defines the benchmark and quantifies the comparison. - Do not repeat management language as analysis. Translate narrative into units, price, mix, margins, cash, capital, and time. - Keep the decision layer concise and move raw evidence, detailed reconciliations, sensitivity tables, and source extracts to appendices. ## Illustrative conclusion pattern > Example structure: "The current price requires approximately X years of Y growth and Z terminal margin. Our evidence supports a lower/higher path because A, B, and C. Under base assumptions the equity value range is $___ to $___, with the largest downside transmission through ___. The view should be reconsidered if ___ occurs by ___." Replace every placeholder with sourced numbers. ## Memo exit standard A senior reviewer should be able to read the first page in under five minutes and identify the decision, market expectation, evidence edge, valuation, downside mechanism, and falsification trigger without opening the model. # APPENDIX I - EARNINGS PREVIEW / POST-EARNINGS TEMPLATE ## Before the print - Freeze the pre-earnings model and timestamp market price/consensus context. - Write the three decision variables that matter most and the expected value/range for each. - Define reaction cases based on operating drivers, not EPS beat/miss alone. - List thesis-break evidence and what would be noise. ## After the print - Reconcile reported statements to filing/source before changing forecasts. - Bridge revenue by price/volume/mix or the appropriate business driver. - Bridge margin by price/cost/mix/utilization and recurring versus temporary items. - Reconcile working capital, capex, cash, debt, shares, taxes, and guidance. - Update the model-change log with old/new assumptions and valuation effect. - Compare actual outcome with the precommitted expectation and record forecast error. ## Call/transcript review - Separate prepared narrative from Q&A evidence. - Track changes in wording, KPI emphasis, definition, confidence, and time horizon. - Do not treat management explanation as verified until it reconciles numerically or with external evidence. ## Required pre-earnings control sheet - Freeze a timestamped model and record company guidance, consensus, market price, option-implied move if available, and the analyst range for the three variables that matter most. - For each key variable write: expected range, bull threshold, bear threshold, source, accounting definition, and valuation sensitivity. Do not use EPS alone as the key variable when the business is driven by units, ARR, NIM, loss ratio, utilization, backlog, or another native KPI. - Pre-write the bull, base, bear and thesis-break operating cases so the analyst cannot move the goalposts after the print. ## Surprise bridge - Revenue surprise = reported revenue minus frozen-model revenue. Rebuild the difference through the native driver bridge, such as units x price x mix, beginning ARR + new - churn +/- expansion, or earning assets x spread. - Margin surprise = price/mix + volume/utilization + variable cost + fixed-cost absorption + one-time/accounting items. Identify the residual rather than labeling the entire difference execution. - FCF surprise = EBITDA/operating-profit surprise + working-capital difference + cash taxes + capex + restructuring/acquisition cash + other financing/operating differences. - Guidance surprise = compare the new range with the prior implied quarterly path, not only with the prior midpoint. ## First 120 minutes after the release - 0-20 minutes: capture filed statements and release, verify units/definitions, reconcile headline numbers and identify only the decision-relevant deltas. - 20-60 minutes: rebuild driver, margin, cash and guidance bridges; update scenarios without changing long-term assumptions that the new evidence did not address. - 60-120 minutes: review call/Q&A, update the assumption register, quantify valuation changes, classify each change as structural, cyclical, timing, accounting, or one-time, and write the post-earnings conclusion. - After 120 minutes: perform source QA, compare actuals with precommitted expectations, log forecast errors, and decide whether the thesis, only the numbers, or neither changed. ## Reaction matrix - Good print / stronger thesis: operating driver and cash corroborate; long-term value increases for an economic reason. - Good print / weaker thesis: headline beats but quality, working capital, retention, reserves, backlog, unit economics, or forward indicators deteriorate. - Bad print / intact thesis: miss is timing/noise and the causal driver remains within the precommitted range. - Bad print / broken thesis: a predeclared falsification condition is met. Do not protect the prior target by extending the horizon or changing the metric. ## Post-earnings exit standard The update is complete only when reported results tie to primary sources, the surprise is decomposed into economic drivers, the frozen-model forecast error is logged, the valuation impact is quantified, and the analyst explicitly states whether the thesis strengthened, weakened, broke, or remained unchanged. # APPENDIX J - 12-MONTH ANALYST TRAINING PROGRAM ## Month 1: SEC filings, source discipline, three statements, and historical reconstruction - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 2: Revenue, working capital, capex, taxes, SBC, and share count - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 3: Forensic accounting, non-GAAP, auditor reports, estimates, and red-flag logs - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 4: Business models, customer behavior, pricing, retention, and unit economics - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 5: Industry structure, market sizing, competitive benchmarking, cycle, and disruption - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 6: Integrated three-statement modeling and model quality control - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 7: DCF, relative valuation, reverse DCF, SOTP, and scenario analysis - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 8: Management, incentives, governance, ownership, and capital allocation - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 9: Earnings, guidance, M&A, restructuring, distress, and catalyst mapping - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 10: Alternative data, expert calls, channel work, scraping, and signal validation - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 11: Investment writing, IC defense, decision journaling, forecast-error analysis, and post-mortems - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. ## Month 12: Full-company initiation under senior review, including model, memo, sector map, risk register, monitoring, and oral defense - Complete one primary-source drill with a source log and definition reconciliation. - Complete one spreadsheet/model exercise with visible error checks and senior review. - Write one two-page variant-view memo separating fact, estimate, judgment, and falsification evidence. - Complete one error review describing what was initially wrong, why it was wrong, and what process control will prevent recurrence. - Defend the work orally against at least five adversarial questions and update the research based on valid objections. # APPENDIX K - SENIOR REVIEW AND INVESTMENT COMMITTEE QUESTION BANK ## Evidence - What is the single most decision-relevant primary-source fact? - Which material claim relies on the weakest source? - Which definition changed over time or differs most across peers? - What evidence is missing that could plausibly reverse the conclusion? ## Economics - What is the causal driver of value rather than the accounting proxy? - What constraint prevents the forecast from being achieved? - What is the incremental return on the next dollar of capital? - Which reported improvement is cyclical, timing-related, or accounting-only? ## Forecast - Which assumption contributes most to valuation sensitivity? - What is the historical forecast error for this variable? - What physical/customer/capacity reality does the forecast imply? - What combination of assumptions is internally inconsistent? ## Valuation - What percentage of value comes from the terminal period? - What growth and return assumptions are embedded in the current price? - Which EV-to-equity item is easiest to miss or double count? - What valuation method would a skeptical investor use and why? ## Risk - How does the largest risk transmit into cash and balance-sheet stress? - What is the earliest observable indicator? - What is the explicit thesis-break condition? - Could the company survive the stress without external capital? ## Behavior - What evidence are we discounting because it conflicts with the thesis? - Did the conclusion change more than the underlying evidence? - Are we anchoring to purchase price, prior target, consensus, or management guidance? - What would we conclude if we had no position and saw the evidence today? ## Additional red-team questions - Accounting: Which balance-sheet account would move first if management were pulling revenue forward or delaying expense? - Accounting: Which non-GAAP exclusion has recurred often enough to be economically normal? - Accounting: What assumption in reserves, useful lives, fair value, taxes, pension, or capitalization produces the largest earnings sensitivity? - Industry: Who sets the marginal price and what evidence shows that today? - Industry: Which capacity addition, substitute, standard, regulation, or technology shift could invalidate the market-size or margin framework? - Industry: Is the company gaining share because of structural advantage or because a competitor is supply constrained? - Management: Which capital-allocation decision created or destroyed the most per-share value over the last five years? - Management: Which compensation metric can be optimized while intrinsic value deteriorates? - Management: Where has management changed definitions, targets, or time horizons after a miss? - Model: If we delete the forecast tab and rebuild from physical/customer drivers, which assumptions change? - Model: Which forecast relies on capacity, sales headcount, customer adoption, regulatory approval, or financing that does not yet exist? - Model: What is the largest unexplained residual in the historical driver bridge? - Valuation: What terminal ROIC and reinvestment rate are implied by the terminal growth assumption? - Valuation: How much value remains if the competitive advantage period is cut in half? - Valuation: What happens to equity value after realistic refinancing, dilution, pension, minority interest and lease/debt-like claims? - Data: Which alternative-data signal has been backtested out-of-sample, and what is its false-positive rate? - Data: Could the data have survivorship, look-ahead, selection, geography, or definition bias? - AI: Which conclusions were produced or transformed by AI, and which deterministic or human verification step independently checked them? - AI: Can every AI-extracted fact be traced to exact source text and version? - Decision: What would make us reverse the view today, and are we actually monitoring it? ## IC pass/fail gate The recommendation does not pass senior review if the analyst cannot answer the highest-sensitivity assumption, strongest contrary evidence, liquidity/downside path, source provenance, and thesis-break question with a number or a specific evidence plan. "We will watch it" is not an acceptable answer without a source, cadence and threshold. # APPENDIX L - INTEGRATED WORKED COMPANY CASE > This fictional case is deliberately simplified enough to calculate by hand but complete enough to demonstrate the workflow from reported statements to operating drivers, cash flow, valuation, reverse expectations, scenarios, and monitoring. Numbers are illustrative and do not describe a real issuer. ## Case company: Orion Grid Systems Orion sells grid-control equipment and recurring monitoring/service contracts. Equipment is 75% of current revenue and service is 25%. Equipment has lower gross margin and consumes inventory/working capital; service has higher recurring margin. The company has $280 million of debt, $80 million of cash, and 50 million diluted shares. | Current-year income statement | $mm | Analyst note | | --- | --- | --- | | Equipment revenue | 750 | Model units, realized price, product mix, and capacity | | Service revenue | 250 | Model installed base, attach, renewal, price, and usage | | Total revenue | 1,000 | Reported revenue base | | COGS | 600 | 40% consolidated gross margin | | Gross profit | 400 | Reconcile equipment/service mix | | SG&A | 180 | Separate variable sales expense from fixed corporate cost | | R&D | 60 | Treat as operating investment analytically when assessing returns | | EBITDA | 160 | Before $40 D&A | | D&A | 40 | Linked to productive asset base | | EBIT | 120 | 12% EBIT margin | | Interest expense | 18 | Financing item, excluded from FCFF | | Pretax income | 102 | | | Tax expense at 25% | 25.5 | Simplified case | | Net income | 76.5 | Equity earnings, not the DCF starting point | | Current-year balance sheet | $mm | Current-year balance sheet | $mm | | --- | --- | --- | --- | | Cash | 80 | Accounts payable | 90 | | Accounts receivable | 150 | Accrued liabilities | 80 | | Inventory | 120 | Deferred revenue | 60 | | Other current assets | 50 | Debt | 280 | | PP&E, net | 500 | Other liabilities | 90 | | Goodwill/intangibles | 100 | Equity | 400 | | Total assets | 1,000 | Total liabilities + equity | 1,000 | ## 1. Statement and cash reconciliation - Net debt is $200 million: $280 million debt less $80 million cash. Whether all cash is excess cash must still be tested; the simplified case assumes it is available. - Assume current-year CFO is $106.5 million: $76.5 million net income + $40 million D&A + $10 million other noncash/SBC - $20 million working-capital use. Capex is $60 million, producing $46.5 million simplified equity free cash flow before debt repayment/buybacks/dividends. - A $10 million credit sale at 60% gross margin initially increases revenue by $10 million, EBIT by $6 million before other costs, receivables by $10 million, inventory/COGS according to shipment cost, retained earnings by after-tax profit, and CFO by less than net income until cash is collected. This is why revenue and cash cannot be analyzed independently. - The reported statements are the immutable layer. Any normalization for SBC, restructuring, acquisition costs, or unusual items must appear in a separate analyst layer with a reconciliation. ## 2. Working-capital diagnostic | Measure | Calculation | Result | Interpretation | | --- | --- | --- | --- | | DSO | $150 AR / $1,000 revenue x 365 | 54.8 days | Collection/billing intensity | | Inventory days | $120 inventory / $600 COGS x 365 | 73.0 days | Manufacturing/channel inventory intensity | | DPO | $90 AP / $600 COGS x 365 | 54.8 days | Supplier financing | | Cash conversion cycle | 54.8 + 73.0 - 54.8 | 73.0 days | Simplified working-capital cycle | | Five-day DSO deterioration | $1,000 / 365 x 5 | $13.7mm cash use | Translate KPI movement into cash | - If revenue rises 12% but receivables rise 25%, do not conclude manipulation. Test billing timing, customer mix, payment terms, milestones, acquisition scope, and collections first. - If DPO rises sharply and CFO improves, inspect supplier-finance disclosures and payment terms before treating the cash improvement as sustainable. - If inventory days rise while management says demand is exceptionally strong, compare backlog, cancellations, lead times, channel inventory, and production plans. ## 3. Driver forecast The simplified base case assumes revenue growth decelerates from 12% to 5%, EBIT margin rises from 12.5% to 14%, D&A stays at 4% of revenue, capex at 5.5% of revenue, net operating working capital at 12% of revenue, and cash tax at 25%. The operating model should normally be built from equipment units/price and service installed-base economics; consolidated growth is shown here only to keep the arithmetic compact. | Year | Revenue | Growth | EBIT margin | NOPAT | FCFF | | --- | --- | --- | --- | --- | --- | | 1 | $1,120.0 | 12.0% | 12.5% | $105.0 | $73.8 | | 2 | $1,232.0 | 10.0% | 13.0% | $120.1 | $88.2 | | 3 | $1,330.6 | 8.0% | 13.5% | $134.7 | $102.9 | | 4 | $1,410.4 | 6.0% | 14.0% | $148.1 | $117.4 | | 5 | $1,480.9 | 5.0% | 14.0% | $155.5 | $124.8 | - Year 1 FCFF = NOPAT $105.0 + D&A $44.8 - capex $61.6 - increase in NWC $14.4 = $73.8 million. - The forecast must be rebuilt if equipment/service mix changes because consolidated margin, capex, inventory, and deferred revenue would no longer scale mechanically with total revenue. - A forecast is not validated because the spreadsheet balances. Validate units, customer demand, price, capacity, lead times, and reinvestment against external and issuer evidence. ## 4. DCF construction and terminal consistency - Use WACC of 9.0%, a five-year explicit period, and midyear discounting for illustration. The explicit FCFF present value is approximately $402.7 million. - Terminal growth is 3.0% and terminal ROIC is 12.5%. The sustainable-growth identity requires a 24% reinvestment rate: 3.0% / 12.5% = 24%. - Year 6 terminal NOPAT is approximately $160.2 million. Terminal FCFF is $160.2 x (1 - 24%) = approximately $121.7 million. - Terminal value at the end of Year 5 is $121.7 / (9.0% - 3.0%) = approximately $2,028.7 million. Discounted at the midyear timing used in this case, its present value is approximately $1,376.6 million. - Enterprise value is therefore approximately $1,779.2 million. Subtract $200 million net debt to get $1,579.2 million equity value. Divide by 50 million diluted shares for approximately $31.58 per share. - The terminal value contributes most of enterprise value, so the analyst must explicitly stress terminal margin, reinvestment/ROIC, growth, and WACC. A precise $31.58 output does not imply precise intrinsic value. | DCF bridge | $mm except per share | | --- | --- | | PV of explicit FCFF | 402.7 | | PV of terminal value | 1,376.6 | | Enterprise value | 1,779.2 | | Less net debt | (200.0) | | Equity value | 1,579.2 | | Diluted shares | 50.0 | | Illustrative value/share | $31.58 | ## 5. Reverse expectations - Assume Orion trades at $25 per diluted share. Equity value is $1,250 million and, adding $200 million net debt, enterprise value is $1,450 million. - Holding the same growth path, reinvestment mechanics, WACC, and terminal assumptions, an approximately 11.44% constant EBIT margin would reproduce a $1,450 million enterprise value in this simplified model. - The research question is therefore not merely whether a $31.58 DCF is above $25. The analyst must determine whether evidence supports an operating path materially better than the roughly 11.4% margin economics embedded in the simplified reverse case, after considering uncertainty and alternative assumptions. - Reverse DCF conclusions should be translated into equipment units, service attach/retention, price, mix, capacity, and cost evidence before becoming a thesis. ## 6. Scenario design | Scenario | Operating assumptions | WACC / terminal | EV | $ / share | | --- | --- | --- | --- | --- | | Bull | Growth 14% -> 7%; EBIT margin -> 16% | 8.5% / 3.25% | $2,516.7 | $46.33 | | Base | Growth 12% -> 5%; EBIT margin -> 14% | 9.0% / 3.0% | $1,779.2 | $31.58 | | Bear | Growth 8% -> 3%; EBIT margin near 11% | 10.0% / 2.5% | $976.8 | $15.54 | | Stress | Flat/negative early growth; EBIT margin trough 7% | 11.5% / 2.0% | $553.0 | $7.06 | - The scenarios alter operating drivers and discounting coherently. They do not simply apply higher or lower valuation multiples. - Probability weights, if used, must be shown separately from the unweighted outcomes and must not hide a liquidity or dilution pathway. - If Orion needed external capital in the stress case, the $7.06 value would be too high unless the share count and financing terms were updated. Scenario integration must include the balance sheet. ## 7. Forensic overlay - Assume management reports adjusted EBIT that excludes $10 million of recurring SBC and $8 million of integration expense for the third consecutive year. Rebuild the adjustment history before accepting either exclusion as nonrecurring. - Assume inventory rises 30% while revenue rises 12%. Test planned capacity ramp, long-lead safety stock, product transition, cancellations, channel inventory, and obsolescence reserve before deciding whether the build is healthy or a warning. - Assume gross margin improves 200 bps while CFO weakens. Reconcile receivables, inventory, deferred revenue, supplier terms, and capex before calling the margin improvement high quality. - Assume the company switches from backlog to a new bookings KPI after backlog conversion slows. Reconstruct both series and treat definition drift itself as an open diligence item. ## 8. Mini investment-memo conclusion A disciplined conclusion would say: Orion can create value if recurring service mix and equipment utilization lift EBIT margin toward 14% while growth remains supported by real grid demand and capacity. The simplified base DCF is $31.58 per share versus a hypothetical $25 market price, but most value resides in the terminal period and the reverse case shows that the market can be rationalized with roughly 11.4% EBIT margin under the same other assumptions. The decisive research work is therefore evidence on service retention/attach, equipment capacity and pricing, working-capital quality, and reinvestment returns, not the arithmetic spread between $31.58 and $25. ## 9. Monitoring sheet - Quarterly equipment units, realized price, backlog conversion, cancellation rate, capacity utilization, and gross margin. - Service installed base, attach, renewal/gross retention, expansion, realized price, and service gross margin. - DSO, inventory days, DPO, deferred revenue, capex, and maintenance/growth split. - Net debt, interest cost, covenant headroom, SBC/dilution, repurchases, and cash taxes. - Explicit thesis break: evidence that normalized EBIT margin cannot exceed the market-implied range without unsustainable working-capital, underinvestment, or customer deterioration. # APPENDIX M - FINAL 99-POINT SELF-AUDIT | Dimension | 99-level requirement | Weight | | --- | --- | --- | | Breadth | Complete end-to-end research system plus sector adaptation and reusable deliverables | 10 | | Source discipline | Primary-source-first, versioned, citable, reproducible evidence trail | 10 | | Accounting/forensics | Three-statement mechanics, complex accounting, non-GAAP, estimates, fraud screens, cash corroboration | 10 | | Business/industry | Unit economics, pricing, retention, capital intensity, industry structure, competition, disruption | 10 | | Modeling | Driver-based integrated statements, schedules, scenarios, QA, no hidden plugs | 12 | | Valuation | DCF, relative, reverse DCF, SOTP, terminal consistency, EV-equity bridge, dilution | 12 | | Risk/decision process | Liquidity, stress, thesis breaks, pre-mortems, post-mortems, decision journal | 10 | | Sector specificity | KPIs, traps, valuation, bottlenecks, bear cases, model tabs for major sectors | 10 | | AI/automation | Provenance, structured extraction, deterministic compute, evals, security, human approval | 8 | | Teachability | Prerequisites, inputs, procedures, formulas, examples, failure modes, exit criteria, training path | 8 | > Target: no dimension below 99% of its intended standard. A static manual should not claim absolute 100% coverage because regulations, technologies, products, and market structures evolve. The required regulatory-change monitor and source-verification system are therefore part of the mastery standard. ## Scoring method - Score each dimension from 0 to 100 using documented evidence. Multiply by the published weight and sum the weighted results. - 99-level means the dimension is executable by another trained analyst without hidden knowledge, material unsupported claims, unexplained model plugs, or missing validation/falsification logic. - No averaging around a critical defect: any automatic-fail condition below caps the entire manual or research package below 70 until corrected. - Reviewers must cite the page/module, defect, economic consequence, remediation and retest result for every score below 99. ## Automatic score caps - Cap below 70 if a material factual claim cannot be traced to a source, historical financials do not reconcile, the balance sheet does not balance, or the EV-to-equity bridge omits a material claim. - Cap below 70 if valuation depends on an internally inconsistent terminal growth/ROIC/reinvestment assumption, an unexplained plug, or a forecast that cannot be translated into business-native operating units. - Cap below 70 if the work uses alternative data or AI-generated extraction without provenance, validation and human review sufficient to reproduce the conclusion. - Cap below 70 if the downside case ignores liquidity, refinancing, covenant, collateral, dilution, or priority-of-claims issues that could dominate common-equity value. - Cap below 70 if sector analysis uses generic revenue/margin assumptions while omitting the sector-native driver, accounting trap, valuation claimant level, or binding constraint. ## 99-point remediation loop 1. 1. Identify every dimension or module below 99 and rank by decision impact, not convenience. 1. 2. Convert the weakness into a concrete deliverable: missing source, reconciliation, formula, worked example, stress case, source map, or review control. 1. 3. Repair the underlying analytical procedure, not merely the prose describing it. 1. 4. Re-run the calculation/model/source QA and have an independent reviewer reproduce the result. 1. 5. Re-score using the same rubric. Repeat until no material dimension is below 99 or document the reason the standard is genuinely not applicable. ## Final certification evidence - Completed scorecard with reviewer name/date and links to evidence. - List of all defects found below 99 and the exact remediation performed. - Re-rendered and visually inspected final document with accessibility and table-header checks complete. - Change log that records content added, removed, consolidated, or corrected during remediation. # PART XVI - MASTERY LABORATORIES - Primary sources before narratives. - Economics before accounting presentation. - Cash, balance sheet, and dilution before adjusted earnings. - Drivers and constraints before line-item extrapolation. - Reinvestment and incremental returns before growth worship. - Ranges and scenarios before false precision. - Disconfirming evidence before conviction. - Liquidity and solvency before multiple compression. - Process quality before outcome worship. - Versioned evidence and model changes before memory. - Automation for speed and consistency, human accountability for judgment. > The following mastery laboratories convert the operating system into zero-to-expert execution. Use them when a core module identifies what must be done but the analyst needs the mechanics, equations, model architecture, worked examples, edge cases, or reviewer tests required to do it correctly. # MASTER LAB 01 - ACCOUNTING MECHANICS, JOURNAL ENTRIES, AND FOOTNOTE RECONSTRUCTION > Purpose: make the analyst capable of reconstructing the accounting from the transaction level upward. The analyst should be able to explain not only what a reported number is, but what journal entries created it, which estimates sit behind it, where the cash moved, what will reverse, and how the item changes valuation. ## Required outputs - A clean historical three-statement reconstruction tied to audited filings and amendments. - A transaction-level bridge for every material accounting area that can alter normalized earnings, cash flow, invested capital, or share count. - A footnote roll-forward workbook for debt, leases, stock compensation, taxes, goodwill and intangibles, pensions, reserves, and commitments. - A GAAP-versus-economic adjustment schedule that preserves both reported accounting and the analyst normalization. - A list of unresolved accounting judgments with sensitivity ranges and source citations. ## Prerequisites - Basic algebra and percentage change. - Understanding of debit and credit mechanics at the level of assets, liabilities, equity, revenue, expense, gains, and losses. - Ability to locate primary statements and footnotes in a 10-K or equivalent annual filing. ## 1. Build the accounting map before reading ratios Start with the transaction, not the ratio. A ratio can look unusual for many reasons. A transaction map asks what the company sold, what it promised, when control transferred, when cash moved, what asset or liability was created, what estimate was required, and what future reversal is expected. For every material line item, create a six-column accounting map: economic event, debit, credit, income-statement effect, cash-flow-statement effect, and balance-sheet carryforward. | Economic event | Typical debit | Typical credit | Income statement | Cash flow | Future analytical question | | --- | --- | --- | --- | --- | --- | | Cash sale of product | Cash | Revenue | Revenue and related COGS recognized | Operating inflow | Is gross margin sustainable and are returns and rebates accrued correctly? | | Sale on credit | Accounts receivable | Revenue | Revenue recognized before cash | No collection yet | Are receivables growing faster than sales and why? | | Cash collected in advance | Cash | Contract liability / deferred revenue | No revenue until performance | Operating inflow | What portion is refundable, cancellable, or tied to future service cost? | | Capital equipment purchase | PP&E | Cash / payable | Depreciation later | Investing outflow | Maintenance or growth capex, useful life, residual value, capacity added? | | Stock award vesting | Compensation expense / equity mechanics | APIC / equity | SBC expense | Usually noncash add-back in CFO, tax effects separate | What is the true dilution and repurchase offset? | | Acquisition | Assets and goodwill | Cash, debt, or equity | Future amortization, integration costs, step-up effects | Investing and financing mix | How much value depends on synergies and purchase accounting? | ## 2. Trial balance to financial statements A novice often learns the three statements as separate reports. An expert treats them as different views of one ledger. Reconstruct at least one quarter from a simplified trial balance. Confirm that every income-statement item closes into retained earnings, every noncash income-statement item has a balance-sheet counterpart, and every cash movement is classified once in the cash flow statement. Core identity: Assets = Liabilities + Equity. Retained earnings ending = Retained earnings beginning + Net income - Common dividends +/- other direct retained-earnings adjustments. Cash ending = Cash beginning + CFO + CFI + CFF + FX or other cash translation effects. Do not plug cash to make a model balance. If the model does not balance, find the missing accounting link. Common causes are omitted retained earnings, incorrect share repurchase treatment, debt issuance or repayment omitted from financing cash flow, acquisition consideration not bridged, lease principal classified incorrectly, or a sign error in working capital. ### Mini-ledger drill 1. Record a $100 credit sale with $60 cost of goods sold. 1. Collect $70 of the receivable. 1. Buy $40 of inventory for cash. 1. Buy $120 of equipment with $20 cash and $100 debt. 1. Record $10 depreciation. 1. Pay $5 interest and accrue $3 of unpaid interest. 1. Repurchase $15 of stock and issue $4 of stock compensation. After the seven entries, rebuild the income statement, balance sheet, and cash flow statement. The drill is complete only when the balance sheet balances without a plug and the change in cash equals the cash flow statement. ## 3. Revenue recognition from contract economics Revenue is not a demand metric by default. The analyst must separate contract signing, order intake, backlog or remaining performance obligation, billing, cash collection, delivery, acceptance, and revenue recognition. For each material revenue stream, document the performance obligation, transaction price, variable consideration, timing of transfer, cancellation rights, refund rights, renewal behavior, principal-agent presentation, and significant financing components. | Signal | What it measures | Can lead revenue? | Can mislead? | Required reconciliation | | --- | --- | --- | --- | --- | | Bookings / order intake | Signed or accepted orders under company definition | Often | Yes, cancellations, nonbinding orders, multi-year scope | Bookings to backlog or RPO to recognized revenue | | Billings | Invoice issuance | Sometimes | Yes, billing terms can change | Billings to receivable or deferred revenue to cash | | Cash collections | Cash received | Sometimes | Yes, prepayments can precede delivery | Cash to contract liability or receivable reduction | | RPO / backlog | Contracted future performance under issuer definition | Usually | Yes, cancellability and duration vary | Opening balance + additions - recognized or cancelled = ending balance | | GAAP revenue | Performance recognized under accounting rules | No, it is the reported result | Yes as a demand proxy | Revenue to cash, receivables, contract assets and liabilities, units and price | For over-time recognition, identify the measure of progress and test whether it can be influenced by cost estimates, milestones, or front-loaded inputs. For point-in-time recognition, identify the control-transfer trigger and whether acceptance clauses, shipment terms, installation, or customer-specific customization delay recognition. For marketplaces and resellers, test gross-versus-net presentation because a presentation change can make reported revenue growth look spectacular while gross profit and economics barely change. ### Revenue quality bridge Organic revenue growth = Reported growth - acquisition/divestiture scope effect - FX effect - presentation/definition effect. Contract asset intensity = Contract assets / Revenue. Rising intensity deserves explanation when revenue growth is strong. Deferred revenue conversion = Revenue recognized from opening deferred revenue / Opening deferred revenue, when disclosed and meaningful. A revenue conclusion is not finished until it reconciles to at least one physical or contractual driver such as units shipped, seats, customers, megawatts, transactions, subscriptions, procedures, tonnage, capacity, or price per unit. ## 4. Working capital as an operating model, not a ratio plug Build working capital from the commercial process. Receivables depend on sales timing, billing, payment terms, customer mix, disputes, and collections. Inventory depends on demand, lead times, safety stock, production cycle, product mix, obsolescence, and channel strategy. Payables depend on procurement, supplier terms, bargaining power, purchase timing, and financing programs. DSO = Average accounts receivable / Revenue x Days in period. Inventory days = Average inventory / COGS x Days in period. DPO = Average accounts payable / COGS or purchases x Days in period, using the denominator that best matches the liability. Cash conversion cycle = DSO + Inventory days - DPO. Do not mechanically forecast each balance as a percentage of revenue if the business mix is changing. A hardware-plus-service company, a marketplace shifting to gross presentation, or a manufacturer entering a build phase can break historical ratios. Use operational drivers where available and use ratios only as a cross-check. ### Supplier finance and factoring Supplier finance can improve reported operating cash flow by extending payment timing while economically introducing financing. Factoring can accelerate cash collections while transferring or retaining credit risk. Read the footnotes for program size, outstanding balances, classification, recourse, fees, and cash-flow presentation. Recast when necessary so operating cash generation is comparable through time. ## 5. Capex, depreciation, useful lives, and asset intensity Separate maintenance capex, growth capex, capitalized software or development, construction in progress, acquired assets, and lease-related asset additions. The maintenance-versus-growth split is an analytical estimate, not a GAAP line. Use physical capacity, age of asset base, depreciation, management disclosures, replacement cycles, and peer economics to constrain the estimate. Net PP&E ending = Net PP&E beginning + capital expenditures + acquired assets + FX/other - depreciation - disposals - impairments. Gross investment intensity = Capex / Revenue. Incremental capital intensity = Change in invested operating assets / Change in revenue, measured over a sensible multi-year window. Useful-life changes can shift earnings without changing near-term cash. Compare disclosed useful lives through time and against peers. If depreciation falls because useful lives are extended, show the counterfactual expense under the prior life. When a company claims asset-light economics, test whether capital expenditure has simply moved into leases, supplier financing, cloud commitments, or acquisition spending. ## 6. Leases and debt-like commitments Create a lease schedule that separates operating lease cost, finance lease interest, amortization, lease liabilities, right-of-use assets, principal payments, and undiscounted future commitments. For valuation, decide consistently whether leases are treated as operating obligations embedded in EBITDA or as debt-like claims with corresponding earnings adjustments. Do not subtract lease debt from enterprise value while also using an EBITDA denominator that already bears full rental expense without considering comparability. Approximate lease debt sensitivity: present value of contractual lease payments discounted at a rate consistent with lease risk, when a full reported liability is unavailable. Also inventory purchase commitments, take-or-pay contracts, cloud commitments, minimum volume commitments, guarantees, securitizations, receivable sales, and off-balance-sheet joint-venture obligations. A liquidity model that includes only funded debt is incomplete. ## 7. Stock-based compensation and diluted share economics Treat stock compensation in two separate dimensions. First, it is compensation expense that can be economically recurring even when management excludes it from adjusted earnings. Second, it creates dilution or consumes cash when the company repurchases shares to offset dilution. Analyze both. Net dilution rate = (Ending diluted shares - Beginning diluted shares - shares issued for acquisitions or capital raises where separately analyzed) / Beginning diluted shares. Buyback offset ratio = Shares repurchased / Gross shares issued or vested from employee equity programs, when share data permit. Build an award inventory for options, RSUs, PSUs, employee stock purchase plans, and convertibles. For options, use treasury-stock-method mechanics where appropriate. For performance awards, model probable payout ranges. For convertibles, test if-converted treatment and capped-call effects where applicable. The goal is a forward share count that responds to price and award assumptions rather than a flat historical diluted share number. ## 8. Income taxes, cash taxes, and deferred tax mechanics Separate statutory tax rates, reported effective tax rate, cash taxes, deferred tax expense or benefit, NOLs, tax credits, valuation allowances, uncertain tax positions, discrete items, geographic mix, and withholding taxes. A low effective rate can be durable, temporary, or purely accounting. Tax provision = Current tax expense + Deferred tax expense, subject to jurisdictional and classification detail. NOPAT for valuation should use a normalized operating cash-tax concept, not automatically the reported effective tax rate. Create a tax-rate bridge by jurisdiction and item. Review valuation allowance changes carefully because releasing a valuation allowance can create a large GAAP tax benefit without an equivalent current-period cash inflow. For NOLs and credits, track jurisdiction, amount, expiration, limitations, and expected utilization. If a transaction changes ownership or legal structure, assess whether tax assets remain usable. ## 9. Acquisitions, goodwill, intangibles, and purchase accounting Rebuild each material acquisition from purchase price to assets acquired and liabilities assumed. Separate consideration paid in cash, debt, equity, contingent consideration, rollover equity, and assumed obligations. Identify identifiable intangibles, useful lives, deferred tax effects, step-ups, restructuring reserves, and goodwill. Goodwill = Consideration transferred + fair value of prior or noncontrolling interests - fair value of identifiable net assets acquired. For analysis, create an acquisition cohort table with purchase date, purchase price, revenue or EBITDA acquired, initial synergy claims, actual integration progress, impairments, disposals, and cumulative return on acquired capital. This prevents repeated acquisitions from hiding weak organic economics. Amortization of acquired intangibles is noncash in the current period but reflects a real acquisition price paid for finite-lived assets. Whether to add it back depends on the analytical question. Never label it economically irrelevant solely because it is noncash. ## 10. Reserves, estimates, warranties, returns, litigation, and restructuring Any estimate that lets management recognize cost earlier or later deserves a roll-forward. Build beginning reserve + provisions - cash or use + acquisitions, FX, or reclasses = ending reserve. Compare provision rates with underlying activity. A declining warranty reserve while unit volumes and defect indicators rise is different from a declining reserve after genuine product quality improvement. | Estimate area | Key management judgment | Analyst cross-check | Red flag | | --- | --- | --- | --- | | Credit loss | Default probability, loss severity, macro scenario | Charge-offs, delinquency, vintage performance | Reserve ratio falls as credit metrics worsen | | Warranty | Failure rate, repair cost, product mix | Claims, installed base, recalls, field data | Expense rate falls before evidence of quality improvement | | Returns / rebates | Return rate, channel inventory, rebate behavior | Sell-through, discounting, channel checks | Reserve rate falls while promotions rise | | Litigation | Probability and estimable loss range | Court filings, regulator actions, settlements | Repeated not-estimable language despite advancing case | | Restructuring | Scope, timing, eligible costs | Cash payments, headcount, facility closures | Recurring restructuring excluded from adjusted earnings every year | ## 11. Pension and postretirement obligations For defined-benefit plans, separate service cost, interest cost, expected or actual asset return, actuarial gains and losses, contributions, benefit payments, funded status, discount rate, asset allocation, and expected future contributions. The accounting can obscure economic leverage. Funded status = Fair value of plan assets - Projected benefit obligation, subject to the exact plan definition. Stress discount rates and asset returns, then estimate cash-contribution implications. In enterprise-to-equity valuation, avoid double counting pension deficits if related expense adjustments already change cash flow. For companies with large plans, include pension sensitivity in balance-sheet stress testing. ## 12. Foreign currency and translation Separate transaction exposure from translation exposure. Revenue translated at a weaker or stronger currency can change reported growth without changing local-currency demand. Transaction exposure can affect gross margin or operating expense depending on where costs and revenues are denominated. Constant-currency growth should be reconstructed from disclosed local-currency or FX bridge data when possible, not accepted as a black-box non-GAAP metric. Track balance-sheet translation in accumulated other comprehensive income where relevant. For companies with debt in multiple currencies, test whether hedges match the underlying exposure and whether hedging gains and losses are operating or financing in economic substance. ## 13. Cash flow statement reconstruction Reconcile CFO from net income through every material noncash item and working-capital movement. Then compare CFO to a direct cash logic: cash collected from customers minus cash paid to suppliers, employees, taxes, and other operating uses. The indirect and direct perspectives should tell the same economic story. Review classification elections and changes. Interest, taxes, restricted cash, discontinued operations, supplier finance, factoring, acquisition-related costs, and lease payments can create comparability issues across companies and accounting regimes. Normalize for analytical comparability while preserving the reported statement. Free cash flow is not a single authoritative GAAP measure. Define the exact construction every time and reconcile it to reported cash flows. A useful analytical stack is CFO, CFO before working-capital changes, capex, maintenance capex estimate, growth capex estimate, cash restructuring, acquisition spending, SBC-related tax and cash effects, and owner-oriented free cash flow. Each answers a different question. ## 14. U.S. GAAP and IFRS comparability Do not assume two companies using different accounting frameworks are directly comparable because the line-item names look similar. Focus on economic differences that can affect margins, assets, cash flows, leverage, and valuation. Common areas requiring explicit review include development-cost capitalization, inventory costing, impairment reversals, lease presentation, pension accounting, revaluation options, and cash-flow classification. Build a comparability bridge only for material differences. The goal is not to rewrite one company into the other accounting framework. The goal is to prevent an accounting-policy difference from being mistaken for an economic advantage. ## 15. Footnote reconstruction protocol 1. Create a source index containing filing date, reporting period, accession or document identifier, amendment status, page or section, and link. 1. For each material note, capture the current table and at least two comparable prior periods or prior filing versions. 1. Standardize units, signs, fiscal periods, and definitions before calculating trends. 1. Build roll-forwards for balances that change through additions, usage, remeasurement, acquisitions, FX, and reclassification. 1. Tie every roll-forward ending balance to the balance sheet or disclosed subtotal. Treat unexplained differences as open issues. 1. Mark each input as reported fact, analyst calculation, management claim, external estimate, or judgment. 1. Document every normalization separately from the reported accounting so another analyst can restore the original data. 1. Write one sentence explaining how the footnote could change revenue, margin, cash flow, invested capital, financing, dilution, risk, or valuation. ## 16. Accounting mastery exercises ### Exercise A: deferred revenue versus demand A software company grows revenue 18%, billings 8%, deferred revenue 3%, and RPO 20%. Build at least four explanations that could produce the pattern. Examples include longer contract duration, billing-frequency change, usage revenue mix, acquisition scope, renewal timing, or weakening new bookings. Identify the additional disclosure required to distinguish them. Do not choose a narrative before reconciling definitions. ### Exercise B: inventory and margins A hardware company reports gross margin expansion while inventory days rise from 70 to 115. Reconstruct price-volume-mix, purchase commitments, obsolescence reserves, channel inventory, expedite cost, and manufacturing utilization. Quantify a scenario in which margin remains strong but cash conversion deteriorates, and another in which inventory is an intentional buffer ahead of constrained supply. ### Exercise C: acquisition economics A serial acquirer reports 12% adjusted EPS growth, 2% organic revenue growth, rising goodwill, and recurring restructuring add-backs. Build an acquisition cohort analysis. Recalculate cumulative invested capital, organic versus acquired operating profit, cash acquisition spending, SBC and dilution, and returns on acquired capital. Determine whether EPS accretion reflects value creation or financing and accounting optics. ### Exercise D: cash tax normalization A company reports a 9% effective tax rate because of stock-compensation windfalls and valuation-allowance release. Build a normalized tax range using jurisdiction mix, statutory rates, durable credits, and expected NOL utilization. Show the impact on NOPAT, FCFF, and DCF value under reported and normalized tax assumptions. ## 17. Accounting completion gate - Every material balance can be traced to a filing, footnote, roll-forward, or explicit analyst estimate. - Every normalization has a reversible bridge to reported accounting. - The three statements reconcile without unexplained plugs. - Revenue, cash, working capital, capex, debt, tax, and share count are linked through transaction mechanics. - At least one skeptical accounting interpretation has been quantified for every thesis-relevant judgment. - The reviewer can change one key accounting assumption and observe the full effect on earnings, cash flow, invested capital, and valuation. # MASTER LAB 02 - INTEGRATED THREE-STATEMENT MODEL BUILD FROM RAW FILINGS > Purpose: teach the complete build process for an auditable public-company model. The target is not a pretty spreadsheet. The target is a model in which a reviewer can trace every historical number to a source, every forecast to an economic driver, every statement through double-entry logic, every scenario through a small set of explicit assumptions, and every valuation output back to the operating model. ## Required model package - Source log and assumptions register. - Historical income statement, balance sheet, cash flow statement, and segment/KPI history. - Driver schedules for revenue, margins, working capital, capex, depreciation, taxes, debt, cash, share count, and material non-operating items. - Integrated forecast statements with automated balance and cash checks. - Base, bull, bear, stress, and thesis-break scenarios that change operating drivers rather than only valuation multiples. - Valuation tabs linked to the operating model, plus a reverse-expectations view. - Visible model checks, change log, and reviewer sign-off block. ## 1. Workbook architecture and non-negotiable conventions Use a modular architecture. The exact tab names may change, but the separation of raw source data, standardized historicals, assumptions, operating schedules, statements, valuation, scenarios, and checks should remain clear. Avoid a single giant sheet where inputs, calculations, and outputs are mixed. | Layer | Typical tabs | Rule | | --- | --- | --- | | Control | Cover, Source Log, Assumptions, Change Log | Every material input has owner, source, date, unit, period, and status. | | Raw data | 10-K, 10-Q, XBRL extracts, KPI extracts | Do not overwrite source data with analyst normalization. | | Normalized history | Historicals, Segments, KPIs | Reported and adjusted views reconcile through explicit bridges. | | Operating schedules | Revenue, Costs, WC, Capex, Tax, Debt, Shares | Forecast from drivers and roll-forwards. | | Statements | IS, BS, CF | No hard-coded forecast subtotals that bypass schedules. | | Scenarios | Scenario assumptions, sensitivity matrix | One scenario switch or explicit scenario set controls the forecast. | | Valuation | DCF, Multiples, Reverse DCF, SOTP if needed | Valuation pulls from statements or schedules, never duplicate hard-coded forecasts. | | Quality control | Checks, Diagnostics | All material reconciliations visible and designed to fail loudly. | Use consistent sign conventions. One workable standard is positive revenue and expenses shown as negative only where the model logic requires it, or all statement lines shown in presentation sign with cash-flow formulas handling direction explicitly. Whichever standard is chosen, document it and never mix conventions silently. Separate inputs from formulas visually using your organization standard, but do not rely on color as the only control. A reviewer should identify an input from cell comments, labels, named ranges, or a dedicated assumption block. Every forecast formula should have a readable causal path. ## 2. Build the source layer before the forecast 1. Record the filing and reporting period for every historical column. 1. Capture the audited annual statements first, then reconcile interim periods to annual totals. 1. Store raw values in reported units and retain the filing definition. 1. Map source line items into standardized model line items through a mapping table rather than rewriting the source. 1. Create an exceptions log for recasts, discontinued operations, segment changes, fiscal-year changes, acquisitions, and accounting-policy changes. 1. Lock the reported historical layer after reconciliation so later analyst adjustments cannot silently replace it. ### Source mapping table | Source label | Model label | Source period | Unit | Normalization | Check | | --- | --- | --- | --- | --- | --- | | Net sales | Revenue | FY2025 | USD mm | None | Ties to statement | | Revenue - Product | Product revenue | FY2025 | USD mm | Segment recast if needed | Segments sum to consolidated | | Share-based compensation | SBC | FY2025 | USD mm | Reclassified by function if disclosed | Matches cash-flow add-back and footnote | | Capital expenditures | Capex | FY2025 | USD mm | Separate acquisitions and finance leases | Matches PP&E roll-forward where possible | ## 3. Historical normalization and definition control Never force historical comparability by deleting reported history. Maintain a reported series and a constant-definition analytical series. If the company recasts segments, changes KPI definitions, acquires a major business, changes fiscal year, or shifts gross-versus-net revenue presentation, document both the company recast and your own analytical bridge. Constant-definition growth = Comparable current-period measure / Comparable prior-period measure - 1. For each adjustment, record whether it changes revenue, operating profit, cash flow, invested capital, or only presentation. The model should be able to reproduce published numbers before any analyst normalization is applied. ## 4. Revenue build: forecast the economic engine Choose the smallest set of drivers that explains the business without hiding risk. A model with 200 product rows can be less useful than a 10-driver model if the 200 rows simply extrapolate historical growth. The correct granularity is the level at which pricing, volume, capacity, customer behavior, or mix can change differently and materially affect value. | Business type | Primary revenue drivers | Useful cross-checks | | --- | --- | --- | | Subscription software | Beginning ARR, new ARR, churn, expansion, price, FX | RPO, billings, deferred revenue, customer count, NRR | | Semiconductor | Units, wafer starts/capacity, yield, ASP, mix, utilization | Inventory, lead times, design wins, foundry capacity, end-market sell-through | | Industrial project | Backlog, bookings, conversion, price, scope, execution timing | Book-to-bill, project milestones, working capital, capacity | | Marketplace | GMV or transactions, take rate, ads/services, buyer/seller activity | Payment volume, active users, cohort behavior, gross-versus-net presentation | | Bank | Average earning assets, loan growth, securities, yields, deposits, funding cost | NIM, deposit beta, credit losses, capital | | Commodity producer | Volume, grade/yield, realized price, benchmark differentials | Reserves, cost curve, hedges, sustaining capex | ### Subscription revenue mechanics Ending ARR = Beginning ARR + New ARR + Expansion ARR - Churn ARR - Contraction ARR + FX or scope effects. Approximate subscription revenue depends on average in-period ARR and revenue-recognition timing, not simply ending ARR. Use cohort or installed-base logic when retention is central. For usage-based models, separate committed minimums from variable consumption. For price increases, model realized price after discounting, seat consolidation, downgrade behavior, and churn response rather than applying list-price change directly to revenue. ### Capacity-constrained revenue Units shipped = MIN(Demand units, Effective capacity units) when capacity is binding. Effective capacity = Nameplate capacity x Utilization x Yield, adjusted for downtime, mix, and ramp losses where relevant. If capacity is the binding constraint, a top-down TAM growth rate is not a forecast. Model commissioning dates, ramp curves, equipment lead times, labor, permits, power, upstream supply, yield, and customer qualification. Revenue cannot exceed the physical system unless the definition changes. ## 5. Price, volume, mix, and FX bridge For businesses with unit data, decompose growth into volume, price, mix, currency, acquisitions, and other scope. Multiplicative interactions mean the components will not always add cleanly. Use a bridge method and keep an explicit residual when disclosure is insufficient rather than forcing a false decomposition. Revenue = Units x Realized price per unit for a simple single-product case. For multi-product portfolios: Revenue = SUM(Units_i x Realized price_i). Mix changes when the weights of products or customers change. A reported price increase is not always pricing power. Separate inflation pass-through, surcharge, scarcity pricing, mix shift, geography, contract repricing, and true willingness-to-pay. Model the margin consequence because price that merely passes through higher cost may lift revenue without improving value. ## 6. Cost and margin architecture Classify costs into variable, fixed, semi-variable, pass-through, step-function, and discretionary buckets. Then map each major cost to its driver. COGS may depend on units, material input prices, yield, utilization, freight, warranty, depreciation, and mix. Operating expense may depend on headcount, compensation per employee, sales capacity, R&D program count, facilities, and variable commissions. Contribution margin = Revenue - truly variable operating costs. Incremental operating margin = Change in operating profit / Change in revenue, over a period where definition and scope are comparable. Do not extrapolate margin solely from historical percentage trends. Identify where fixed costs step up. A data center, fab, sales organization, or distribution network can create periods of strong incremental margin followed by discrete capacity investment. ### Utilization and absorption For manufacturing, low utilization can depress gross margin through under-absorbed fixed costs. Model fixed manufacturing cost separately from variable unit cost. When utilization rises, gross margin can improve even without price or product-cost improvement. That effect may reverse if capacity expands ahead of demand. ## 7. Headcount and compensation build When labor is material, model average headcount by function, hires, attrition, compensation per employee, variable compensation, payroll taxes, and stock compensation. For rapidly scaling companies, ending headcount multiplied by annual compensation can overstate expense because hiring occurs throughout the year. Use average headcount or cohort timing. Personnel expense ≈ Average headcount x cash compensation per employee + benefits/payroll burden + SBC, with function-specific assumptions where material. Link sales headcount to bookings capacity only after accounting for ramp time and quota attainment. Link engineering headcount to roadmap capacity cautiously because output is not linear in headcount. Use headcount primarily as a cost and capacity constraint, not a simplistic productivity guarantee. ## 8. Working-capital schedules Build receivables, inventory, payables, contract assets, deferred revenue, and other material operating balances from their drivers. Use days or turnover ratios only where the denominator matches the economic process and where business mix is stable. Accounts receivable ending can be modeled from DSO and revenue, but collections and billing schedules are superior when available. Inventory ending can be modeled from forward COGS and target days, adjusted for lead time, safety stock, seasonality, launch inventory, and obsolescence. Deferred revenue depends on prebilling and recognition timing, not revenue alone. Add cash-flow sign checks. An increase in an operating asset generally consumes cash; an increase in an operating liability generally provides cash, subject to classification and specific account economics. Make the schedule show both balance and period change so sign mistakes are visible. ## 9. Capex, depreciation, and asset roll-forward Separate capital spending into maintenance, growth, capitalized software or development, construction in progress, and acquisition-related asset additions. Build depreciation from asset cohorts or a simplified useful-life schedule when material. A flat depreciation percent of revenue can become wrong quickly when capex cycles are lumpy. Gross PP&E ending = Gross PP&E beginning + Capex + acquired assets + FX/reclasses - gross disposals. Accumulated depreciation ending = Beginning accumulated depreciation + depreciation expense - accumulated depreciation on disposals +/- reclasses. For capacity projects, model spend curve, in-service date, ramp, depreciation start date, and revenue capacity separately. Construction cash can leave years before the asset reaches mature utilization. ## 10. Taxes and NOL schedule Build pretax book income, permanent differences, temporary differences, jurisdiction mix, NOL usage, tax credits, valuation allowance changes, and cash taxes at the level needed for the thesis. For most companies, a simplified tax schedule can be accurate if the major drivers are explicit. For companies with large NOLs, cross-border structures, or major stock-compensation tax benefits, the schedule must be more detailed. Cash taxes = Current taxes payable adjusted for payment timing, refunds, settlements, withholding, and material discrete items. Do not use a single reported effective rate as the perpetual tax rate without examining why the rate differs from statutory. The model should distinguish near-term shield utilization from steady-state tax economics. ## 11. Debt, interest, liquidity, and cash sweep Create a debt schedule by instrument with opening balance, mandatory amortization, optional repayment, new issuance, maturity, base rate, spread, floor, fixed coupon, hedges, original issue discount where material, and covenant classification. Link cash interest to average debt balance and instrument rate. Interest expense by instrument ≈ Average debt balance x effective interest rate, plus amortization and fees where material. Model revolver draws as a liquidity backstop only after defining minimum cash, operating cash needs, covenant capacity, borrowing base, and other restrictions. A model that automatically draws infinite revolver capacity hides distress. ### Circularity control Interest depends on debt, debt may depend on cash flow, and cash flow depends on interest. Use a controlled circularity switch or an explicit iterative mechanism. The model must converge under all scenarios. If circularity is disabled, use beginning or average balances with an explicit approximation and quantify the error. ## 12. Share count, repurchases, issuance, options, RSUs, and convertibles Build basic shares, weighted-average shares, awards vesting, option exercise dilution, performance share payout, employee plan issuance, repurchases, acquisition shares, capital raises, and convertible treatment. Distinguish share count for EPS from period-end shares used for market capitalization. Equity value per share = Equity value / fully diluted forward share count appropriate to the valuation date and award mechanics. When repurchases are modeled, connect cash spent, average repurchase price, shares retired, and offsetting employee issuance. Avoid using repurchases as a mechanical EPS accretion plug. Capital allocation should be a separate assumption with liquidity and valuation discipline. ## 13. Build the statements only after the schedules The income statement should pull revenue and cost schedules, not forecast them independently. The balance sheet should pull working capital, PP&E, debt, leases, taxes, acquisitions, equity, and cash schedules. The cash flow statement should calculate from net income and balance-sheet movements, then reconcile ending cash to the balance sheet. This architecture ensures that a driver change flows through all statements. 1. Complete revenue and operating-cost schedules. 1. Calculate operating profit and taxes. 1. Complete working capital and capex/depreciation schedules. 1. Complete debt, interest, and share schedules. 1. Build the income statement. 1. Build the balance sheet excluding ending cash if cash is the balancing output. 1. Build the cash flow statement from net income and schedule movements. 1. Calculate ending cash from beginning cash plus cash flows. 1. Link ending cash to the balance sheet. 1. Run the balance check and investigate any nonzero difference. ## 14. Scenario system design A scenario is a coherent economic state, not a list of independent optimistic or pessimistic percentages. Write the causal story first. For example, a bear case might combine slower end demand, lower utilization, price pressure, inventory build, delayed capacity expansion, tighter working capital, and reduced capex. Some variables may improve in a bear case, such as growth capex or labor needs. Preserve those causal interactions. | Scenario | Demand | Price/mix | Margins | Working capital | Capital / financing | Purpose | | --- | --- | --- | --- | --- | --- | --- | | Base | Most evidence-supported path | Contract and market assumptions | Normalized execution | Normal operating range | Funded plan | Central underwriting case | | Bull | Evidence-based upside state | Better mix or pricing where causal | Operating leverage with constraints | May consume more WC | May require more growth capex | Test upside and bottlenecks | | Bear | Plausible adverse state | Discounting or weaker mix if supported | Under-absorption or cost pressure | Cash conversion often weakens | Liquidity still expected to hold | Downside without thesis break | | Stress | Severe but plausible shock | Price/volume stress | Fixed-cost and restructuring burden | Adverse WC and collections | Refinancing/covenant pressure | Solvency and liquidity | | Thesis break | Mechanism proving original thesis wrong | Structural impairment | Lower normalized earnings power | Depends on mechanism | May require recapitalization | Decision exit or full re-underwrite | ## 15. Model checks that must fail loudly - Balance sheet check equals zero in every historical and forecast period. - Cash-flow change in cash equals balance-sheet change in cash after FX and reclassification effects. - Segments plus eliminations reconcile to consolidated revenue and profit. - Debt ending balances tie to the debt schedule and disclosed maturities. - Interest expense directionally responds to debt and rate assumptions. - Share count responds to issuance, repurchases, options, RSUs, and converts. - Retained earnings roll-forward ties to net income and distributions. - Tax schedule reconciles reported provision and modeled cash-tax bridge within explained differences. - No forecast formula contains an unexplained hard-coded plug. - Scenario switch changes only intended assumptions and does not overwrite historical data. - Valuation pulls from model outputs rather than duplicate forecasts. - A sources or last-updated field exists for every material external input. ## 16. Model error taxonomy and debugging sequence Debug in layers. First check source mapping and units. Second check statement signs and roll-forwards. Third check schedule-to-statement links. Fourth check scenario switches. Fifth check circularities. Sixth check valuation bridges. Do not fix a balance-sheet error by adding a balancing plug because that destroys the diagnostic signal. | Symptom | Likely cause | First check | | --- | --- | --- | | Balance sheet off by constant amount | Opening balance or missing equity item | Historical opening balances and retained earnings | | Cash mismatch grows each year | Working-capital sign, capex, debt or FX link | Cash-flow statement versus schedule changes | | Interest explodes or oscillates | Circular debt-cash logic | Circularity switch, minimum cash, revolver formula | | EPS changes but equity value does not | Share-count or valuation denominator mismatch | Weighted-average versus period-end diluted shares | | DCF changes when display units change | Unit inconsistency | Currency and scale across FCFF, debt, cash, share count | | Segment total differs from consolidated | Eliminations or definition change | Segment reconciliation and footnotes | ## 17. Build-from-scratch training case Use a fictional company with three years of historical statements and two operating segments. Recreate history from source-like exhibits, then forecast five years. Segment A is subscription service driven by beginning ARR, new bookings, churn, and price. Segment B is hardware driven by units, ASP, utilization, and yield. Build shared corporate costs, working capital, capex, taxes, debt, and share count. The case is complete only after the model balances and a change in churn, utilization, or ASP flows through cash and valuation without manual intervention. ### Minimum test changes - Increase churn by 200 basis points and confirm revenue, deferred revenue, margins, cash flow, and valuation update. - Delay a capacity project by two quarters and confirm capex timing, depreciation, units, revenue, working capital, and liquidity update. - Increase interest rates by 150 basis points and confirm floating-rate debt, interest expense, cash, covenant headroom, and equity value update. - Increase stock compensation grants and confirm expense, cash-flow add-back, dilution, repurchase needs, and per-share value update. - Turn off all growth after the explicit forecast and confirm terminal value responds through reinvestment and mature economics rather than a hard-coded multiple. ## 18. Reviewer completion gate - A reviewer can trace any material historical cell to a source and any material forecast cell to a named driver. - The model can reproduce reported historical statements before normalization. - All model adjustments are reversible and documented. - No statement balances through unexplained plugs. - All five scenarios remain mathematically stable and economically coherent. - Valuation changes can be traced to operating assumptions rather than hidden overrides. - The model contains enough visible checks that a material sign, unit, statement, share-count, or circularity error is likely to be detected before publication. # MASTER LAB 03 - VALUATION FROM FIRST PRINCIPLES, REVERSE EXPECTATIONS, AND SPECIAL CASES > Purpose: make valuation an extension of business economics rather than a multiple-selection exercise. The analyst must be able to move from operating drivers to cash flows, discounting, reinvestment, terminal economics, capital structure, dilution, and per-share value, then reverse the current market price into the operating expectations it implies. ## Valuation decision hierarchy - Start with the asset or claim being valued: enterprise operations, common equity, a financial institution, a resource asset, a project, a segment, or a security with optionality. - Choose a method that matches the cash-flow and capital-structure economics. - Use at least one intrinsic framework and one market-relative or expectations framework when the business permits it. - Treat valuation as a range conditioned on scenarios, not a point estimate. - Reconcile methods. A large disagreement is a research question, not an invitation to average outputs blindly. ## 1. Enterprise DCF build sequence 1. Forecast operating revenue, margins, taxes, working capital, capex, and other operating reinvestment. 1. Convert operating profit to NOPAT using normalized operating taxes. 1. Add back noncash operating charges only when the corresponding economic investment is captured elsewhere. 1. Subtract all operating reinvestment needed to produce the forecast growth. 1. Define an explicit forecast horizon long enough for growth, margins, capital intensity, and competitive returns to move toward mature economics. 1. Estimate the cost of capital consistently with currency, leverage, and risk. 1. Discount interim FCFF using period-appropriate timing. 1. Estimate terminal value using internally consistent growth, reinvestment, and return assumptions. 1. Bridge enterprise value to common equity through all debt-like and non-operating claims or assets. 1. Divide by a forward diluted share count that captures awards, options, converts, and expected issuance or repurchases. FCFF = NOPAT + noncash operating charges - net investment in fixed assets - investment in working capital - other operating investment. NOPAT = EBIT x (1 - normalized operating tax rate), with explicit treatment for unusual tax items and jurisdictions where needed. The exact FCFF construction can vary with accounting and business model. The invariant principle is that cash available to all capital providers equals operating after-tax cash generation after the reinvestment required to support the forecast. Avoid mixing financing cash flows into FCFF. ## 2. Reinvestment, growth, and return consistency Growth is not free. A DCF that forecasts growth without the capital, working capital, R&D, acquisition spending, customer acquisition, or other investment necessary to support it can materially overstate value. Connect growth to the economic investment that creates it. For a steady-state framework: Growth rate = Reinvestment rate x Return on incremental invested capital. Therefore, Reinvestment rate = Growth rate / Return on incremental invested capital, when the steady-state relation is applicable. If terminal growth is 3% and mature incremental ROIC is 12%, the implied reinvestment rate is 25% of after-tax operating profit. A model that simultaneously assumes 3% perpetual growth, finite returns, and zero reinvestment is internally inconsistent. If growth is generated by intangible investment expensed through the income statement, adjust the economic model so the reinvestment is not accidentally omitted. ### ROIC fade Competitive advantage rarely disappears instantly at the terminal date. Model a fade period when the business is expected to earn returns materially above its cost of capital for a finite period. The fade can be implemented through declining margins, increasing capital intensity, slowing growth, or explicit ROIC convergence. The economic mechanism should match the moat thesis. ## 3. Explicit forecast horizon The explicit period should extend through the portion of the business cycle or competitive transition that materially affects value. A five-year template is not automatically sufficient. Early-stage capacity builds, long patent cliffs, multiyear contract ramps, resource depletion, or restructuring can require a longer horizon. Conversely, a stable mature utility may need less detail if terminal assumptions are already near steady state. Track the percentage of enterprise value coming from terminal value. A high terminal-value share is not automatically wrong, but it means the valuation is especially sensitive to mature economics. In that case, spend more research time on terminal ROIC, reinvestment, margins, and cost of capital than on small near-term quarterly differences. Terminal value concentration = Present value of terminal value / Enterprise value. ## 4. Discounting timing, stub periods, and midyear convention Match discounting to when cash is economically received. If valuation occurs between fiscal year ends, use an actual or approximate stub fraction. If annual cash flows are earned throughout the year, midyear convention often better reflects timing than year-end discounting. Year-end discount factor for period t = 1 / (1 + WACC)^t. Midyear discount factor for annual period t = 1 / (1 + WACC)^(t - 0.5), adjusted for a valuation-date stub where necessary. Do not mix a midyear terminal value with year-end interim cash flows unless the timing convention is explicitly reconciled. Small timing errors can matter when the valuation duration is long or the discount rate is high. ## 5. Cost of equity and beta CAPM baseline: Cost of equity = Risk-free rate + Beta x Equity risk premium, plus any separately justified risk components that do not double count existing inputs. Use a risk-free rate consistent with the currency of the cash flows, not simply the company headquarters. Use an equity risk premium appropriate to the market exposure and methodology. Beta is an estimate, not a fact. Review the regression period, frequency, index, leverage, business mix, and structural changes. ### Bottom-up beta For companies undergoing major business or leverage change, estimate asset risk from comparable businesses. Unlever peer equity betas, take a robust central estimate, then relever to a target capital structure. This is often more stable than relying on one noisy historical regression. Simplified unlevered beta = Levered beta / [1 + (1 - tax rate) x Debt/Equity], where the assumptions behind debt beta and tax shield are acceptable. Simplified relevered beta = Unlevered beta x [1 + (1 - tax rate) x target Debt/Equity]. When debt is risky, the zero-debt-beta simplification becomes less reliable. Distressed companies may require a more explicit asset-risk and debt-risk treatment. ## 6. Cost of debt and WACC WACC = Weight of equity x Cost of equity + Weight of debt x After-tax cost of debt + other capital components as applicable. Use market-value weights when possible. The cost of debt should reflect the marginal financing cost for the forecast risk, not only the coupon on legacy low-cost debt. Consider maturity structure, fixed versus floating mix, hedges, credit spreads, and refinancing needs. Apply the tax shield only to the extent interest deductibility is economically expected. ### Capital structure consistency WACC, FCFF, and the enterprise-to-equity bridge must describe the same capital structure. If the forecast assumes major deleveraging or recapitalization, a constant WACC may be inappropriate. Either model changing capital structure or explain why a steady target structure is a reasonable approximation. ## 7. Terminal value using perpetuity growth Terminal value at end of year N = FCFF_(N+1) / (WACC - g), where WACC > g and FCFF reflects steady-state economics. The terminal year must be normalized before applying the perpetuity formula. Normalize margins, taxes, reinvestment, working capital, cyclicality, capital structure, and one-time items. The numerator should be a sustainable next-period cash flow, not a peak-cycle or transition-year result. Long-run nominal growth should be constrained by the currency and economy in which cash flows are generated. A perpetual growth assumption above the sustainable nominal growth of the relevant economic base requires exceptional justification because the company would eventually become implausibly large relative to that base. ## 8. Terminal multiple as a cross-check, not a substitute Calculate the multiple implied by the perpetuity-growth terminal value. If the terminal value implies a mature EV/EBITDA, EV/EBIT, P/E, or FCF yield inconsistent with the terminal growth, margins, returns, and risk, investigate. A terminal multiple can also be used directly in some settings, but it should be tied to mature economics rather than copied from a current peer median. Implied terminal EV/EBITDA = Terminal enterprise value / Terminal EBITDA. A high-growth company can deserve a high current multiple and a much lower terminal multiple. A cyclical company can trade at a deceptively low multiple on peak earnings. Context matters more than the raw number. ## 9. Enterprise-to-equity bridge The bridge is one of the most common sources of valuation error. Build a line-by-line claim schedule. Start with enterprise value, subtract debt and debt-like claims, add excess cash and non-operating assets, adjust for minority interests or noncontrolling claims, preferred stock, pension deficits, environmental or restructuring obligations when debt-like, investments, associates, and other material assets or liabilities. | Bridge item | Typical direction | Key question | | --- | --- | --- | | Gross debt | Subtract | Is debt already reflected in a financial-company equity model instead of enterprise value? | | Cash | Add only excess or non-operating cash | How much cash is required for operations, regulation, collateral, or trapped jurisdictions? | | Lease liabilities | Depends on operating metric convention | Are lease expenses and peer multiples consistently treated? | | Pension deficit | Often subtract when debt-like | Are future contributions already embedded in FCFF? Avoid double counting. | | Minority interest | Usually subtract claim not owned by common | Does consolidated operating profit include subsidiaries not wholly owned? | | Equity investments | Add non-operating value | Is income from the investment removed from operating earnings? | | Options/RSUs | Capture through diluted shares or claim value | Do not ignore in-the-money employee claims. | | Convertibles | Debt or equity depending on treatment | Ensure numerator and denominator treatment are internally consistent. | ## 10. Dilution and per-share value Per-share value requires a capitalization table, not a single diluted share figure copied from the latest income statement. Model current basic shares, treasury stock, options by strike, unvested RSUs, performance shares, employee plans, convertibles, warrants, acquisition earn-outs payable in stock, and expected capital raises where material. Use treasury-stock or option-value methods appropriate to the purpose. When valuation per share is far above the current market price, more options can become in the money. Therefore, dilution can be valuation-dependent. Test this rather than assuming a fixed diluted count. ## 11. Reverse DCF and expectations investing A reverse DCF asks what operating path makes the current price reasonable. This converts valuation from an argument about the correct multiple into a testable set of business expectations. Solve for one or two key variables at a time while holding the rest at defensible values. 1. Start with current enterprise value and capital structure. 1. Use a normalized cost of capital and explicit forecast horizon. 1. Choose the variable most connected to the thesis, such as revenue CAGR, terminal margin, incremental ROIC, retention, units, capacity, or reinvestment. 1. Solve for the path that equates present value to market value. 1. Compare the implied path with company history, peer economics, industry capacity, customer demand, and physical constraints. 1. Identify the earliest observable KPI that would confirm or falsify the market-implied path. ### Example reverse expectation Suppose the market price implies revenue must compound 22% for seven years while operating margin expands from 8% to 24%, and the business must maintain 25% incremental ROIC. The analytical question is not whether 22% sounds high. The question is whether customer adoption, capacity, pricing, competition, and required investment can jointly support that combination. Compare the required outcome with the addressable market and with the strongest competitor response. ## 12. Relative valuation done correctly Peer multiples are useful only after normalizing denominator definitions and economic differences. Compare growth, margins, returns on capital, capital intensity, leverage, accounting policy, geographic mix, customer concentration, cyclicality, and durability. A median multiple is not intrinsic truth. | Multiple | Best suited for | Major weakness | | --- | --- | --- | | P/E | Equity claims with comparable leverage and accounting | Distorted by leverage, tax, non-operating gains, buybacks, and accounting differences | | EV/EBIT | Operating businesses where depreciation is economic | Requires consistent lease and capitalization treatment | | EV/EBITDA | Capital structures with different D&A patterns | Can hide capex and working-capital intensity | | EV/Revenue | Early businesses with meaningful gross-profit potential | Ignores gross margin, opex burden, capital intensity, and dilution | | EV/Gross profit | Businesses with widely varying gross margins | Still ignores operating cost and capital intensity | | FCF yield | Mature cash-generative equities | Can be distorted by working-capital timing and underinvestment | | P/TBV | Banks and insurers where book capital is economically relevant | Book quality, reserve adequacy, ROE, and risk can differ sharply | | NAV / asset value | Resources, property, holding companies | Highly sensitive to asset values, discount rates, and corporate costs | Use regression or matched-pair logic only if the sample is economically comparable. A statistical relation does not prove causality. The analyst should be able to explain why a variable such as growth, ROIC, or margin should influence the multiple and why the relation is stable enough to use. ## 13. Sum-of-the-parts valuation Use SOTP when segments have materially different economics, peer sets, capital intensity, growth, risk, or ownership. Value each segment using the most appropriate method, then subtract central costs, taxes, debt-like claims, minority interests, and other corporate items. Equity value = Sum of segment enterprise or equity values - central debt-like claims - present value of unallocated corporate costs + non-operating assets, with consistent ownership adjustments. Do not value segments on peer multiples and then forget shared corporate costs or required holding-company capital. If a segment is not separately financeable, account for structural dependencies. For spin-off analysis, include stranded costs, separation costs, tax leakage, debt allocation, and transition-service arrangements. ## 14. Cyclical companies Do not capitalize peak or trough earnings as if they are normal. Build mid-cycle volume, price, utilization, margins, working capital, and capex from a full cycle. Use replacement capacity, industry cost curve, supply additions, and demand elasticity to inform the normalized state. Mid-cycle earnings should reflect normalized industry conditions, not a simple arithmetic average if the historical period contains structural change. Multiples can invert in cyclicals. A low P/E or EV/EBITDA at peak earnings can signal risk, while a high multiple at trough earnings can coexist with attractive normalized value. Use normalized free cash flow and asset economics as anchors. ## 15. Commodity and resource valuation Model production, reserve life, grade, recovery, decline, realized price, basis differentials, royalties, operating cost, sustaining capex, growth capex, reclamation, taxes, and hedges. Use a commodity price deck with explicit near-term market assumptions and long-run incentive economics. Stress the price because it often dominates value. Project NAV = Present value of after-tax project cash flows less required development and closure obligations, adjusted for probability and ownership as appropriate. Do not treat resources and reserves as interchangeable. Account for mine life, depletion, permitting, development probability, infrastructure, jurisdiction, and capital intensity. Corporate overhead and future exploration spending can materially change equity value relative to asset NAV. ## 16. Financial institutions Banks and insurers are generally better valued through equity cash flows, book value, excess capital, ROE, and franchise economics than through an industrial enterprise-value DCF because debt-like liabilities are operating inputs. Model net interest income, credit losses, fee income, expenses, capital requirements, and distributions directly. A simple residual-income concept: Equity value = Current book value + present value of future residual income, where residual income = (ROE - cost of equity) x beginning equity. For banks, normalize credit costs and deposit funding. For insurers, analyze underwriting margin, reserve adequacy, investment income, catastrophe exposure, reinsurance, and capital. Book value deserves a premium only when sustainable returns exceed the required return with acceptable risk. ## 17. REITs and property companies Use property-level NOI, occupancy, rent growth, lease maturity, tenant credit, cap rates, development pipeline, recurring capex, and balance-sheet leverage. FFO and AFFO definitions vary, so reconstruct them rather than accepting labels. NAV and cash-flow approaches should reconcile with implied cap rates and financing costs. Property value ≈ Stabilized NOI / Capitalization rate, with asset-specific adjustments for growth, capex, vacancy, and quality. A lower cap rate increases asset value but also usually coincides with a financing environment. Stress both asset value and debt costs. For development-heavy REITs, separate stabilized assets from construction and land value. ## 18. Early-stage and negative-FCF businesses A negative current cash flow does not prevent intrinsic valuation, but it increases dependence on the path to positive unit economics and financing. Build customer or unit economics, gross margin progression, opex scaling, capital needs, dilution, and probability of reaching a sustainable state. Use scenario probability explicitly when survival or product-market fit is uncertain. Avoid valuing an early-stage company solely on a revenue multiple without connecting the multiple to eventual margins, reinvestment, dilution, and duration. Reverse the multiple into the mature economics required to justify it. ## 19. Distressed and restructuring valuation When solvency is uncertain, enterprise DCF must be supplemented by claim-priority analysis. Build a liquidity runway, debt maturity schedule, covenant map, collateral package, security ranking, restricted subsidiaries, guarantees, and potential debtor-in-possession or rescue financing. Value the business under going-concern and liquidation or restructuring scenarios. Recovery by claim class depends on reorganized or liquidation enterprise value, administrative and priority claims, collateral, structural priority, contractual ranking, and the negotiated plan. A common equity valuation can be near zero even when enterprise value is substantial if senior claims consume the value. Conversely, an equity stub can retain option value if there is a credible path to higher enterprise value before liquidity is exhausted. Do not confuse option value with base-case value. ## 20. M&A valuation and accretion versus value creation Accretion is not synonymous with value creation. Analyze purchase price, standalone value, control premium, synergies, integration cost, financing, tax attributes, purchase accounting, dilution, and opportunity cost. A low-cost debt-funded acquisition can be EPS-accretive while destroying value if the buyer overpays. Value created for buyer ≈ Present value of realizable synergies and strategic benefits - premium paid over target standalone value - integration and transaction costs - financing or execution costs not already included. Build synergy timing and probability explicitly. Revenue synergies usually deserve more skepticism than cost synergies because they depend on customer behavior. Include antitrust remedies, divestitures, employee attrition, customer churn, systems conversion, and delayed close where material. ## 21. Scenario valuation and probability weighting Do not assign probabilities merely to create a weighted average. Each scenario needs an economic narrative, a coherent set of drivers, a valuation method, and evidence that would move probability. Probabilities should update as evidence arrives. Keep the unweighted scenario values visible so a probability assumption cannot hide the downside. Probability-weighted value = SUM(Scenario value_i x Probability_i), where probabilities sum to 100% and scenario definitions are mutually coherent. Include a thesis-break case separately when the business model can fail structurally. The thesis-break case is not just a more pessimistic bear case. It describes evidence that invalidates the original causal thesis. ## 22. Sensitivity design Sensitivity tables should vary assumptions that actually drive value. Typical DCF tables vary WACC and terminal growth, but those are often downstream abstractions. Also vary revenue duration, normalized margin, reinvestment, ROIC, retention, capacity, price, or capital intensity. If one operating variable dominates, focus the research program on reducing uncertainty in that variable. Value sensitivity can be expressed as change in equity value for a defined change in an assumption, or as elasticity = % change in value / % change in assumption where interpretable. ## 23. Valuation error checks - WACC exceeds perpetual growth in a standard perpetuity-growth model. - Terminal growth and reinvestment are consistent with terminal incremental ROIC. - Terminal margins and returns are economically plausible for a mature competitive state. - Debt, cash, leases, pensions, minorities, investments, options, and converts are not double counted or omitted. - Cash flows and discount rate use the same currency and inflation basis. - Nominal cash flows are discounted with nominal rates; real cash flows with real rates. - The per-share denominator matches the valuation date and dilution mechanics. - Peer multiples use comparable denominator definitions and capital-structure treatment. - The DCF implied terminal multiple is reviewed, and a multiples valuation is reverse-checked against cash-flow economics. - The current price is translated into operating expectations and compared with physical, customer, and competitive constraints. ## 24. Worked DCF example Assume a company generates $150 million of next-year NOPAT. Depreciation is $40 million, capex is $65 million, and working-capital investment is $15 million. Other operating reinvestment is $10 million. Next-year FCFF is therefore $100 million. Suppose FCFF grows through explicit operating forecasts for seven years, WACC is 9%, and the mature business can grow 3% with 12% incremental ROIC. Year 1 FCFF = 150 + 40 - 65 - 15 - 10 = $100 million. Terminal reinvestment rate = 3% / 12% = 25% of terminal NOPAT under the steady-state relation. The terminal cash flow must reflect that reinvestment. If terminal NOPAT is $300 million, steady-state reinvestment would be approximately $75 million before considering the exact composition of noncash charges and investment. The analyst then calculates terminal FCFF, applies the perpetuity formula, discounts interim and terminal cash flows, and completes the enterprise-to-equity bridge. Repeat under at least two WACC and mature-ROIC assumptions. The output is a range, not a single target price. ## 25. Valuation completion gate - The selected method matches the economics of the security or operating asset being valued. - Every key valuation input is linked to the operating model, market evidence, or an explicit assumption. - Growth, reinvestment, and returns are mathematically consistent. - The enterprise-to-equity bridge is complete and free of double counting. - Dilution is dynamic where material. - At least one reverse-expectations analysis shows what the current price requires. - At least one market-relative framework is normalized for accounting and economic differences where peers exist. - Bull, base, bear, stress, and thesis-break values are visible with causal scenario definitions. - A senior reviewer can change any major operating or discount-rate assumption and trace the entire value impact without hidden overrides. # MASTER LAB 04 - FORENSIC ACCOUNTING, EARNINGS QUALITY, AND MANIPULATION DETECTION > Purpose: identify when reported performance, adjusted performance, cash generation, or management narrative diverges from the underlying economics. The analyst is not trying to prove fraud from ratios. The analyst is building a disciplined evidence chain that distinguishes benign timing, aggressive accounting, weak controls, structural deterioration, and possible misconduct. ## 1. Forensic operating principle Start with multi-year economics and cash, not with a fraud score. Most accounting anomalies have several possible explanations. The job is to identify the anomaly, quantify it, trace it to primary evidence, generate competing explanations, and seek discriminating evidence. Never accuse a company or person of misconduct merely because a screening ratio is elevated. - Preserve the reported numbers before adjustments. - Use at least five years or a full relevant cycle when available. - Compare income statement, balance sheet, cash flow, footnotes, segment data, and non-GAAP reconciliation together. - Read disclosure changes, not just numbers. - Escalate combinations of incentive, opportunity, discretion, weak controls, and inconsistent cash evidence. - Quantify valuation impact under both benign and adverse explanations. ## 2. Earnings-to-cash bridge Accrual component of earnings ≈ Net income - Cash flow from operations, with normalization for classification and non-operating items where needed. Persistent earnings growth without corresponding cash generation deserves explanation, but cash flow itself can be managed through working-capital timing, supplier finance, receivable sales, tax timing, or classification. Build a multi-year bridge from net income to CFO and from CFO to a normalized operating cash measure. | Pattern | Benign explanation | Adverse explanation | Discriminating evidence | | --- | --- | --- | --- | | CFO below net income | Rapid growth and working-capital investment | Aggressive revenue, weak collections, inventory build | DSO, contract assets, inventory, customer terms, aging | | CFO above net income | Deferred revenue, strong collections, noncash charges | Payables stretch, supplier finance, underinvestment | DPO, financing programs, capex, payable aging | | FCF strong despite weak operations | Working-capital release after prior build | Capex deferral, factoring, asset sales mischaracterized | Maintenance needs, receivable sales, cash-flow classifications | | Adjusted FCF much above GAAP FCF | Legitimate one-time restructuring | Recurring excluded cash costs | History of exclusions and recurrence by category | ## 3. Accrual quality and balance-sheet growth Analyze total accruals and working-capital accruals over several years. High accruals are a prompt for deeper work, not proof of manipulation. Rapidly growing companies can generate legitimate accruals. The key question is whether balance-sheet assets created by earnings convert to cash at rates consistent with the stated economics. Simple working-capital accrual = Change in noncash current operating assets - Change in current operating liabilities, adjusted for acquisitions, FX, and reclassifications when material. Asset growth rate = Change in average or ending assets / prior assets, interpreted with acquisitions and capital intensity. Watch for receivables, contract assets, inventory, capitalized costs, deferred commissions, prepaid assets, or other assets growing materially faster than the activity they support. Ask whether the growth reflects capacity creation, contract timing, a business-model shift, or capitalization of current-period cost. ## 4. Revenue forensic tests 1. Reconcile revenue growth to units, price, mix, FX, acquisitions, and accounting presentation. 1. Compare receivable and contract-asset growth with revenue and billings. 1. Review customer concentration and changes in payment terms. 1. Inspect channel inventory, returns, rebates, concessions, and side-agreement risk where relevant. 1. Compare backlog or RPO additions, cancellations, and conversion with reported demand claims. 1. Search filings for changes in revenue policy, principal-agent conclusions, contract modification language, or significant judgments. 1. Check quarter-end seasonality and unusual acceleration in the final month or quarter if data are available. ### Receivable diagnostic Receivables growth gap = Accounts receivable growth - Revenue growth. A positive gap can be normal when billing mix changes, but persistent widening requires a collection explanation. Calculate DSO on a comparable basis, inspect allowance for credit losses, compare bad-debt expense, and read aging disclosures when available. If the company sells receivables, add sold balances back for an economic DSO view where appropriate. ### Contract asset diagnostic Contract assets can rise because work is performed before an unconditional billing right exists. Test the balance against revenue recognition method, project milestones, change orders, disputed claims, and cash conversion. A contract asset is economically different from a standard receivable and should not be analyzed as interchangeable cash due. ## 5. Inventory forensic tests Inventory can reveal demand weakness, supply-chain buffering, new-product launch preparation, inflation, or production inefficiency. Decompose raw materials, work in process, and finished goods when disclosed. Compare inventory growth with units, sales, lead times, backlog, capacity, purchase commitments, and obsolescence reserves. Inventory growth gap = Inventory growth - COGS or sales growth, using the denominator that best matches the business. Rising inventory and rising gross margin can coexist if fixed-cost absorption improves while sell-through weakens. That combination deserves special attention because the income statement may look stronger before the cash or write-down consequence appears. ## 6. Payables, supplier finance, and cash-flow optics A company can temporarily improve CFO by extending supplier terms or using a bank-supported supplier-finance program. Identify balances, payment terms, program size, cash-flow classification, and whether suppliers are paid earlier by a financing intermediary. Recast debt-like supplier financing when it materially changes operating-cash comparability. Compare payable growth with purchases or COGS and with supplier commentary. A sudden DPO expansion can be bargaining power, deliberate cash conservation, distress, or financing reclassification. The interpretation depends on commercial terms and liquidity context. ## 7. Capitalized-cost tests Capitalization moves current-period cost from the income statement to the balance sheet. Review software development, internal-use software, commissions, film/content costs, exploration, interest, contract acquisition costs, customer implementation, and other capitalized expenditures relevant to the business. Capitalization intensity = Capitalized cost / related operating activity or revenue, tracked through time on a consistent definition. Build a counterfactual expense if the capitalization rate rises materially. Then estimate the effect on operating margin and cash flow. Capitalization does not automatically improve cash flow because the cash may still leave the business, but it can improve reported earnings and EBITDA depending on presentation. ## 8. Depreciation, useful lives, and impairment timing Compare asset useful lives, residual values, depreciation methods, and impairment triggers through time. Extending useful lives can reduce current depreciation and increase asset carrying value. Delayed impairment can keep assets and earnings higher until a later write-down. Neither proves manipulation, but both require economic evidence. Test whether asset age, utilization, replacement spending, technology change, product obsolescence, or market prices support the carrying values. For goodwill and indefinite-lived intangibles, compare reporting-unit performance with acquisition assumptions and market valuation. ## 9. Reserve releases and estimate management Reserves create opportunities for earnings smoothing because current estimates can release prior-period expense. Build roll-forwards for warranty, returns, credit losses, litigation, restructuring, insurance, environmental obligations, rebates, and other material estimates. Calculate provision rates relative to the underlying activity. Reserve coverage ratio = Ending reserve / relevant exposure, with exposure defined specifically for the reserve. A reserve release during operational weakness is not automatically inappropriate. The analyst must understand whether the underlying exposure genuinely declined. The red flag is favorable estimate movement without supporting operating evidence, especially when it helps meet guidance or incentive thresholds. ## 10. Non-GAAP reconstruction Create a category-level reconciliation for at least five years. Do not accept the issuer labels of recurring or nonrecurring. Classify each adjustment by economic nature, cash versus noncash, frequency, controllability, acquisition linkage, and whether excluding it improves comparability or merely presentation. | Adjustment | Questions | | --- | --- | | Stock compensation | Is it recurring compensation? What is dilution? How much cash is spent to offset it? | | Restructuring | How often has it occurred? Is restructuring part of the normal operating model? | | Acquisition costs | Is acquisition activity recurring? Are integration and retention costs necessary to realize acquired earnings? | | Intangible amortization | What asset was purchased? Is replacement through future acquisition spending likely? | | Legal settlements | Is the issue isolated or part of recurring business/regulatory exposure? | | Gains/losses | Are favorable gains excluded as consistently as unfavorable losses? | | Transformation costs | Has the transformation persisted for multiple years? What is the cumulative cash cost? | Build your own normalized earnings view. The objective is not to be mechanically conservative. The objective is to estimate repeatable economic earnings and cash flow under a clearly stated definition. ## 11. Acquisition masking and organic reconstruction Frequent acquisitions can reset growth rates, margins, reserves, depreciation, and segment definitions. Build a transaction calendar and acquisition cohort table. Separate reported growth into acquired, organic, price, volume, FX, and presentation. Track cash purchase price, shares issued, debt assumed, earn-outs, integration costs, restructuring, and goodwill. When an acquisition closes near quarter-end, compare the acquired contribution with purchase-accounting adjustments and pro forma disclosures. Watch for cases where acquired deferred revenue or inventory step-ups alter post-close reported margins and growth. ## 12. Segment and KPI definition drift Compare each filing with the prior version. Search for new qualifiers, removed metrics, changed definitions, restated segment boundaries, altered non-GAAP reconciliations, or changes in which KPIs management emphasizes. Disclosure drift can be informative even when every individual change is permitted. 1. Archive prior filing text or structured extracts. 1. Run a section-level diff. 1. Classify changes as mandatory update, business change, definition change, risk-language change, or unexplained removal. 1. Quantify whether the changed disclosure affects a thesis variable. 1. Ask what evidence would make the benign explanation more likely than the adverse explanation. ## 13. Auditor, critical audit matters, and control signals Read the auditor report, critical audit matters, material-weakness disclosures, restatements, auditor changes, audit-fee trends, and audit-committee disclosures. Critical audit matters identify areas involving especially challenging, subjective, or complex audit judgment. They are research maps, not accusations. A material weakness increases the risk of a material misstatement, but the analytical impact depends on the process affected, remediation, duration, and whether prior statements require restatement. Track remediation milestones rather than assuming the issue disappears because management says it is being addressed. ## 14. Related parties and unusual counterparties Inventory related-party customers, suppliers, lenders, lessors, joint ventures, executives, founders, family entities, and significant shareholders. Compare transaction terms, balances, guarantees, loans, and asset transfers. Related-party transactions can be legitimate, but they deserve heightened scrutiny because arm-length pricing and collectability may be harder to establish. For customer or supplier concentration, inspect whether a counterparty is financially dependent on the company, financed by the company, or connected through ownership. Revenue quality is weaker when apparent external demand is economically circular. ## 15. Liquidity signals hidden inside accounting Distress often appears in working capital and footnotes before it appears in EBITDA. Monitor receivable sales, supplier finance, covenant amendments, minimum liquidity, revolver draws, restricted cash, collateral changes, delayed capex, stretched payables, customer deposits, tax deferrals, and unusual asset sales. A company can report adjusted EBITDA growth while losing financial flexibility. Build a cash runway using cash, committed availability, mandatory payments, interest, capex, working capital, restructuring, and realistic access to capital. ## 16. Screening models: use as triage, never verdict Statistical screens such as Beneish-style manipulation indicators or Altman-style distress scores can identify observations worth investigating. They can also produce false positives in industries with unusual working capital, high growth, acquisitions, or regulated capital. Recalculate inputs from clean source data and use the result as a prompt for primary-source work. Never write that a fraud screen proves fraud. Instead write that the screen identifies specific components such as receivable growth, asset quality, depreciation changes, or accrual intensity that require further explanation. ## 17. Management incentives and threshold behavior Map compensation metrics, debt covenants, earn-outs, acquisition milestones, analyst consensus, guidance ranges, and financing needs. Incentives do not prove accounting bias, but they tell the analyst where discretion matters most. Pay special attention when a subjective estimate changes just enough to meet a threshold. Compare compensation definitions with the company non-GAAP framework. If management is paid on adjusted EBITDA that excludes recurring restructuring or SBC, understand whether the incentive encourages economic performance or metric management. ## 18. Forensic issue register | Field | Required content | | --- | --- | | Issue | Specific observed anomaly, not a conclusion | | Magnitude | Dollar, margin, cash-flow, balance-sheet, or per-share effect | | Primary evidence | Filing section, footnote, exhibit, auditor language, transaction data | | Benign explanation | Most plausible non-adverse explanation | | Adverse explanation | Accounting, operational, control, or integrity concern | | Discriminating test | Evidence that would separate explanations | | Status | Resolved, open-immaterial, open-material, thesis-break | | Model treatment | No change, sensitivity, normalization, scenario, liquidity adjustment | | Next trigger | Dated filing, KPI, cash conversion, audit update, or other evidence | ## 19. Worked forensic case Fictional company Atlas Devices reports revenue growth of 24%, adjusted EPS growth of 35%, and adjusted FCF growth of 40%. Over the same year, receivables grow 48%, contract assets grow 60%, inventory grows 52%, supplier-finance balances rise sharply, capitalized software increases 80%, and SBC is excluded from adjusted earnings. Management explains the working-capital build as preparation for strong demand. 1. Reconstruct reported and adjusted earnings for five years. 1. Calculate DSO, contract-asset intensity, inventory days, DPO, supplier-finance-adjusted DPO, capitalization intensity, dilution, and cash conversion. 1. Build a benign case in which product launch and constrained supply explain the balance-sheet build. 1. Build an adverse case in which demand is weaker, revenue is pulled forward, costs are capitalized, and supplier financing flatters CFO. 1. Identify three near-term pieces of evidence that distinguish the cases, such as collections, sell-through, cancellations, reserve changes, or capitalized-cost amortization. 1. Quantify revenue, margin, cash, liquidity, and valuation under each interpretation. The conclusion should not be fraud or no fraud. The conclusion should state which economic interpretation is best supported, what remains unresolved, how much value is exposed, and what future evidence would force an update. ## 20. Forensic completion gate - Reported earnings reconcile to cash, working capital, and balance-sheet changes over a multi-year window. - Every recurring non-GAAP adjustment has been reconstructed by category and cash nature. - Material reserves and capitalized costs have roll-forwards and exposure-based ratios. - Revenue quality is cross-checked against units, price, collections, contract balances, backlog/RPO, or other operating evidence. - Acquisition effects are separated from organic economics. - Auditor, control, related-party, and disclosure-drift evidence has been reviewed. - At least one benign and one adverse explanation have been quantified for each thesis-relevant anomaly. - No manipulation or fraud conclusion is based solely on a screening ratio. - The issue register identifies what evidence resolves each open item and how the model changes if the adverse explanation is correct. # MASTER LAB 05 - INDUSTRY STRUCTURE, COMPETITIVE ADVANTAGE, MARKET SIZING, AND TECHNOLOGY TRANSITIONS > Purpose: convert broad industry narratives into an investable map of customers, suppliers, economics, bottlenecks, profit pools, competitive behavior, technology change, regulation, and measurable leading indicators. The output must explain not just how large an industry could become, but who captures value, what constrains growth, and what evidence would show the structure is changing. ## 1. Build the value-chain map Start by drawing the physical, contractual, and economic flow from raw input to final customer. Identify every material layer that can capture margin, create a bottleneck, control standards, or impose switching cost. Do not begin with the public companies you already know because that anchors the map around listed incumbents rather than the actual system. | Layer | Questions to answer | | --- | --- | | Inputs | What raw materials, data, IP, labor, components, land, power, or licenses are required? | | Production / transformation | Where are capacity, yield, utilization, quality, and scale decisive? | | Distribution / channel | Who owns customer access, inventory, financing, installation, or service? | | Customer | Who makes the purchase decision, who uses the product, who bears failure cost, and who has bargaining power? | | Complements | What adjacent infrastructure or software must exist for adoption? | | Regulators / standards | Which rules, certifications, reimbursement systems, interconnection standards, or technical protocols control access? | | Capital providers | Where does financing availability affect adoption or capacity expansion? | For each layer, estimate revenue pool, gross profit pool, capital employed, market concentration, pricing mechanism, capacity, growth, and bargaining power. Profit pools matter more than revenue pools. A fast-growing layer can destroy capital if entry is easy and returns are competed away. ## 2. Market sizing from first principles Build bottom-up TAM, serviceable market, and realistic adoption rather than accepting promotional TAM slides. Start with a countable unit, usage rate, capacity requirement, replacement cycle, price, or spend pool. Use at least two independent constructions when valuation depends heavily on market size. Simple bottom-up market size = Addressable units x Penetration or usage x Annual units per customer x Realized price. Capacity market size = Required system capacity x Installed cost or annual revenue per capacity unit, adjusted for utilization and replacement. Separate stock from flow. Installed base is not annual revenue unless products are replaced or monetized repeatedly. Separate one-time hardware, recurring software/service, maintenance, consumables, financing, and aftermarket revenue. Avoid double counting layers of the same value chain. ### TAM sanity checks - Compare implied spend with customer budgets or revenue. - Compare implied physical capacity with supply-chain and infrastructure limits. - Compare implied installed base with replacement cycles. - Compare implied market share with plausible competitor count and customer concentration. - Compare implied growth with adoption history for analogous technologies. - Translate the company valuation into required share of industry profit, not only industry revenue. ## 3. Demand model and adoption curve Identify the customer problem, current alternative, economic benefit, payback period, implementation burden, switching cost, regulation, organizational friction, and failure risk. Adoption accelerates when the total value proposition crosses a threshold, not simply because awareness rises. Customer ROI ≈ Incremental economic benefit - incremental operating cost - switching/implementation cost, evaluated over the relevant life and risk. For new technologies, separate innovators, early adopters, mainstream buyers, and laggards only if the segmentation maps to real customer economics. Use actual purchase triggers such as total cost of ownership, performance, reliability, regulation, subsidy, labor scarcity, or capacity shortage. ### Adoption constraint tree - Product readiness: performance, reliability, interoperability, safety. - Supply readiness: manufacturing capacity, qualified suppliers, yield, materials. - Infrastructure: power, network, charging, interconnection, data center, logistics, installation. - Customer economics: payback, financing, operating savings, resale value. - Organization: training, procurement cycle, integration, change management. - Regulation: permits, standards, reimbursement, export controls, local content. - Competing technology: incumbent cost declines, substitutes, leapfrog risk. ## 4. Supply, capacity, and bottleneck analysis A growth market can still produce poor equity returns if capacity expands faster than demand. Build a supply ledger by producer, geography, technology, announced capacity, funded capacity, construction status, expected commissioning, qualification time, ramp yield, and realistic utilization. Discount press-release capacity that lacks financing, permits, customers, equipment, or infrastructure. Effective industry supply = SUM(Nameplate capacity x ramp factor x utilization x yield x availability), measured on a consistent unit. Identify the true bottleneck. In one cycle it may be semiconductors, in another transformers, grid interconnection, HBM, skilled labor, permits, specialized test equipment, or customer qualification. A bottleneck can shift the profit pool toward a small upstream layer even when downstream revenue is much larger. ## 5. Cost curves and marginal pricing In commodity-like markets, the marginal producer often helps set price. Build a cost curve by producer or technology where data permit. Include cash operating cost, freight, royalties, sustaining capital, carbon or regulatory costs, and quality differentials as relevant. In manufactured goods, construct an approximate cost curve using scale, yield, labor, input cost, depreciation, and logistics. A structural cost advantage matters only if competitors cannot replicate it cheaply and if customers do not capture all of the advantage through lower prices. Estimate how much of the cost advantage is retained as margin versus passed through. ## 6. Competitive-advantage testing | Potential moat | Evidence that supports it | Evidence that weakens it | | --- | --- | --- | | Switching costs | High retention despite price increases, integration burden, workflow lock-in | Easy migrations, multi-vendor use, low churn cost | | Network effects | Product value rises with relevant participants or data | Users multi-home, network quality does not scale with size | | Scale economies | Unit cost or service quality improves with scale | Diseconomies, localized competitors, fixed cost not material | | Brand / trust | Willingness to pay, lower acquisition cost, risk-sensitive purchase behavior | Commodity bidding, promotions drive share | | Cost advantage | Persistent delivered-cost gap from proprietary process, resource, scale, or location | Competitors rapidly match cost or customers force pass-through | | Data advantage | Unique data improves product and compounds with use | Data broadly available or not causally tied to product quality | | Regulatory / license | Scarce authorization or hard-to-replicate compliance infrastructure | Rule change, easier licensing, mandated interoperability | | Ecosystem / standard | Complementors build around platform and switching is costly | Open standard commoditizes control point | A moat is a mechanism that supports excess returns on incremental capital for longer than competitors can erode them. Do not equate market share, brand awareness, patents, or high margins with a moat without showing the causal barrier and the reinvestment economics. ## 7. Competitor benchmarking Normalize peer definitions before comparing. Build a peer table containing growth, gross margin, operating margin, FCF, ROIC, capital intensity, working capital, leverage, retention or unit KPIs, price, capacity, customer mix, geography, and valuation. Annotate definition differences instead of hiding them. Use matched comparisons. A vertically integrated producer should not be compared mechanically with an asset-light designer. A recurring software product should not be compared with perpetual-license revenue without adjusting timing. The objective is to identify why economics differ and whether the difference is structural. ## 8. Profit-pool migration Technology transitions often shift value between layers. Track which layer controls scarcity, standards, data, customer relationship, or performance. For example, a system transition may shift profit from commodity hardware to specialized components, software control, power electronics, networking, test, service, or financing. The most obvious end-market winner may not be the best value-capture point. 1. Map current industry revenue and gross profit by layer. 1. Identify which technical or economic constraint is becoming more important. 1. Identify the suppliers whose product removes that constraint. 1. Estimate customer value created versus supplier revenue captured. 1. Test whether competitors can enter before scarcity rents fade. 1. Model the duration of excess returns and the capital required to expand supply. ## 9. Technology S-curves and learning curves Separate performance improvement from cost decline and adoption. A technology can improve rapidly but adopt slowly because infrastructure or customer economics lag. Conversely, adoption can accelerate with modest technical progress if regulation or economics change. Learning-curve concept: Unit cost may decline by a roughly consistent percentage for each doubling of cumulative production in some technologies. Estimate empirically rather than assume a universal rate. Track performance per dollar, energy density, compute per watt, throughput, yield, reliability, cycle life, latency, efficiency, or other domain-specific frontier metrics. Identify whether improvement is physical, architectural, manufacturing, software-driven, or subsidy-driven. Each has different durability. ## 10. Standards, regulation, and policy Regulation can expand a market, restrict it, reallocate profit, or create compliance costs. Build a rule map with regulator, legal authority, proposal/final status, effective date, affected population, transition period, enforcement, and litigation risk. Separate enacted rules from proposals and political commentary. Model policy economics explicitly. A subsidy may accelerate demand but also invite supply, compress customer prices, or expire. A tariff can protect domestic pricing while raising input cost. Local-content rules can favor some producers but increase capital intensity. Do not treat policy support as pure margin. ## 11. Customer concentration and bargaining power A concentrated customer base can create efficient distribution or dangerous bargaining power. Analyze the number of customers, purchase frequency, contract duration, switching cost, procurement sophistication, dual-sourcing behavior, qualification cycles, and share of customer cost. A component that is a small fraction of customer cost but critical to system performance can have strong pricing power even with large customers. Track concentration by revenue and by economic dependence. A supplier with 20% of revenue from one customer may still have leverage if the customer cannot qualify an alternative quickly. Conversely, a vendor with hundreds of customers may have weak power if the product is commoditized. ## 12. Supplier power and upstream fragility Map sole-source components, geographic concentration, lead times, minimum orders, tooling, qualification, commodity exposure, and financial health of suppliers. Determine whether the company can pass cost through, redesign around the input, carry buffer inventory, or vertically integrate. A strong end market can hurt margins when upstream scarcity absorbs the economics. Model the pass-through lag and the working-capital cost of inventory buffers. ## 13. Cycle mapping Build a cycle clock using orders, bookings, backlog, lead times, inventory, utilization, pricing, cancellations, capex announcements, capacity additions, and customer inventory. Identify which indicators lead revenue and which lag it. The same metric can change role across cycles. | Cycle phase | Typical evidence | Analytical risk | | --- | --- | --- | | Early recovery | Orders stabilize, inventory clears, utilization rises | Mistaking restocking for structural demand | | Expansion | Backlog, price, utilization, capex rise | Extrapolating scarcity pricing permanently | | Late cycle | Capacity announcements surge, customer inventory builds | Underestimating new supply and elasticity | | Downturn | Cancellations, lead times, price, utilization fall | Capitalizing trough earnings as permanent weakness | | Reset | Capex cuts, consolidation, inventory normalization | Missing improving supply discipline | ## 14. Industry source hierarchy Use issuer filings and competitor filings as the base, then triangulate with government data, regulators, industry bodies, customer procurement disclosures, supplier filings, customs or trade data where lawful, technical standards, patent or product data, job postings, pricing, and other alternative data. Every source has a population, definition, lag, revision behavior, and incentive. Create a source dictionary that records what each dataset measures, what it does not measure, update frequency, geographic coverage, historical revisions, survivorship risk, and known breaks. A time series without a definition history can be more dangerous than no time series. ## 15. Industry model Industry demand = Installed base additions + replacement + utilization or consumption growth, using sector-specific units. Industry supply = Existing effective capacity + qualified new capacity - closures or outages. Utilization = Demand units / Effective supply capacity, when units are consistent. Connect utilization to pricing and margins only after testing historical relationships. In some markets contracts delay pricing response. In others spot pricing moves quickly. Capacity can be geographically or technically nonfungible, making global utilization misleading. ## 16. Hidden-subsector discovery process 1. Start with a large end-market growth thesis. 1. Decompose the system into required components, software, infrastructure, test, service, finance, and maintenance. 1. Rank each layer by growth, scarcity, gross profit pool, switching cost, capital intensity, competitive concentration, and visibility. 1. Identify layers where demand growth is mechanically linked to the end market but investor attention and capacity expansion are lower. 1. Search for public companies with material pure-play or underappreciated exposure. 1. Test whether the exposure is large enough to move consolidated earnings and whether valuation already discounts the opportunity. 1. Define what would cause the bottleneck or profit pool to migrate elsewhere. The goal is not to find obscure companies for its own sake. The goal is to find a causal profit pool where end-market growth, constrained supply, customer value, and competitive structure can support returns above the cost of capital for long enough to matter. ## 17. Industry thesis-break conditions - Demand elasticity is materially worse than modeled at the required price. - Capacity expands faster or cheaper than expected, eliminating scarcity economics. - A substitute reaches sufficient performance and cost to cap price or share. - Customer vertical integration removes the supplier profit pool. - A standard or regulation shifts control to another layer. - The supposed bottleneck is solved by architecture change, not by more spending on the incumbent solution. - Capital intensity rises enough that growth no longer creates attractive incremental returns. - The market size requires implausible customer budget, infrastructure, power, labor, materials, or financing. ## 18. Worked industry case: grid-scale energy infrastructure Assume electricity demand from data centers and electrification is expected to rise. Do not stop at utilities or battery manufacturers. Map generation, interconnection, substations, transformers, switchgear, power conversion, grid-forming controls, thermal management, protection equipment, engineering, construction, maintenance, and financing. For each layer, identify lead times, capacity, technical qualification, customer concentration, pricing, and capital intensity. Suppose large transformers have multi-year lead times while several power-electronics categories have shorter lead times but higher technical differentiation. The investment question is not which category grows fastest. It is which category captures durable economic value after capacity response. Model new entrants, factory expansion lead times, customer qualification, price pass-through, and the risk that standards or architecture change the required component content. ## 19. Industry completion gate - The value chain includes all economically important layers and the customer decision process. - TAM is reconstructed bottom-up with at least one independent cross-check. - Demand and supply are expressed in compatible physical or economic units. - The binding bottleneck and capacity response are explicit. - The profit pool, not just revenue growth, is mapped across layers. - Competitive advantages are supported by causal evidence and linked to incremental returns. - Technology, regulation, substitution, and customer vertical integration have thesis-break tests. - The industry cycle has leading and lagging indicators with source definitions. - The company forecast is constrained by industry capacity, customer economics, and plausible market share. # MASTER LAB 06 - MANAGEMENT, GOVERNANCE, INCENTIVES, AND CAPITAL ALLOCATION > Purpose: evaluate management and governance through observable decisions, incentives, execution, disclosure behavior, and returns on capital rather than charisma, stock-price hindsight, or management-access impressions. The objective is to determine how decision makers are likely to allocate incremental capital and respond when evidence contradicts their plan. ## 1. Management chronology before judgment Build a chronology covering at least one relevant business cycle or the full tenure of the current leadership when shorter. Record major guidance, strategic targets, acquisitions, divestitures, restructuring plans, capital returns, leverage targets, product launches, capacity projects, executive changes, and material misses. Capture the original statement before reading the outcome. | Date | Management commitment | Controllable inputs | Outcome | Explanation after outcome | Analyst assessment | | --- | --- | --- | --- | --- | --- | | T0 | Example: 20% three-year capacity CAGR at 15%+ ROIC | Capex timing, project execution, customer contracting | Populate from evidence | Populate from later statements | Was the original process sound, was the miss foreseeable, and was accountability clear? | Avoid outcome bias. A good decision can have a bad outcome because of external shocks, and a poor decision can look good because the market bailed it out. Score the quality of the process using information reasonably available at the time. ## 2. Execution scorecard 1. Identify the 5 to 10 operating variables management directly influences. 1. Track targets versus actual outcomes on a constant definition. 1. Separate external factors such as commodity prices, FX, macro demand, and regulation from controllable execution. 1. Measure how quickly management identifies and corrects misses. 1. Record whether targets are reset transparently or definitions change after a miss. 1. Compare execution with direct peers facing the same environment. Examples of controllable variables include product roadmap timing, capacity commissioning, cost reduction, working-capital discipline, customer service, sales productivity, integration milestones, leverage, and buyback price discipline. Use sector-specific metrics, not a generic management score. ## 3. Candor and disclosure quality Candor is observable through consistency between claims and evidence, willingness to quantify problems, prompt correction of errors, stable KPI definitions, and acknowledgment of tradeoffs. A polished earnings call is not evidence of candor. Compare language before and after negative outcomes and read prepared remarks against footnotes and cash data. - Does management distinguish demand from shipment or revenue timing? - Does management reconcile changed KPI definitions? - Are missed targets explained with measurable drivers rather than broad macro language? - Are negative developments disclosed before they become unavoidable? - Are favorable and unfavorable one-time items treated consistently? - Does the company provide enough segment and cash detail for investors to test the narrative? ## 4. Incentive architecture Read the proxy compensation tables and footnotes. Map base salary, annual bonus, time-based equity, performance equity, option grants, retention awards, severance, change-in-control provisions, and special awards. Then map the performance metrics, measurement windows, thresholds, targets, maximums, discretion, and modification rights. | Incentive feature | Potential benefit | Potential distortion | | --- | --- | --- | | Revenue growth target | Encourages scale and market capture | Can encourage low-quality or acquisition-driven growth | | Adjusted EBITDA | Focuses operating profit | Can reward exclusions, underinvestment, or working-capital neglect | | EPS target | Links to per-share result | Can reward leverage, buybacks, tax timing, or acquisition accretion | | ROIC / ROE | Encourages capital efficiency | Definition can be gamed through exclusions or denominator choices | | TSR | Aligns with shareholder outcome over time | Can reward market beta and encourage short-window timing | | FCF | Focuses cash | Can encourage capex deferral or working-capital harvesting if poorly designed | | Operational milestones | Can focus strategic execution | Milestones may be subjective or detached from economic value | Recalculate performance metrics under a more economic definition. For example, include recurring restructuring, normalize working-capital timing, or include acquisition capital when evaluating return on invested capital. The question is whether management gets paid for creating durable value or for optimizing a reported metric. ## 5. Insider ownership and transactions Distinguish founder ownership, purchased shares, vested compensation, unvested awards, options, pledged shares, 10b5-1 plan sales, tax withholding, diversification sales, and open-market purchases. Gross insider ownership can overstate economic alignment when most holdings are unvested or routinely sold. Open-market buying can be informative but is not a stand-alone thesis. Evaluate transaction size relative to compensation and net worth, timing, repeated behavior, and whether multiple insiders independently purchase. For sales, avoid assuming negative intent. Focus on pattern changes and context. ## 6. Board structure and control Map board independence, tenure, expertise, committee assignments, related-party relationships, dual-class voting, classified boards, poison pills, supermajority provisions, lead independent director authority, and succession planning. Governance features matter most when they alter accountability or capital-allocation control. Read director biographies for domain expertise relevant to the company. A technically complex or regulated business should have appropriate oversight capability. Board independence on paper does not guarantee effective challenge, especially when directors have long personal or business ties to management. ## 7. Capital allocation framework Every dollar of after-tax operating cash has a destination: reinvest in the existing business, build new capacity, R&D, customer acquisition, acquire another business, repay debt, hold cash, pay dividends, repurchase shares, or return capital through another mechanism. Evaluate each use against its expected risk-adjusted return and strategic necessity. Value creation from reinvestment depends on incremental return on invested capital relative to the required return and the duration over which reinvestment opportunities persist. High historical ROIC does not guarantee high incremental ROIC. Separate legacy assets from new investment. A mature franchise can report excellent average ROIC while deploying new capital at mediocre returns. ## 8. Organic reinvestment Map growth capex, R&D, sales and marketing, product development, capacity, distribution, data, and other internal investment. Estimate the incremental revenue, margin, and capital created by those investments. Some important investments are expensed under accounting rules, so reported capex can materially understate total reinvestment. For intangible-heavy businesses, consider capitalizing a portion of R&D or customer-acquisition spending analytically when doing so improves the measurement of invested capital and incremental returns. The adjustment must be systematic and reversible, not chosen to create a preferred ROIC. ## 9. M&A track record Build an acquisition ledger for every material deal: price, financing, target revenue/profit, valuation multiple, synergy target, integration timeline, retention plan, acquired intangibles, goodwill, restructuring, impairments, divestitures, and current performance. Compare original deal claims with realized outcomes. Approximate acquired-capital return = Normalized after-tax operating profit attributable to the acquired business and realized synergies / Total acquisition capital including consideration and required integration investment. Beware of denominator erasure. An impairment does not mean the capital was never invested. Preserve original purchase price when judging historical acquisition returns. Similarly, selling a weak acquired asset does not erase the loss. ## 10. Buybacks A repurchase creates value per remaining share when shares are bought below a reasonable estimate of intrinsic value, liquidity remains adequate, and the capital has no higher-return use. Repurchases can destroy value when used to offset excessive dilution, support EPS targets, or buy expensive stock while balance-sheet risk rises. Shares retired = Cash used for repurchase / Average repurchase price, adjusted for excise taxes and transaction details where material. Net share reduction = Shares repurchased - employee/transaction shares issued. Build a historical buyback table showing dollars spent, average price, shares retired, gross employee issuance, net share change, leverage, and estimated value range at the time. Do not evaluate buybacks only by whether the stock later went up. ## 11. Dividends and leverage Dividends are appropriate when the company lacks enough high-return reinvestment opportunities and can maintain resilience. Evaluate payout through cycle, not only against current earnings. For debt policy, compare leverage with cash-flow durability, cyclicality, asset collateral, refinancing access, and strategic need for flexibility. A low-cost debt balance is not free capital. The relevant question is whether leverage increases equity risk enough to reduce strategic flexibility or force poor decisions in a downturn. Stress interest, covenant, maturity, and collateral headroom before endorsing aggressive buybacks or acquisitions. ## 12. Cash balances and optionality Separate operating cash, regulatory or restricted cash, customer or collateral balances, trapped foreign cash, and truly excess cash. A large cash balance can be valuable optionality in a cyclical or acquisition-driven industry, but persistent idle cash can dilute returns if management lacks a disciplined use. Do not automatically add all cash to enterprise value. Estimate minimum operational liquidity and the cost to distribute or access cash where relevant. ## 13. Founder-controlled and dual-class companies Founder control can support long-horizon investment and protect a strategy from short-term pressure. It can also weaken external accountability. Analyze voting control, related-party transactions, board independence, succession, capital allocation, and treatment of minority holders. Avoid ideological assumptions in either direction. ## 14. Governance event playbook - CEO/CFO departure: identify timing, stated reason, succession plan, prior internal-control or performance issues, and whether departure changes the thesis. - Auditor change: read the filing carefully for disagreements, reportable events, and timing. - Board refresh: assess whether new directors add relevant expertise or merely change optics. - Activist involvement: separate operational ideas, capital-structure proposals, governance demands, and short-term financial engineering. - Related-party transaction: reconstruct economics and compare with arm-length alternatives. - Compensation redesign: test whether new metrics better align with long-term value or simply lower the performance bar. ## 15. Management red-team questions - Which promise has management made repeatedly without delivering? - Which metric improved while the underlying economic driver worsened? - Which acquisition, project, or product would management be least willing to admit failed? - Where does compensation reward a result management can influence through accounting or timing? - What capital-allocation decision would be hardest to reverse in a downturn? - Which disclosure became less specific after performance deteriorated? - What would a skeptical former employee, competitor, supplier, or customer say about execution? - What evidence would cause the analyst to upgrade management quality, not only downgrade it? ## 16. Worked capital-allocation case Fictional company Meridian Systems generates $500 million of annual normalized FCF. It can invest $200 million internally at an estimated 18% incremental ROIC, repay debt costing 6%, acquire a competitor for $1.2 billion at 14x EBITDA with projected cost synergies, or repurchase shares at a valuation implying a 5% FCF yield. The company has moderate cyclicality and debt equal to 2.5x EBITDA. 1. Fund the high-return internal projects first, subject to evidence that 18% incremental ROIC is real and scalable. 1. Stress liquidity before choosing debt reduction versus repurchase. 1. Value the acquisition standalone and with probability-weighted synergies rather than using EPS accretion. 1. Compare the repurchase yield with the company cost of equity and with internal project returns. 1. Model a recession case to determine whether the chosen capital allocation leaves enough flexibility. 1. Document why the ranking changes if the stock price, acquisition price, or project ROIC changes. The exercise is not intended to produce one universal ordering. It demonstrates that capital allocation is contingent on return, risk, valuation, liquidity, and opportunity set. The analyst should be able to show the decision boundary at which one use of capital becomes superior to another. ## 17. Completion gate - Management is evaluated through a dated record of decisions and outcomes rather than impression. - Controllable execution is separated from external factors. - Compensation metrics are reconstructed and tested for economic alignment. - Insider ownership and transactions are interpreted by economic exposure and context. - Board structure and governance provisions are mapped to actual decision rights. - Organic reinvestment, M&A, debt, dividends, buybacks, and cash are compared through expected returns and resilience. - Historical acquisition and repurchase outcomes preserve original capital invested instead of erasing mistakes. - At least one adverse capital-allocation scenario is modeled through liquidity, dilution, and valuation. - The analyst can state which future management action would materially change the assessment. # MASTER LAB 07 - ALTERNATIVE DATA, EXPERT CALLS, CHANNEL CHECKS, AND SIGNAL VALIDATION > Purpose: use nontraditional evidence to test public-company hypotheses without confusing noisy correlation with fact or crossing legal and compliance boundaries. Alternative data should reduce uncertainty around a defined question. It should never become a collection of interesting charts with no causal link to revenue, margins, cash flow, or risk. ## 1. Start with the hypothesis and the causal chain Write the precise hypothesis before collecting alternative data. Example: "North American enterprise demand for Product X is decelerating because renewal-seat contraction is offsetting list-price increases." Then map the causal chain from customer behavior to observable signal to reported KPI to financial statement. A data source is useful only if the chain is plausible and the signal arrives early enough to matter. Signal value ≈ Expected decision impact x probability the signal resolves uncertainty x timeliness - data cost - false-positive cost - compliance risk. This expected-value framing prevents expensive datasets from becoming status symbols. A cheap source that resolves a material assumption can be more valuable than a sophisticated feed with weak linkage to the thesis. ## 2. Compliance and MNPI boundary Alternative-data work must operate within applicable law, contractual terms, privacy rules, platform terms, and the firm compliance program. Do not solicit or use material nonpublic information. Do not encourage an expert to breach an employer confidentiality duty, customer agreement, government restriction, or other obligation. When the line is unclear, stop and escalate to compliance. - Define prohibited topics before expert calls. - Use approved expert-network and research procedures where required. - Avoid asking for unreleased earnings, bookings, customer-specific confidential spend, unannounced product details, material contract awards, or other nonpublic company information. - Record the source, date, population, and method used to collect a signal. - Separate public aggregation from information that is nonpublic because of the source or method. - Respect robots.txt, access controls, rate limits, licensing, and contractual restrictions for web data. ## 3. Expert-call design Expert calls are hypothesis tests, not fishing expeditions. Build a question tree from broad market structure to specific observable behavior. Ask for process, ranges, comparisons, and examples rather than confidential point estimates. Use multiple independent experts with different positions in the value chain to reduce individual bias. 1. Define the hypothesis and the evidence that would change it. 1. Screen experts for relevant role, geography, customer size, product generation, and time period. 1. Write prohibited topics and conflict notes. 1. Begin with background and definitions so terminology is consistent. 1. Ask open-ended process questions before revealing your hypothesis. 1. Probe contradictions and ask for examples or ranges. 1. At the end, ask what would make the expert wrong and which data source they trust most. 1. Code notes by claim, source role, confidence, and whether the claim is independently corroborated. ### Good versus weak questions | Weak question | Better question | | --- | --- | | Are bookings weak this quarter? | How has the customer purchase process changed over the past six to twelve months, and what public or observable indicators would show that change? | | What will Company A report? | Which vendors are gaining or losing consideration in current evaluations, and what factors drive the decision? | | How much did your company spend? | How is budget allocation changing across the category, and what ranges or priorities are observable across customers like yours? | | Is the new product good? | Compared with the prior generation and alternatives, where does the product perform better or worse on cost, reliability, deployment, and support? | ## 4. Customer research Segment customers by size, geography, use case, maturity, contract type, and strategic importance. A sample of five enthusiastic early adopters does not represent the installed base. Ask about purchase trigger, procurement cycle, alternatives considered, price paid, implementation effort, realized value, renewal, expansion, churn, and support. When the company sells through channels, distinguish sell-in from sell-through. Distributor inventory can make supplier revenue look strong while end demand weakens. Track channel inventory, weeks of supply, discounting, return rights, and vendor incentives where observable and compliant. ## 5. Supplier research Suppliers can reveal order patterns, lead times, capacity, input shortages, redesign activity, customer concentration, and cost pressure. However, one supplier may see only a small program or may be losing share. Weight the evidence by exposure and corroborate with other suppliers, customer commentary, inventory, and company disclosures. Map whether a supplier is sole-source, dual-source, qualified backup, commodity, or highly specialized. Changes in supplier lead time can reflect demand, capacity expansion, inventory normalization, or bargaining power. Do not assume one cause. ## 6. Job postings and workforce data Hiring data can indicate growth priorities, geographic expansion, product investment, sales capacity, or restructuring. It can also be noisy because postings are duplicated, evergreen, cancelled, outsourced, or filled internally. Build a baseline and track unique normalized roles, not raw posting counts. Normalized hiring index = Unique active roles adjusted for duplicate postings, location, function, seniority, and known evergreen behavior. Map functions to economic drivers. Sales hiring may precede bookings only after ramp and quota attainment. Engineering hiring may signal product investment but not near-term revenue. Manufacturing hiring may be constrained by capacity commissioning. Backtest the lead relationship before using it in a forecast. ## 7. Web traffic, app data, search interest, and engagement Digital activity can be useful for consumer, marketplace, software, and media businesses, but source methodology often changes. Track data-provider coverage, bot filtering, device mix, geography, panel composition, and historical revisions. A traffic increase can reflect marketing spend rather than monetization or retention. Monetization bridge example: Revenue = Active users x transactions per user x average transaction value x take rate, or another business-specific decomposition. Use digital signals to test components of the bridge, not to substitute for it. If traffic rises but conversion and monetization fall, revenue may not follow. If the company shifts activity from web to app, a single-channel series can give a false decline. ## 8. Pricing, promotions, and inventory scraping Create a product dictionary with SKU, model, region, channel, list price, observed price, discount, availability, delivery time, seller, and timestamp. Preserve raw observations. Separate permanent price change from promotional discount, bundle, financing subsidy, coupon, channel markdown, and mix. For inventory, distinguish listed availability from actual inventory. E-commerce stock indicators, shipping times, dealer counts, and SKU availability can be proxies. Backtest which measure predicts reported units or channel commentary. Do not interpret one retailer as the whole market. ## 9. Trade, shipment, geospatial, and physical data Customs, shipping, satellite, foot traffic, power use, construction, and other physical data can be valuable when they map to a real asset or flow. The key risks are classification errors, incomplete geography, transshipment, entity mapping, revisions, and changes in reporting coverage. Create entity and unit maps before aggregating. A shipment record may be counted by weight, value, containers, or units and can include internal transfers. A satellite observation can show activity but not necessarily ownership or profitability. Triangulate with company capacity and customer demand. ## 10. Survey design A survey must define the target population, sampling frame, response rate, weighting, question order, and confidence limits. Convenience samples can produce precise-looking but biased numbers. Use neutral wording and separate awareness, consideration, trial, usage, satisfaction, retention, and willingness to pay. Report the sample and uncertainty with every conclusion. A statistically significant change can still be economically immaterial. Conversely, a small sample can be useful qualitatively when the effect is large and corroborated by other evidence. ## 11. Alternative-data validation protocol 1. Define the exact reported KPI or economic driver the signal is supposed to predict. 1. Freeze the data vintage to what would have been available at the historical decision date. 1. Normalize the signal for methodology changes, coverage, seasonality, and obvious confounders. 1. Backtest lead and lag relationships across multiple periods, including weak periods. 1. Measure false positives and false negatives, not only correlation. 1. Test whether the relationship survives out-of-sample periods or peer companies. 1. Document revisions and whether historical data are restated by the vendor. 1. Set a minimum effect size required before the signal changes the model. 1. Continue tracking signal performance after deployment and retire signals that decay. A high in-sample correlation is not enough. Measure out-of-sample forecast error, directional hit rate, revision stability, and incremental explanatory power versus the existing model. ## 12. Avoiding look-ahead and survivorship bias Historical alternative datasets are often cleaner than they were in real time because providers revise entity mapping, remove bad records, or add coverage. A valid backtest uses the vintage available at the time. Survivorship bias appears when failed products, stores, companies, or web properties disappear from the sample. Preserve dead entities and historical coverage where possible. ## 13. Confounders and causal skepticism Alternative data rarely come from randomized experiments. Search traffic can rise because of a scandal, marketing campaign, product launch, or seasonal event. Job postings can rise because turnover increased. Prices can fall because mix changed. Always list at least two alternative causes and identify a second data source that helps distinguish them. ## 14. Signal weighting and evidence matrix | Signal | Coverage | Lead time | Historical reliability | Causal proximity | Current direction | Model impact | | --- | --- | --- | --- | --- | --- | --- | | Example customer survey | Enterprise buyers, North America | 1-2 quarters | Medium | High | Weakening | Reduce new bookings only after corroboration | | Example web traffic | Global public web | Weeks | Low-Medium | Low | Rising | No direct model change | | Example supplier lead time | Critical component | 1-3 quarters | High after backtest | Medium-High | Falling | Test inventory normalization and pricing | Do not average signals mechanically. Weight by relevance to the specific driver, population coverage, lead time, reliability, and causal proximity. A high-quality primary-source operating KPI should usually outrank a noisy digital proxy. ## 15. Worked channel-check case A consumer-electronics company reports strong channel shipments and guides to continued growth. Retail scraping shows rising discounting and faster delivery times. Two distributors report higher weeks of inventory, while search interest remains strong. Build competing explanations: healthy demand with supply normalization, or sell-in exceeding sell-through. The next evidence should include retailer sell-through, returns, channel inventory, company receivables, promotional allowances, and production schedules. Quantify both cases. In the adverse case, model a one-quarter channel correction, lower unit shipments, promotional margin pressure, and working-capital release. In the benign case, model lower lead times with stable sell-through and normalized inventory. Do not change the long-term thesis solely because one high-frequency indicator moves. ## 16. Completion gate - Every alternative-data source is tied to a defined hypothesis and financial driver. - Collection method, population, coverage, terms, revisions, and known biases are documented. - Expert work follows legal and compliance boundaries and does not solicit MNPI. - Signals are validated against reported outcomes using real-time vintages where possible. - False positives, false negatives, and methodology changes are tracked. - No signal changes a material forecast without a pre-defined effect-size threshold and causal explanation. - Primary-source evidence retains priority when it directly measures the same underlying variable. - The analyst can explain what would cause the signal relationship to break and when the signal should be retired. # MASTER LAB 08 - AI RESEARCH ENGINEERING, AUTOMATION, EVALUATION, AND MODEL RISK > Purpose: design an AI-assisted public-equity research system that is faster than a manual workflow without weakening evidence quality, reproducibility, compliance, or human accountability. The system must preserve source provenance, use deterministic computation for math and state changes, test model outputs against known answers, and require human approval where judgment or external action matters. ## 1. Architecture: separate retrieval, transformation, calculation, decision, and action Do not build one prompt that searches, interprets, calculates, writes the model, decides the investment view, and takes an action. Separate the system into layers with explicit permissions. This reduces error propagation and makes failures diagnosable. | Layer | Function | Preferred control | | --- | --- | --- | | Source ingestion | Acquire filings, exhibits, transcripts, regulator data, market data, approved alternative data | Immutable raw store, timestamp, document hash, licensing and access controls | | Parsing / extraction | Convert source documents into text, tables, structured fields | Schema validation, page/section spans, parser version, confidence | | Retrieval | Select relevant approved source passages | Source whitelist, metadata filters, provenance, recency and version rules | | Language-model transformation | Classify, summarize, compare, extract, generate hypotheses | Structured output, evidence requirement, no hidden arithmetic | | Deterministic calculation | Reconcile statements, calculate ratios, build valuation, write files | Tested code or formulas, units and sign checks, reproducible inputs | | Decision support | Compare evidence, scenarios, thesis, risk | Human-owned assumptions and sign-off | | External action | Send email, update production model, publish research, trade or other action | Explicit authorization, scoped permissions, logging, approval gates | ## 2. Immutable source layer The raw-source layer is the foundation. Store the exact document used, not only extracted text. Minimum metadata should include issuer/entity identifier, source type, accession or document identifier, URL or connector reference, publication time, period end, retrieval time, file hash, parser version, language, page count, and whether the source supersedes a prior version. Never silently replace a filing with an amended filing in historical research. Preserve both and mark the amendment relationship. A model built on an earlier version must remain reproducible even after the source changes. ### Canonical source record | Field | Example purpose | | --- | --- | | entity_id | Stable company mapping independent of ticker changes | | document_id | Accession, regulator ID, or internal source identifier | | source_type | 10-K, 10-Q, 8-K, proxy, debt agreement, transcript, regulator release | | period_start / period_end | Financial-period alignment | | published_at / retrieved_at | Point-in-time reproducibility and stale-data control | | hash | Detect source or parsing changes | | parser_version | Reproduce extraction behavior | | raw_location | Immutable original file or approved repository reference | | supersedes | Link amendments or corrected data | ## 3. Entity resolution and identifier discipline Ticker is not a sufficient primary key. Tickers change, can be reused, and differ by exchange. Maintain stable entity identifiers plus CIK or relevant regulator identifier, LEI where useful, exchange/ticker history, legal name, former names, subsidiaries, brands, and major acquired or divested entities. Entity mapping errors can contaminate every downstream result. 1. Resolve issuer identity before retrieval. 1. Resolve period and fiscal calendar before comparing data. 1. Map segments and subsidiaries with effective dates. 1. Preserve historical tickers and names for alternative-data joins. 1. Flag ambiguous names for human review instead of fuzzy-matching silently. ## 4. Document parsing and table extraction Prefer structured regulator data such as XBRL for machine-readable facts, but always reconcile material values to the human-readable filing because tags can be misapplied, dimensions can differ, and presentation context matters. For tables, retain row labels, column labels, units, scale, period, and footnotes. A number without those fields is unsafe. When OCR is unavoidable, attach OCR confidence and route low-confidence fields to review. Never use OCR output as silently authoritative when a native text or structured source exists. ## 5. Structured extraction schema Require model extraction to emit a schema that captures the semantics needed for audit. Free-form prose is insufficient for model writes. | Field | Required behavior | | --- | --- | | entity | Stable entity identifier and displayed name | | metric | Canonical metric plus source label | | value | Raw numeric or categorical value | | unit / scale | USD, shares, %, units, thousands, millions, etc. | | period | Start/end date and fiscal label | | segment / dimension | Segment, geography, product, customer class if applicable | | source_id | Exact document version | | source_span | Page, section, table, or exact text span | | definition | Issuer definition when nonstandard | | confidence | Extraction confidence separate from economic confidence | | transformation | Any calculation or mapping applied after extraction | Reject or quarantine records missing period, unit, or source span when those fields are necessary. Most catastrophic financial-data errors are mundane: wrong period, wrong scale, wrong sign, wrong segment, or wrong entity. ## 6. Retrieval-augmented generation that preserves provenance Retrieval should prioritize approved primary sources and exact versions. Chunking should respect document structure such as sections, tables, footnotes, and exhibits. A chunk must retain document metadata and location. The answer layer should cite the exact retrieved evidence used for each factual claim. - Filter by issuer and reporting period before semantic ranking when the question is period-specific. - Prefer current filing plus directly comparable prior filing for definition-drift questions. - Use debt agreements for covenant terms rather than summaries when available. - Retrieve the footnote table and surrounding narrative together when definitions depend on context. - Require the system to state when evidence is absent rather than infer a fact from nearby text. ## 7. Prompt injection and untrusted content Treat instructions embedded inside retrieved webpages, documents, filings, emails, or datasets as untrusted data. The system prompt and authorization layer determine behavior, not text found in a source. Retrieval content should never be able to grant new tool permissions, override compliance rules, expose credentials, or trigger external actions. 1. Separate system instructions from retrieved content. 1. Escape or delimit source text clearly. 1. Disable or restrict action tools during untrusted-document analysis unless explicitly needed. 1. Allow only approved domains, connectors, or repositories for automated retrieval where appropriate. 1. Log tool calls and authorization decisions. 1. Test the system with adversarial source text that attempts to redirect behavior. ## 8. Deterministic math versus language-model judgment Use code or spreadsheet formulas for arithmetic, reconciliations, statement roll-forwards, unit conversions, valuation, optimization, and file writes. A language model can propose the mapping or formula, but deterministic execution should calculate the result. This separates reasoning uncertainty from arithmetic correctness. Example: the model may identify that a lease liability should be included in a debt-like claim bridge. Code should retrieve the balance and perform the enterprise-to-equity calculation. The language model then explains the result and its assumptions. ## 9. Numeric reconciliation engine Create automatic checks before extracted data can enter the production model. Checks should operate at the statement, footnote, segment, and time-series levels. - Assets = liabilities + equity within a defined tolerance. - Segment revenue plus eliminations = consolidated revenue. - Cash-flow change in cash = balance-sheet change in cash after FX/reclassification. - Current period + prior comparable period agree with source table headings and fiscal calendar. - Units and scale are consistent within a calculation. - Sign conventions are normalized explicitly. - Amended filings supersede but do not erase prior versions. - XBRL value is cross-checked with presented filing for material facts. - Share counts distinguish basic, diluted weighted average, and period-end shares. - Debt balances reconcile with maturity and instrument tables. ## 10. Definition-drift detection Automate comparison of KPI definitions, segment descriptions, non-GAAP reconciliations, risk factors, accounting policies, and guidance language across periods. Use text diff plus semantic classification. The output should show the exact old and new language and classify the change as formatting, clarification, scope change, methodology change, or potentially thesis-relevant removal/addition. Do not let an AI system summarize a definition change without preserving the source text. Reviewers need the original wording to judge significance. ## 11. Filing-event pipeline 1. Detect new filing or approved source event. 1. Fetch and hash the source. 1. Classify form/event type and issuer. 1. Parse text, tables, XBRL, and exhibits. 1. Run materiality triage based on known thesis variables and event type. 1. Diff against the prior comparable filing. 1. Extract changed financial facts and definitions into a structured schema. 1. Run reconciliation and anomaly checks. 1. Generate a draft impact memo showing facts, calculations, open questions, and potential model lines affected. 1. Require human approval before writing material assumption changes to the production model. 1. Version the model change, memo, evidence, and reviewer decision. ## 12. Earnings-call and transcript workflow Transcripts are management claims, not primary audited facts. Use AI to classify questions, guidance, KPI changes, contradictions, and new risk statements, then reconcile each material claim with the filing and earnings release. Track what management was asked, what was not answered, and whether the same issue recurs. A useful transcript system creates a management-claim ledger with claim date, topic, metric, time horizon, confidence language, later outcome, and contradiction status. This supports management-quality analysis and reduces hindsight bias. ## 13. Spreadsheet and model-write controls Automated model writes are high risk because one wrong cell can propagate through valuation. Use a staging layer. The system proposes changes with source, old value, new value, unit, period, cell/range, formula impact, and confidence. Human review approves or rejects material writes. After write, run the entire model check suite and record the diff. - Never overwrite formulas with values unless the workflow explicitly expects an input cell. - Reject writes to protected historical or formula ranges. - Validate workbook version before applying a patch. - Calculate before-and-after model outputs and flag large changes. - Preserve a rollback copy and change log. - Require deterministic address mapping rather than asking a model to guess cell locations from prose. ## 14. Model and prompt versioning Record model provider/version, reasoning configuration where relevant, system prompt version, extraction schema version, retrieval configuration, code commit or script hash, source-set version, and evaluation-set version. An AI result that cannot be reproduced under the same versioned inputs is not audit-ready. When an upstream model changes, run regression tests before promoting it to production. A newer model can improve prose while degrading table extraction or citation precision. ## 15. Evaluation framework Build a golden set of representative tasks with known expected outputs. Include simple and adversarial cases, tables, footnotes, amendments, unusual fiscal years, segment recasts, negative values, parenthetical amounts, multi-currency disclosures, and documents containing misleading instructions. | Metric | Definition | | --- | --- | | Extraction accuracy | Exact or tolerance-based correctness of values, units, period, and dimensions | | Citation support | Percentage of factual outputs supported by the cited source span | | Numeric reconciliation pass rate | Percentage of outputs that pass statement and roll-forward checks | | Hallucination rate | Unsupported factual assertions per task or per 1,000 claims | | Definition accuracy | Correct identification of issuer-defined KPI meaning and changes | | False-positive anomaly rate | Benign items incorrectly escalated as red flags | | False-negative rate | Known material issues missed by the workflow | | Analyst correction rate | Share of outputs requiring substantive human correction | | Net automation yield | Analyst time saved after review and correction cost | | Regression stability | Change in task performance after model, prompt, parser, or retrieval update | ## 16. Evaluation-set construction 1. Select tasks from real historical workflows and anonymized or public sources. 1. Include multiple industries and filing structures. 1. Create ground truth from senior analyst review or deterministic calculations. 1. Freeze source versions and expected answers. 1. Split development and holdout sets. 1. Add adversarial and edge cases continuously from production failures. 1. Score factual fields separately from narrative quality. 1. Require minimum thresholds by task type before production deployment. A single average score can hide a dangerous weakness. A system that is 99% accurate on narrative classification but 93% accurate on unit scale may still be unsafe for automated financial model writes. Set task-specific gates. ## 17. Confidence and abstention Confidence should be calibrated to the task and evidence, not copied from the language model self-assessment. Use objective signals such as source agreement, parser confidence, schema completeness, reconciliation pass, ambiguity count, and historical error rate. When evidence conflicts or key fields are missing, the correct output is an escalation or abstention. Example confidence components: source quality, extraction reliability, reconciliation status, definition match, and independent corroboration. Document the weighting rather than treating confidence as a magic probability. ## 18. Human approval gates | Action | Default gate | | --- | --- | | Retrieve and summarize public filing | May be automated with provenance and logging | | Extract values into staging table | May be automated with validation | | Change a material forecast assumption | Human approval | | Change production valuation model | Human approval plus model checks | | Publish investment conclusion | Human accountable reviewer | | Use expert/channel information with compliance implications | Compliance-approved process and human review | | Send external communication or take market action | Explicit authorization under firm policy | Human review should be targeted, not ceremonial. The reviewer should see the source span, proposed conclusion, calculation, uncertainty, and model impact. A checkbox without enough evidence to evaluate the change is not a meaningful control. ## 19. Private data, credentials, and access Use least-privilege access. Keep secrets in an approved secret manager rather than prompts, notebooks, or documents. Separate public research data from confidential firm data and client information. Log access to sensitive sources. Respect retention, regional, and contractual requirements. Do not send confidential or restricted material to an AI service unless the firm has approved the service, data handling, retention, and contractual terms. Model capability does not override data-governance policy. ## 20. Cost, latency, and model routing Use the least expensive and fastest model that reliably passes the task evaluation. Simple classification or extraction may not require the most capable reasoning model. Complex contradiction analysis or ambiguous accounting may. Route by task difficulty, risk, and required latency. Automation economics = Analyst time saved + error reduction value - model/API cost - infrastructure cost - review cost - expected error cost. Cache immutable retrieval and deterministic calculations. Do not pay repeatedly to re-summarize an unchanged filing when the stored structured result is valid. Recompute only when source, prompt, model, schema, or analytical question changes. ## 21. Multi-agent or workflow orchestration Multiple agents do not automatically improve quality. Use specialized stages only when separation creates a check or expertise benefit. One useful pattern is extractor, reconciler, skeptic, calculator, and reviewer. Each stage receives limited inputs and produces a structured output. The skeptic should search for contradictory evidence, not merely rephrase the first agent. Avoid agent loops with no stop condition. Define a maximum number of retries, escalation criteria, and what constitutes resolved evidence. Store each stage output so the final answer can be audited. ## 22. Research memory and decision journal Separate factual memory from thesis memory. Factual memory stores versioned source facts and definitions. Thesis memory stores analyst assumptions, probability judgments, scenario choices, and reasons for changes. Never allow a generated summary to overwrite the underlying source record. When new evidence arrives, compare it with the prior thesis state. Record what changed, why it changed, whether the change was anticipated, and which forecast lines moved. This prevents the system from rewriting history after the fact. ## 23. Automated contradiction hunting 1. Extract management claims and guidance from prior periods. 1. Identify current reported outcomes for the same metric or promise. 1. Compare current narrative with prior wording. 1. Search competitor, customer, supplier, and regulator sources for evidence pointing the other direction. 1. Generate alternative explanations and list evidence needed to distinguish them. 1. Route material contradictions to the analyst with source spans and quantified model sensitivity. The contradiction engine should not decide that management is wrong. It should make it hard for the analyst to miss evidence that challenges the current thesis. ## 24. AI-assisted valuation and scenario generation AI can generate candidate scenario narratives, sensitivity dimensions, and missing-risk questions, but the operating model and valuation calculations should remain deterministic. Ask the system to explain the causal chain for each scenario and to identify internal inconsistencies such as growth without reinvestment, margin expansion despite falling utilization, or a debt path that violates liquidity constraints. Use AI to search for hidden assumptions, not to manufacture precise target prices. The senior analyst owns the selected assumptions and probability weights. ## 25. Production incident response Treat material AI errors like model incidents. Preserve the failed input and output, classify root cause, determine whether any published or model output was affected, correct the result, update the evaluation set, patch the workflow, and rerun regression tests. Do not quietly fix the one example without addressing the class of failure. | Failure class | Example | System fix | | --- | --- | --- | | Retrieval | Wrong period filing retrieved | Metadata filter and period assertion | | Extraction | $1.2 billion read as $1.2 million | Unit/scale schema plus reconciliation | | Entity resolution | Subsidiary confused with parent | Stable entity IDs and ambiguity review | | Reasoning | One-time gain treated as recurring | Accounting rules plus skeptic/reviewer stage | | Calculation | DCF arithmetic error | Deterministic code and unit tests | | Citation | Claim cites nearby but unsupported passage | Claim-level evidence validation | | Action | Wrong model cell overwritten | Staging, protected ranges, explicit approval and rollback | ## 26. Worked AI workflow: 10-Q to model update 1. Watcher detects the new 10-Q and stores the exact file, accession, timestamp, and hash. 1. Parser extracts statements, footnotes, tables, XBRL facts, and exhibits. 1. Entity/period validator confirms issuer, fiscal quarter, currency, and scale. 1. Reconciliation engine ties statements and segments, flagging any discrepancy. 1. Definition-diff system compares KPIs, segments, and accounting policies with prior filing. 1. Language model extracts changed facts and produces source-linked structured records. 1. Deterministic code calculates growth, margins, working capital, cash flow, share count, and variance versus the existing model. 1. Skeptic stage searches for contrary footnote, cash, or competitor evidence. 1. System produces a proposed model-change package with old value, new value, source, reason, confidence, and sensitivity. 1. Analyst approves, rejects, or edits each material assumption change. 1. Model patch applies only approved inputs, recalculates, and runs all checks. 1. A draft earnings note is generated from the approved state and cited evidence. 1. Reviewer signs off and the system archives source, prompts, model versions, changes, and final output. ## 27. AI completion gate - Every material factual output can be traced to an immutable source version and exact supporting span. - Math, model writes, and reconciliations are deterministic and tested. - Entity, period, unit, scale, and definition errors are actively checked. - Retrieved content cannot change system permissions or execute instructions. - The system has a representative golden evaluation set with task-specific thresholds. - Model, prompt, parser, schema, source, and code versions are logged. - Production failures become regression tests. - Human approval is required for material assumptions, publication, compliance-sensitive evidence, and external actions. - The workflow measures net analyst time saved after correction cost, not just raw automation volume. - The analyst remains accountable for the conclusion and can reconstruct exactly why the system produced it. # MASTER LAB 09 - RISK, LIQUIDITY, DISTRESS, DECISION JOURNALING, AND RESEARCH SIZING INPUTS > Purpose: convert risk from a generic list into quantified transmission mechanisms, leading indicators, liquidity consequences, and thesis-break conditions. Risk analysis is not complete when the analyst can name what might go wrong. It is complete when the analyst can show how the event reaches revenue, margins, cash, financing, dilution, and value, how quickly it arrives, and what evidence will reveal it. ## 1. Risk taxonomy | Risk class | Examples | Required analytical output | | --- | --- | --- | | Demand | Customer slowdown, churn, budget cuts, substitution | Volume or retention sensitivity, leading indicators, cash effect | | Pricing / mix | Discounting, commoditization, adverse mix | Price-volume-mix bridge and incremental margin | | Supply / operations | Input shortage, yield, outage, quality failure | Capacity loss, cost, working capital, recovery time | | Technology | Obsolescence, standard shift, architecture change | Share, price, stranded capital, terminal economics | | Financial | Leverage, rates, refinancing, covenant, liquidity | Runway, maturity, interest, dilution or recovery | | Accounting / control | Estimate bias, restatement, material weakness | Normalized earnings, confidence, financing and governance effect | | Regulatory / legal | Rule change, litigation, antitrust, reimbursement | Affected revenue/cost, timing, probability, remedies | | Governance | Poor capital allocation, incentives, related parties | Expected return erosion, dilution, control response | | Valuation | Duration, rate sensitivity, terminal concentration | Value sensitivity and market-implied expectations | | Exogenous | FX, commodity, geopolitical, disaster | Scenario transmission and hedges | ## 2. Risk register construction For every material risk, record the event, causal transmission, exposed metric, leading indicator, time horizon, severity, likelihood range, mitigants, management control, financial model line, and thesis-break threshold. Avoid single-number probabilities when evidence does not support them. Use qualitative bands or ranges with rationale. | Field | Example | | --- | --- | | Risk event | Top customer delays next-generation qualification | | Transmission | Volume slips, utilization falls, fixed-cost absorption weakens | | Leading indicator | Qualification milestone, customer capex, supplier order pattern | | Financial exposure | Revenue -8%, EBIT margin -350 bps in bear case | | Liquidity effect | Lower CFO, inventory build, no covenant breach in base bear | | Mitigant | Alternative customer ramp, flexible capex | | Thesis break | Second major customer delay extends underutilization beyond funded runway | | Next review | Customer filing, company 10-Q, supplier data, project milestone | ## 3. Risk interaction and second-order effects Risks are not independent. A demand shock can reduce utilization, pressure margins, increase inventory, weaken cash flow, raise leverage, reduce covenant headroom, force capex cuts, delay product development, and worsen competitive position. Build chains rather than isolated sensitivities. Compound downside is not generally equal to the sum of isolated percentage shocks because variables interact through operating leverage, working capital, financing, and dilution. When scenarios contain correlated risks, make the causal link explicit. Do not apply every historical worst case simultaneously unless the combined state is economically coherent. ## 4. Liquidity runway Liquidity is cash plus truly available committed financing less required uses. Do not count an undrawn revolver at face value if covenants, borrowing bases, collateral, material adverse change clauses, or other restrictions may reduce availability. Separate unrestricted cash from restricted or operationally trapped balances. Liquidity runway = Unrestricted cash + usable committed capacity + expected operating cash generation - mandatory debt service - required capex - working-capital needs - restructuring/other committed uses. Model monthly or quarterly liquidity when runway is tight. Annual periods can hide a midyear cash trough. Include seasonal working capital, interest payment dates, tax payments, customer prepayments, inventory builds, and maturity dates. ## 5. Debt maturity and refinancing map For each instrument, record principal, maturity, coupon, floating base rate, spread, floor, secured status, guarantees, collateral, covenants, call schedule, amortization, and market price or yield when relevant. Distinguish legal maturity from expected refinancing need. A company with no near-term maturity can still be risky if interest coverage collapses or collateral availability shrinks. Conversely, a large maturity may be manageable if the business has strong cash generation and capital-market access. Analyze both capacity and timing. ## 6. Covenant stress Covenant headroom = Covenant limit - calculated covenant metric, using the exact contractual definition and permitted add-backs. Do not calculate covenants from headline GAAP or adjusted EBITDA unless that matches the credit agreement. Debt documents can define EBITDA, debt, cash, unrestricted subsidiaries, baskets, and cure rights differently. Build the contractual calculation and then an economic leverage measure separately. - Test operating downside before assuming lender waiver. - Identify springing covenants tied to revolver utilization. - Track restricted-payment and debt-incurrence baskets when capital allocation matters. - Identify cross-default and acceleration provisions. - Review collateral coverage and structural subordination. ## 7. Interest-rate and refinancing sensitivity Model floating-rate exposure, hedges, maturities, and new-debt spreads. The refinancing rate is a function of the risk state at refinancing, not simply the current coupon. In a downside case, EBITDA can fall at the same time spreads widen. Approximate interest sensitivity = Floating-rate debt x rate change, adjusted for hedges, floors, caps, and debt changes. For highly levered companies, iterate interest and debt because lower cash flow increases borrowing and borrowing increases interest. Confirm the model converges or use explicit period logic. ## 8. Distress waterfall If the company can exhaust liquidity, move from standard equity valuation to claim-priority analysis. Identify administrative claims, secured debt, structurally senior claims at subsidiaries, unsecured debt, preferred stock, and common equity. Map collateral and guarantees. Build a going-concern enterprise value range and allocate value according to legal and economic priority, recognizing that negotiated outcomes can differ. Illustrative recovery = Value available to class / Allowed claims in class, capped at 100%, after senior claims and required priority costs. Equity can have option value even when current enterprise value is below debt, but that does not make the option a base-case value. Estimate probability, time, dilution, and cash required to survive until the upside state. ## 9. Permanent impairment versus volatility Classify negative outcomes into temporary earnings volatility, thesis delay, permanent reduction in normalized cash flow, permanent increase in required return, or capital-structure impairment. Market-price decline alone does not identify which occurred. The research response should depend on the mechanism. A lost quarter from shipment timing may delay value realization. Loss of a patent, customer, cost advantage, or technology standard can reduce long-term value. A liquidity event can transfer value from equity to creditors even if the operating business survives. ## 10. Thesis-break design Write thesis-break conditions before the evidence arrives. A good condition is specific, observable, causal, and tied to value. Avoid vague statements such as "growth slows". Define the level, duration, or mechanism that invalidates the original thesis. | Weak thesis break | Better thesis break | | --- | --- | | Revenue growth below plan | Two consecutive quarters of net retention below 95% caused by customer churn rather than FX, with no offsetting acquisition or pricing evidence, reducing mature growth below the reverse-DCF requirement | | Competition increases | Two qualified competitors reach cost/performance parity and win more than 20% of new designs, eliminating the modeled pricing premium | | Balance sheet worsens | Stress case shows unrestricted cash below minimum operating need before the next maturity with no committed financing capacity | ## 11. Pre-mortem Before finalizing a positive thesis, assume the investment thesis failed three years later. Write the most plausible causal stories. Then identify which evidence should have been visible earlier. This surfaces neglected mechanisms and improves the monitoring plan. 1. Assume normalized value is 50% below the current base case. 1. List five distinct mechanisms that could produce the outcome. 1. For each mechanism, identify the earliest observable indicator. 1. Check whether the model currently contains that indicator and sensitivity. 1. Add missing risks only if they are economically plausible and material. ## 12. Decision journal Record the decision question, evidence available, selected assumptions, scenario probabilities, valuation range, strongest disconfirming evidence, unresolved items, thesis-break conditions, and next expected datapoints. Date the journal. Future review should compare actual evidence with the prior record, not with a reconstructed memory of what the analyst meant. When the view changes, label the reason: new fact, corrected fact, definition change, modeling error, probability update, valuation input change, portfolio-context change, or thesis break. This produces an error taxonomy over time. ## 13. Forecast error decomposition After each reported period, decompose forecast error into volume, price, mix, margin, working capital, tax, share count, financing, and timing. Then classify whether the error came from source data, model mechanics, assumption quality, unexpected event, or judgment. Repeated errors in one category should change the research process. Forecast error = Actual - Prior forecast, decomposed through the driver tree rather than only the headline EPS difference. ## 14. Research-team sizing inputs Research can provide inputs relevant to position sizing without making the mandate decision. Useful inputs include expected value range, downside severity, thesis-break probability, liquidity, catalyst timing, duration, balance-sheet risk, crowding, factor exposure, and correlation with existing portfolio risks. The portfolio manager or mandate owner applies the portfolio rules. | Research input | What the analyst should provide | | --- | --- | | Upside distribution | Scenario values and evidence supporting each state | | Downside | Bear, stress, and thesis-break values plus liquidity path | | Confidence | Evidence quality and unresolved assumptions, not a personality score | | Time horizon | When evidence should validate or falsify the thesis | | Liquidity | Trading liquidity and company financial liquidity as separate concepts | | Correlation drivers | Common commodity, rate, macro, customer, or technology exposures | | Catalyst path | Dated events and what each can resolve | ## 15. Monitoring dashboard The monitoring system should contain only variables that can change the thesis, valuation, or risk state. For each metric, store definition, source, frequency, expected range, warning threshold, thesis-break threshold, and owner. Separate leading indicators from reported outcomes. - Operating KPIs and customer behavior. - Pricing, utilization, capacity, backlog, or retention. - Cash conversion, working capital, capex, and liquidity. - Debt, rates, maturities, covenants, and credit-market signals. - Competitive launches, share, cost or performance benchmarks. - Regulatory and legal milestones. - Management promises and capital-allocation actions. - Valuation and market-implied expectations. ## 16. Risk review cadence Review the full risk register after each material filing, earnings event, acquisition, capital raise, product event, regulatory change, or thesis-level data point. Review critical high-frequency indicators more often only when the information can change a decision. Avoid monitoring every available metric merely because automation makes it possible. ## 17. Worked liquidity stress case Fictional company Nova Components begins with $220 million unrestricted cash and a $150 million revolver. Base CFO is $120 million, annual capex is $90 million, interest is $35 million, and a $250 million term loan matures in 18 months. A demand shock lowers revenue 20%, turns CFO negative $40 million, creates a $70 million inventory build, and widens expected refinancing spread by 400 basis points. 1. Build quarterly cash through the maturity date. 1. Include the inventory build and a realistic minimum cash need. 1. Determine when the revolver is drawn and whether a covenant springs into effect. 1. Test capex cuts and restructuring cash costs. 1. Estimate refinancing capacity under the stressed EBITDA and rate. 1. If refinancing fails, build a claims waterfall and dilution/restructuring scenario. 1. Define the earliest indicator that would make the stress case more likely. The equity thesis is not complete until the analyst knows whether the company survives the adverse operating path without value-destructive financing. ## 18. Risk completion gate - Every material risk has a transmission mechanism, leading indicator, model line, and thesis-break condition. - Liquidity is modeled with timing, restrictions, mandatory uses, and usable financing capacity. - Debt and covenant analysis uses contractual definitions where material. - Correlated risks are modeled coherently rather than added mechanically. - Distress analysis recognizes claim priority and potential value transfer. - The decision journal preserves prior assumptions and disconfirming evidence. - Forecast errors are decomposed and feed back into process changes. - The monitoring dashboard is limited to decision-relevant indicators with explicit thresholds. - A senior reviewer can state what evidence would turn the base case into bear, stress, or thesis break and how quickly that transition would affect cash and value. # MASTER LAB 10 - INVESTMENT MEMOS, INVESTMENT COMMITTEE DEFENSE, CHARTS, AND CONTINUOUS REVIEW > Purpose: convert deep research into a decision-useful communication system that preserves evidence, uncertainty, and auditability. A good research product is concise where the decision requires concision and detailed where the evidence requires detail. The reader should know what matters, why it matters, how much it matters, what is uncertain, and what would change the view. ## 1. The two-layer research product Maintain a short decision layer and a full evidence layer. The short layer should be understandable in minutes and point to the exact assumptions that drive value. The evidence layer should preserve source detail, calculations, model schedules, contrary evidence, and review history. Do not force senior decision makers to read hundreds of pages to understand the thesis, and do not delete the audit trail to achieve brevity. | Decision layer | Evidence layer | | --- | --- | | One-sentence thesis | Source log and dated evidence | | Variant perception | Historical reconstruction and KPI definitions | | Key operating drivers | Full model and scenario schedules | | Valuation range | DCF, reverse DCF, multiples, SOTP detail | | Top risks and thesis breaks | Risk register and stress tests | | Catalysts / validation timeline | Monitoring dashboard and event log | | What changed since prior view | Decision journal and model-change log | ## 2. One-sentence thesis The thesis should state the causal mispricing, not merely that a company is good, cheap, or growing. It should connect an underappreciated operating fact to a financial consequence and explain why the current price does not fully reflect it. Template: "The market appears to discount [embedded expectation], while evidence from [primary drivers] supports [different operating path]; if [specific validation] occurs, normalized [cash flow/earnings/returns] should reach [range], while [defined thesis-break] would invalidate the view." A thesis that cannot be falsified is a narrative, not an underwriting statement. ## 3. Variant perception State what the market appears to expect using reverse valuation, consensus, management guidance, or observable positioning. Then state the evidence-based disagreement. Avoid claiming the market believes one precise narrative unless there is evidence. Often the market price is consistent with several combinations of growth, margins, and returns. - What assumptions are embedded in price? - Which assumption differs from your model? - Why is the difference not already obvious or arbitraged away? - What evidence should emerge if your view is correct? - How long should the market need to recognize the evidence? ## 4. Evidence labeling | Label | Meaning | Communication rule | | --- | --- | --- | | Reported fact | Directly sourced from issuer, regulator, audited or authoritative record | Cite source and period | | Analyst calculation | Reproducible transformation of sourced inputs | Show formula or components | | Management claim | Issuer statement not independently verified | Attribute and test | | External estimate | Third-party data or estimate | Name source, definition, and uncertainty | | Analyst judgment | Interpretation, probability, normalization, or assumption | State rationale and sensitivity | Do not blend categories in one sentence. Readers should be able to distinguish what is known from what is inferred. ## 5. Memo structure 1. Decision summary: thesis, variant view, value range, key risk, validation timeline. 1. Business model: customer, problem, product, economics, capital, bottleneck. 1. Industry: value chain, market size, competition, cycle, technology, regulation. 1. Historical economics: revenue, price-volume-mix, margins, cash conversion, ROIC, capital allocation. 1. Forecast: explicit drivers, constraints, base/bull/bear/stress paths. 1. Valuation: intrinsic, relative, reverse expectations, bridge to per-share value. 1. Risk: transmission mechanisms, liquidity, thesis breaks, mitigants. 1. Management and governance: execution, incentives, capital allocation. 1. Catalysts and monitoring: dated evidence and update rules. 1. Appendix: source log, model checks, detailed schedules, contrary evidence. ## 6. Quantify every debate that matters Convert qualitative debates into model variables. If the debate is "pricing power," show price, volume response, gross margin, retention, competitor behavior, and value sensitivity. If the debate is "AI opportunity," show units, content per system, capacity, margin, reinvestment, competitive supply, and the portion of enterprise value dependent on the opportunity. When a debate cannot be quantified reliably, state the limitation and use a scenario range. False precision is worse than an explicit uncertainty interval. ## 7. Chart design A chart must answer a question faster than prose. Start the title with the conclusion or question, define units and period, source the data, and avoid decoration that obscures comparison. Use the simplest visual that reveals the relationship. | Question | Useful visual | | --- | --- | | Is revenue quality weakening? | Revenue growth versus receivable/contract-asset growth and cash conversion over time | | Is pricing power real? | Price, volume, mix, gross margin, and retention bridge | | Is capacity the constraint? | Demand, effective capacity, utilization, lead time, and planned additions | | What does price imply? | Reverse-DCF required growth/margin path versus history and peer range | | Is capital allocation creating value? | Incremental investment versus incremental NOPAT/FCF and ROIC by period | | Is liquidity adequate? | Quarterly cash runway, committed capacity, maturities, and minimum cash | Avoid dual axes unless the scales are necessary and clearly explained. Avoid truncating axes in ways that exaggerate small changes. Mark definition changes and acquisitions on time-series charts. ## 8. Investment committee preparation Assume the committee will attack the weakest assumption, not the most impressive analysis. Before the meeting, identify the five assumptions with the highest value sensitivity and the weakest evidence. Prepare a source, counterargument, sensitivity, and answer for each. ### IC red-team categories - What is the strongest evidence the thesis is wrong? - What does the market know that the model might miss? - Which assumption is most sensitive to management guidance? - What happens if growth is right but margins are wrong? - What happens if margins are right but capital intensity is worse? - Which cash-flow or balance-sheet item could invalidate adjusted earnings? - What breaks first in the stress case? - Which competitor response is under-modeled? - What evidence arrives before the next formal review? - Why is the expected reward sufficient for the uncertainty and downside? ## 9. Answering uncertainty honestly Do not defend every assumption as if it were certain. A strong IC answer distinguishes high-confidence evidence from judgment, states the range, and explains why the conclusion remains or does not remain robust. If a question reveals a missing issue, mark it open and assign the research required. Improvised certainty is a process failure. ## 10. Earnings preview 1. Write the market and internal expectation for revenue, margins, EPS, FCF, KPIs, and guidance. 1. Identify the two or three variables that matter most for value, not just for the quarter. 1. Define upside, base, and downside interpretations before the print. 1. State what result would require a model change versus only a timing change. 1. Identify footnotes, balance-sheet lines, and cash items to check beyond headline earnings. 1. Prepare management questions that resolve thesis uncertainty without asking for nonpublic information. The preview prevents post-event narrative fitting. After the release, compare the result with prewritten interpretations. ## 11. Post-earnings update sequence 1. Archive the release, filing, presentation, transcript, and data version. 1. Update reported historicals before changing forecasts. 1. Reconcile revenue, margins, working capital, capex, taxes, share count, and cash. 1. Identify definition changes and disclosure drift. 1. Update only assumptions touched by new evidence. 1. Recalculate all scenarios and valuation. 1. Compare the new thesis state with the prior decision journal. 1. Write what changed, what did not change, and what remains unresolved. 1. Update monitoring triggers and next expected evidence. ## 12. Catalyst map A catalyst is an event that can resolve uncertainty or change the distribution of value, not merely a date on the calendar. Examples include capacity qualification, product launch, regulatory decision, debt refinancing, customer renewal cycle, margin inflection, spin-off, asset sale, or evidence that a bottleneck clears. | Catalyst | Date/window | Variable resolved | Positive evidence | Negative evidence | Model line affected | | --- | --- | --- | --- | --- | --- | | Example customer qualification | Q2-Q3 | Capacity utilization / market share | Production award and ramp | Delay or competitor award | Units, utilization, margin | ## 13. Monitoring without noise Set thresholds before the data arrive. A dashboard that turns every small move into an alert trains the analyst to ignore alerts. For each signal, define normal range, warning threshold, thesis-break threshold, and required action. Where a metric is volatile, use rolling averages or cumulative measures that match the business. Automate retrieval and calculation, but preserve human judgment for interpretation. The analyst should receive an alert containing the changed metric, source, historical range, model sensitivity, and relevant thesis statement, not merely "metric changed." ## 14. Research handoff A new analyst should be able to assume coverage without reconstructing the entire research history from email or memory. Maintain an onboarding pack with business model, source index, historical model, assumptions register, top debates, open research items, management chronology, risk register, valuation, and monitoring schedule. - Current thesis and one-sentence variant view. - Top five value drivers and sensitivities. - Top five thesis risks and breaks. - Most important source links and filing sections. - Known model quirks or circularities. - Definition changes and historical recasts. - Upcoming events and open questions. - Decision journal for major prior view changes. ## 15. Post-mortem Perform post-mortems on both successful and unsuccessful outcomes. Separate thesis quality from outcome luck. A correct outcome generated by a bad process should not be celebrated as evidence the process works. A well-reasoned thesis can fail because a low-probability event occurred. | Error type | Example | Process response | | --- | --- | --- | | Source error | Wrong period or stale filing | Strengthen source/version controls | | Model error | Share dilution omitted | Add capitalization-table check | | Forecast error | Overestimated retention | Improve cohort/leading-indicator work | | Thesis error | Moat mechanism was wrong | Revise competitive-advantage test | | Timing error | Catalyst took longer but economics intact | Improve duration/liquidity planning | | Risk error | Liquidity path ignored | Require stress runway before valuation | | Behavioral error | Contrary evidence dismissed | Strengthen precommitment and red-team review | | Luck | Unexpected external shock | Do not rewrite prior reasoning; improve resilience analysis if applicable | ## 16. Writing quality control - Lead with conclusion and numbers, then evidence. - Use specific nouns and verbs rather than vague adjectives. - Separate fact, calculation, claim, estimate, and judgment. - State units and periods. - Avoid unsupported superlatives. - Use ranges where uncertainty is material. - Cite claims at the point of use. - Do not repeat the same point in multiple sections unless repetition serves a different decision purpose. - Remove charts or paragraphs that do not change understanding. - End with what evidence would change the view. ## 17. Worked memo compression exercise Take a 30-page initiation draft and compress it into a two-page decision memo without losing the thesis, variant view, operating drivers, valuation, downside, or validation path. Move evidence tables, footnote detail, and secondary debates to the appendix. Then ask a reviewer to answer six questions after reading only the two pages: What is mispriced? Why? How much is it worth? What can go wrong? What breaks the thesis? What happens next? If any answer is unclear, revise the decision layer. ## 18. Completion gate - The short memo and full evidence pack are consistent. - The thesis is causal, quantified, and falsifiable. - Market-implied expectations are distinguished from analyst assumptions. - Every material claim is labeled by evidence type and sourced. - Charts answer a defined analytical question and preserve units and definitions. - IC preparation targets high-sensitivity weak-evidence assumptions. - Earnings updates compare new evidence with precommitted interpretations. - Catalysts resolve uncertainty and are linked to model variables. - Monitoring thresholds are predefined and decision relevant. - Post-mortems feed specific process improvements back into the research system. # MASTER LAB 11 - SECTOR-SPECIFIC DRIVER, LEADING-INDICATOR, AND FAILURE-MODE ENCYCLOPEDIA > Purpose: supplement the 36 sector playbooks with a compact sector-specific crosswalk. Each entry identifies the modeling unit, the highest-value leading indicators, the accounting or comparability issue most likely to mislead, the valuation frame that best matches the economics, and the sector-specific thesis breaks that should be monitored. ## Software and SaaS ### Modeling core Model recurring revenue through beginning ARR, new ARR, gross churn, contraction, expansion, price, usage, and FX. Convert bookings and RPO to revenue only after modeling contract duration and billing terms. ### Leading indicators Track renewal cohorts, net retention, sales capacity, cloud-consumption trends, customer optimization, deferred revenue, RPO duration, and seat-versus-usage mix. ### Accounting and comparability risk Capitalized commissions, SBC, restructuring, acquisitions, usage revenue timing, and multi-year prepayments can distort FCF and growth comparisons. ### Valuation nuance Use DCF, EV/FCF, and revenue or gross-profit multiples only with explicit mature margin and reinvestment assumptions. Reverse the current multiple into retention, growth duration, and terminal margin. ### Thesis-break focus Thesis breaks often emerge through retention, competitive displacement, rising customer acquisition cost, or a shift from scarce software to commoditized functionality. ## Semiconductors ### Modeling core Model units, wafer starts, die size, yield, ASP, node mix, utilization, packaging/test, and inventory across end markets. Separate fabless economics from foundry and IDM capital intensity. ### Leading indicators Track book-to-bill where meaningful, lead times, distributor inventory, foundry utilization, equipment orders, memory pricing, design wins, product qualification, and customer capex. ### Accounting and comparability risk Channel inventory, customer concentration, purchase commitments, capitalized manufacturing cost, government incentives, and rapid obsolescence can distort cycle signals. ### Valuation nuance Use mid-cycle earnings, DCF, EV/EBIT, and FCF yield with normalized utilization and capex. Peak-cycle low multiples can be value traps. ### Thesis-break focus Key breaks include architecture displacement, node or packaging disadvantage, customer insourcing, excess capacity, and product cycles that fail to convert design wins into volume. ## AI Accelerators and Compute ### Modeling core Model accelerator units, ASP, HBM content, networking, rack density, power per rack, software attach, cloud utilization, and customer concentration. Constrain shipments by packaging, memory, foundry, power, and deployment capacity. ### Leading indicators Track hyperscaler capex, backlog, lead time, CoWoS or advanced-packaging capacity, HBM supply, rack deliveries, power procurement, model efficiency, utilization, and custom silicon adoption. ### Accounting and comparability risk Supply allocation, customer prepayments, concentrated demand, rapid product transitions, warranty, capitalized software, and channel inventory can make short-term revenue nonrepeatable. ### Valuation nuance Reverse DCF should solve for unit growth, ASP/content, mature margin, and duration. Test value under custom silicon, efficiency gains, and lower accelerator intensity per workload. ### Thesis-break focus Thesis breaks include demand concentration reversing, model-efficiency reducing compute intensity faster than workload growth, customer vertical integration, or supply scarcity rents normalizing sooner than expected. ## Cloud and Data Centers ### Modeling core Model commissioned MW, booked MW, utilization, revenue per MW or rack, power cost, PUE, land, interconnection, construction cost, and lease duration. Separate powered shell, wholesale, colocation, and cloud service economics. ### Leading indicators Track utility interconnection queues, transformer/switchgear lead times, land and power contracts, preleasing, customer capex, construction starts, financing spreads, and regional power prices. ### Accounting and comparability risk Backlog can include long-dated options; construction in progress, capitalized interest, leases, and development JV accounting can flatter near-term operating metrics. ### Valuation nuance Use project-level DCF/NAV plus corporate valuation. Cap rates and EV/EBITDA need normalization for development pipeline and capital intensity. ### Thesis-break focus Thesis breaks include power unavailable on schedule, customer concentration, overbuilding in a region, financing cost exceeding project returns, or technology reducing space/power demand per workload. ## Industrial Machinery ### Modeling core Model units, installed base, price, mix, aftermarket, utilization, backlog conversion, dealer inventory, labor hours, and material cost. Distinguish original equipment from higher-margin service. ### Leading indicators Track PMI/capex indicators, dealer inventories, lead times, book-to-bill, rental utilization, construction/manufacturing activity, freight, commodity inputs, and order cancellations. ### Accounting and comparability risk Percent-completion, backlog quality, dealer financing, restructuring, pension, and working-capital swings can obscure normalized earnings. ### Valuation nuance Use mid-cycle EBIT/FCF, DCF, and SOTP where aftermarket differs materially. Normalize margin for utilization and price-cost cycle. ### Thesis-break focus Breaks include dealer destocking, capacity overbuild, aftermarket disruption, aggressive price-cost assumptions, or structurally lower end-market capital intensity. ## Aerospace and Defense ### Modeling core Model program units, shipsets, content per platform, backlog, production rates, aftermarket flight hours, contract type, cost curves, and government budgets. ### Leading indicators Track OEM build rates, supplier deliveries, engine removals, flight hours, defense appropriations, contract awards, program milestones, and quality/regulatory actions. ### Accounting and comparability risk Program accounting, loss reserves, customer advances, pension, cost-to-complete estimates, and supplier concessions require deep footnote work. ### Valuation nuance Use DCF, EV/EBIT, FCF yield, and SOTP. Value long-cycle defense backlog differently from commercial aftermarket and new-platform ramps. ### Thesis-break focus Breaks include certification/quality failure, supplier bottlenecks, program cancellation, fixed-price cost overrun, or production-rate assumptions that exceed supply-chain capability. ## Airlines ### Modeling core Model ASM, RPM, load factor, yield, ancillary revenue, fuel, labor, fleet, utilization, maintenance, and capacity by region. ### Leading indicators Track booking curves, fare data, TSA/passenger data, corporate travel, competitor capacity, fuel, aircraft deliveries, maintenance events, and credit-card remuneration. ### Accounting and comparability risk Sale-leasebacks, loyalty-program economics, maintenance capitalization, pension, and deferred ticket revenue can complicate cash and leverage. ### Valuation nuance Use normalized EV/EBITDAR or EV/EBIT, FCF through cycle, and asset/liquidity analysis. Peak travel margins should not be capitalized perpetually. ### Thesis-break focus Breaks include capacity oversupply, labor inflation, aircraft constraints, fuel spikes without pricing, balance-sheet stress, or loyalty economics weakening. ## Automotive OEMs ### Modeling core Model wholesale units, retail sell-through, ASP, incentives, mix, dealer inventory, warranty, financial-services contribution, plant utilization, and capex. ### Leading indicators Track registrations, dealer days supply, incentives, used-car values, order banks, production schedules, battery/material costs, fleet regulation, and finance delinquencies. ### Accounting and comparability risk Captive finance, pension, warranty reserves, lease residuals, incentives, inventory financing, and restructuring can move earnings materially. ### Valuation nuance Use mid-cycle earnings, DCF, SOTP for captive finance, and asset/liquidity stress. Treat peak pricing and low incentives as cyclical until proven structural. ### Thesis-break focus Breaks include price war, residual-value collapse, platform transition failure, excess capacity, warranty/recall burden, or financing losses. ## EV and Battery Manufacturers ### Modeling core Model vehicle or cell units, kWh, ASP, chemistry mix, yield, utilization, material cost per kWh, warranty, credits, and capex by plant. ### Leading indicators Track registrations, order lead times, cell pricing, lithium/nickel costs, plant ramp, yield, incentives, charging infrastructure, and competitor price changes. ### Accounting and comparability risk Government incentives, capitalized development, supplier prepayments, warranty, inventory, and rapid technology obsolescence can distort profitability. ### Valuation nuance Use long-horizon DCF with explicit financing/dilution, unit economics, and plant-level returns. Revenue multiples require a credible mature margin and capital-intensity bridge. ### Thesis-break focus Breaks include demand elasticity, slower yield ramp, chemistry displacement, subsidy change, capital shortfall, or price declines outrunning cost reductions. ## BESS Electrical Balance-of-System ### Modeling core Model MW/MWh deployed, inverter or PCS content, switchgear/transformer content, engineering/service attach, project timing, price per kW, warranty, and backlog conversion. ### Leading indicators Track interconnection queues, battery project awards, transformer and switchgear lead times, PCS orders, grid-code changes, utility procurement, fire/safety standards, and storage economics. ### Accounting and comparability risk Project milestones, customer advances, warranties, long lead procurement, supplier concentration, and percentage-of-completion can distort revenue and cash timing. ### Valuation nuance Use DCF, EV/EBIT, and backlog-adjusted scenarios with explicit normalized margins after supply scarcity. Separate equipment from recurring service/software value. ### Thesis-break focus Breaks include vertical integration by battery suppliers, standardization commoditizing PCS, transformer bottleneck shifting value elsewhere, project delays, or warranty failures. ## Grid Equipment and Electrification ### Modeling core Model units or capacity for transformers, switchgear, breakers, conductors, protection, power electronics, and service. Tie demand to utility capex, load growth, interconnection, replacement, and data-center/industrial projects. ### Leading indicators Track utility rate-base plans, interconnection queues, lead times, factory expansions, copper/electrical steel prices, order backlog, cancellations, and manufacturing capacity. ### Accounting and comparability risk Long-cycle backlog, price-escalation clauses, advance payments, pension, project accounting, and capacity-expansion capex require normalization. ### Valuation nuance Use DCF, EV/EBIT, and FCF with margin normalization as scarcity eases. Evaluate replacement-cost and capacity economics for new entrants. ### Thesis-break focus Breaks include capacity response collapsing price, regulatory delays, utility affordability constraints, project cancellations, or a change in grid architecture reducing equipment content. ## Electric Utilities ### Modeling core Model rate base, allowed ROE, load growth, generation mix, fuel, purchased power, capex, depreciation, financing, and customer rates by jurisdiction. ### Leading indicators Track rate cases, regulatory orders, load forecasts, weather, data-center connections, fuel spreads, capex plans, credit metrics, and political affordability pressure. ### Accounting and comparability risk Regulatory assets/liabilities, securitization, storm costs, pension, decommissioning, and capitalized AFUDC can make GAAP earnings differ from cash. ### Valuation nuance Use regulated-asset and dividend/FCFE frameworks, P/E relative to growth and allowed returns, and credit-sensitive DCF. Capital needs and dilution are central. ### Thesis-break focus Breaks include adverse regulatory outcomes, cost overruns, wildfire/nuclear liabilities, financing pressure, load forecasts that fail to materialize, or customer affordability backlash. ## Renewable Developers ### Modeling core Model project MW, capacity factor, PPA price, merchant exposure, construction cost, tax credits, financing, curtailment, degradation, and operating expense. ### Leading indicators Track interconnection, queue progress, equipment prices, interest rates, tax-credit transfer pricing, PPAs, permitting, transmission, and project-sale markets. ### Accounting and comparability risk Tax equity, development gains, project sales, nonrecourse debt, capitalized interest, and unconsolidated JVs complicate reported earnings. ### Valuation nuance Use project NAV/DCF and corporate SOTP. Stress discount rates, merchant tails, curtailment, construction cost, and financing availability. ### Thesis-break focus Breaks include interconnection failure, cost inflation, financing spread, PPA repricing, policy change, or lower capacity factor/greater curtailment. ## Oil and Gas E&P ### Modeling core Model production by commodity, decline curves, realized price, differentials, hedges, lifting cost, royalties, drilling/completion cost, inventory depth, and maintenance capital. ### Leading indicators Track rigs, frac spreads, permits, basin takeaway, storage, commodity curves, service costs, well productivity, decline, and operator capex. ### Accounting and comparability risk Reserve revisions, successful-efforts/full-cost accounting, impairments, derivative marks, asset retirement obligations, and acquisition adjustments matter. ### Valuation nuance Use NAV by acreage/project, mid-cycle FCF, EV/EBITDA only with maintenance-capex context, and commodity sensitivities. ### Thesis-break focus Breaks include lower resource productivity, cost inflation, basis widening, regulatory limits, weak balance sheet, or inventory exhaustion. ## Midstream Energy ### Modeling core Model volumes, contracted capacity, tariff, commodity sensitivity, minimum commitments, expansion capex, maintenance capex, and counterparty credit. ### Leading indicators Track producer activity, basin differentials, contract renewals, pipeline utilization, project approvals, regulatory rulings, and customer leverage. ### Accounting and comparability risk MLP or partnership structures, noncontrolling interests, equity-method JVs, maintenance-capex definitions, and distributable cash flow adjustments require reconciliation. ### Valuation nuance Use DCF/distributable cash flow, EV/EBITDA with contract quality, and project returns. Include debt and distribution coverage. ### Thesis-break focus Breaks include contract roll-off at lower rates, producer distress, regulatory blockage, overbuild, or maintenance needs higher than reported. ## Refiners ### Modeling core Model throughput, utilization, crude differentials, crack spreads, product yields, RIN/renewable obligations, turnaround, and working capital. ### Leading indicators Track crack spreads, inventories, utilization, outages, crude differentials, product demand, refinery closures/additions, and regulatory credit prices. ### Accounting and comparability risk Inventory accounting, turnaround capitalization, environmental obligations, and working-capital swings create large earnings/cash timing differences. ### Valuation nuance Use mid-cycle earnings/FCF and asset replacement economics. Peak cracks should not be capitalized as permanent. ### Thesis-break focus Breaks include structural demand decline, new low-cost capacity, feedstock disadvantage, environmental capex, or prolonged utilization weakness. ## Chemicals ### Modeling core Model volume, price, mix, feedstock, energy, utilization, capacity additions, maintenance outages, and product-specific spreads. ### Leading indicators Track operating rates, inventories, feedstock spreads, China/global capacity, freight, end-market production, and plant outages. ### Accounting and comparability risk Inventory, pension, environmental liabilities, JV accounting, restructuring, and maintenance turnaround timing require normalization. ### Valuation nuance Use mid-cycle EV/EBITDA, DCF, FCF yield, and replacement-cost context. Normalize for cycle and feedstock advantage. ### Thesis-break focus Breaks include sustained global overcapacity, feedstock disadvantage, regulation, substitution, or structurally lower demand in key end markets. ## Mining and Metals ### Modeling core Model production, grade, recovery, realized price, treatment charges, cash cost, sustaining capex, growth capex, reserves, and mine life. ### Leading indicators Track benchmark prices, inventories, treatment charges, production guidance, grade, permitting, labor, power, freight, and new project supply. ### Accounting and comparability risk Stripping, reserve estimates, rehabilitation liabilities, JVs, royalties, impairment, and exploration capitalization can distort comparisons. ### Valuation nuance Use project NAV, commodity sensitivities, mid-cycle FCF, and asset quality. Corporate overhead and future development capital matter. ### Thesis-break focus Breaks include reserve/grade disappointment, project overrun, jurisdiction change, lower commodity incentive price, or capital needs exceeding balance-sheet capacity. ## Banks ### Modeling core Model average loans/securities, yields, deposits, deposit beta, funding mix, NIM, fees, expenses, charge-offs, provisions, capital, and buybacks. ### Leading indicators Track deposit flows and pricing, loan growth, credit delinquencies, charge-offs, securities marks, yield curve, funding markets, capital ratios, and regulatory actions. ### Accounting and comparability risk CECL/reserve assumptions, AOCI, held-to-maturity marks, nonaccruals, loan modifications, and capital treatment are central. ### Valuation nuance Use P/TBV, P/E, residual income, and excess-capital approaches tied to normalized ROE and cost of equity. ### Thesis-break focus Breaks include deposit franchise weakening, credit losses above reserve, capital shortfall, funding stress, regulatory constraint, or asset-liability mismatch. ## Property and Casualty Insurance ### Modeling core Model premiums, exposure units, rate, retention, loss ratio, expense ratio, catastrophe load, reserve development, reinsurance, float, and investment income. ### Leading indicators Track rate filings, claims inflation, catastrophe events, renewal retention, reinsurance pricing, reserve development, and bond yields. ### Accounting and comparability risk Loss reserves, reinsurance recoverables, catastrophe accounting, prior-year development, and investment marks dominate quality. ### Valuation nuance Use P/B or P/TBV relative to normalized ROE, underwriting-cycle earnings, and excess capital. ### Thesis-break focus Breaks include adverse reserve development, social inflation outrunning pricing, reinsurance unavailable, catastrophe concentration, or capital erosion. ## Life Insurance ### Modeling core Model policies or account value, premiums, spreads, lapses, mortality/morbidity, hedging, statutory capital, and investment yield. ### Leading indicators Track lapse rates, credit spreads, mortality, annuity sales, new-money yields, hedge effectiveness, regulatory capital, and reinsurance. ### Accounting and comparability risk Actuarial assumptions, DAC, market-risk benefits, reinsurance, statutory versus GAAP capital, and investment impairments require specialist treatment. ### Valuation nuance Use book-value/ROE, distributable capital, embedded-value concepts where appropriate, and earnings normalized for assumption updates. ### Thesis-break focus Breaks include lapse shock, asset-liability mismatch, reserve strengthening, hedge failure, credit losses, or capital/regulatory constraint. ## Asset Managers and Brokers ### Modeling core Model AUM by asset class, market beta, net flows, fee rate, performance fees, advisor/broker activity, compensation, and capital return. ### Leading indicators Track fund performance, net flows, market levels, fee compression, advisor recruiting, trading/underwriting volumes, and client cash balances. ### Accounting and comparability risk Seed investments, consolidation, fair-value marks, principal investments, performance-fee timing, and compensation accruals affect comparability. ### Valuation nuance Use P/E, FCF, SOTP, and percent-of-AUM context. Separate market appreciation from organic flows and fee-rate mix. ### Thesis-break focus Breaks include persistent outflows, fee compression, poor performance, advisor attrition, regulatory changes, or capital trapped in low-return businesses. ## Payments and Fintech ### Modeling core Model payment volume, transactions, active users/merchants, take rate, value-added services, credit losses if applicable, incentives, and processing cost. ### Leading indicators Track consumer spend, cross-border travel, merchant adds, payment volume, interchange/regulation, fraud losses, funding costs, and network tokenization. ### Accounting and comparability risk Gross-versus-net revenue, customer incentives, pass-through network fees, credit receivables, securitization, and SBC can distort growth/margins. ### Valuation nuance Use DCF, EV/FCF, and growth/margin frameworks with mature take rate and capital needs. Separate networks, processors, lenders, and software-like models. ### Thesis-break focus Breaks include take-rate compression, regulation, merchant/customer concentration, fraud/credit losses, platform disintermediation, or customer acquisition economics deteriorating. ## REITs ### Modeling core Model same-store NOI, occupancy, rent spread, lease maturity, tenant credit, development, acquisitions/dispositions, recurring capex, debt, and shares. ### Leading indicators Track leasing activity, market rents, vacancy, cap rates, transaction volumes, financing spreads, construction supply, and tenant health. ### Accounting and comparability risk FFO/AFFO definitions, straight-line rent, tenant improvements, leasing commissions, unconsolidated JVs, and development capitalization need reconciliation. ### Valuation nuance Use NAV, implied cap rate, AFFO/FCF, and property-level DCF with leverage. ### Thesis-break focus Breaks include tenant distress, refinancing at uneconomic rates, oversupply, development overruns, cap-rate expansion, or recurring capex above AFFO assumptions. ## Homebuilders ### Modeling core Model orders, closings, backlog, ASP, incentives, gross margin, lots, land spend, cycle time, cancellations, and mortgage-rate affordability. ### Leading indicators Track new-home sales, permits, starts, mortgage rates, resale inventory, incentives, community count, order pace, land prices, and cancellations. ### Accounting and comparability risk Land impairments, optioned versus owned lots, mortgage operations, incentives, and capitalized interest affect cycle comparisons. ### Valuation nuance Use normalized P/E, P/B, FCF, land value, and through-cycle ROE. ### Thesis-break focus Breaks include affordability shock, land overcommitment, cancellations, resale inventory normalization, or margin compression from incentives. ## Restaurants ### Modeling core Model units, same-store sales from traffic and ticket, restaurant margin, labor, food, occupancy, new-store ramp, closures, franchise mix, and capex. ### Leading indicators Track traffic, menu pricing, promotions, commodity inputs, wage inflation, unit openings, franchisee health, digital mix, and consumer spending. ### Accounting and comparability risk Franchise versus company-store mix, gift cards, leases, closure costs, and preopening expense change comparability. ### Valuation nuance Use unit-level DCF, EV/EBITDA/FCF, and growth-adjusted frameworks tied to new-unit returns and mature store economics. ### Thesis-break focus Breaks include traffic elasticity to price, unit cannibalization, labor inflation, weak franchisee economics, or declining new-store cash-on-cash returns. ## Retail ### Modeling core Model stores, square footage, traffic, conversion, units per transaction, ASP, e-commerce, gross margin, shrink, inventory turns, rent, labor, and capex. ### Leading indicators Track foot traffic, card spend, promotions, inventory, markdowns, freight, vendor terms, store openings/closures, and consumer credit. ### Accounting and comparability risk Lease obligations, vendor allowances, gift cards, inventory reserves, private-label credit, and supplier finance can distort cash/margins. ### Valuation nuance Use mid-cycle FCF, EV/EBIT, and SOTP for credit/e-commerce where relevant. Inventory and lease intensity matter. ### Thesis-break focus Breaks include sustained traffic loss, markdown cycle, inventory obsolescence, lease burden, e-commerce economics, or vendor tightening. ## Consumer Packaged Goods ### Modeling core Model volume, price, mix, distribution, market share, commodities, advertising, trade spend, and working capital by category/geography. ### Leading indicators Track scanner data, retailer inventory, price gaps, promotion, commodity costs, share, distribution points, consumer confidence, and private-label penetration. ### Accounting and comparability risk Trade promotions, pension, restructuring, brand intangibles, FX, and acquisition accounting can flatter organic comparisons. ### Valuation nuance Use DCF, P/E, EV/EBIT, and FCF yield with durable brand/share and reinvestment assumptions. ### Thesis-break focus Breaks include volume elasticity, private-label share gain, retailer bargaining power, brand underinvestment, or price increases masking unit decline. ## Pharmaceuticals ### Modeling core Model prescriptions/patients, price, gross-to-net, indication, geography, patent/exclusivity, R&D, milestones, royalties, and launch curves. ### Leading indicators Track prescriptions, formulary access, trial readouts, competitor data, payer coverage, patent litigation, manufacturing, and regulatory milestones. ### Accounting and comparability risk Gross-to-net reserves, acquired IPR&D, collaboration revenue, contingent consideration, milestone accounting, and patent lives are critical. ### Valuation nuance Use product-level DCF/SOTP with probability, patent cliffs, and replacement pipeline. Mature P/E alone can hide concentration. ### Thesis-break focus Breaks include clinical failure, reimbursement restriction, safety signal, patent loss, competitor superiority, or pipeline unable to replace cliffs. ## Biotechnology ### Modeling core Model patient population, diagnosis, eligible share, penetration, price, persistence, trial probability, launch timing, cash burn, and financing. ### Leading indicators Track enrollment, trial timelines, biomarker data, regulatory interactions, competitor readouts, manufacturing readiness, KOL adoption, and cash runway. ### Accounting and comparability risk R&D expense, acquired IPR&D, milestone payments, collaboration accounting, warrant/convertible dilution, and going-concern risk dominate. ### Valuation nuance Use probability-adjusted product DCF/SOTP and explicit financing dilution. Avoid revenue multiples without clinical and capital path. ### Thesis-break focus Breaks include trial failure, safety, regulatory delay, financing shortfall, competing therapy superiority, or commercial adoption below the required patient share. ## Medical Devices ### Modeling core Model procedures, installed base, units per procedure, ASP, disposable pull-through, service, surgeon/hospital adoption, and gross margin. ### Leading indicators Track procedure volumes, hospital capex, utilization, trial/approval milestones, reimbursement, sales-force productivity, and competitor launches. ### Accounting and comparability risk Consignment inventory, warranty, capitalized development, acquisition amortization, and distributor inventory can affect comparisons. ### Valuation nuance Use DCF, EV/EBIT, and growth/ROIC frameworks. Separate installed-base capital from recurring consumables/service economics. ### Thesis-break focus Breaks include safety/recall, reimbursement, slower utilization, surgeon switching, hospital budget pressure, or installed base failing to generate expected pull-through. ## Managed Care ### Modeling core Model members by product, premiums, medical cost trend, MLR, risk adjustment, pharmacy, administrative cost, stars/quality, and capital. ### Leading indicators Track enrollment, utilization, provider rates, drug trend, government rate notices, risk adjustment, star ratings, and policy changes. ### Accounting and comparability risk Medical claims reserves, risk adjustment, pharmacy rebates, acquisitions, statutory capital, and government program timing are key. ### Valuation nuance Use P/E/FCF and DCF with normalized margin by product and capital needs. Growth without adequate pricing can destroy value. ### Thesis-break focus Breaks include medical-cost trend outrunning pricing, reimbursement cuts, quality-rating decline, risk-adjustment change, or regulatory constraints. ## Telecom ### Modeling core Model subscribers, gross adds, churn, ARPU, device economics, network usage, spectrum, capex, tower/fiber lease, and promotions. ### Leading indicators Track porting, subscriber adds, pricing/promotions, handset cycles, spectrum auctions, capex, network quality, fixed-wireless/fiber adds, and competition. ### Accounting and comparability risk Device financing, leases, spectrum capitalization, pension, tower transactions, and customer acquisition costs complicate FCF. ### Valuation nuance Use DCF, EV/EBITDA with capex/leasing normalization, and sum-of-parts where infrastructure assets differ. ### Thesis-break focus Breaks include price war, churn spike, network disadvantage, capex intensity, spectrum cost, or leverage limiting investment. ## Internet Platforms and Marketplaces ### Modeling core Model users/buyers/sellers, engagement, transactions or GMV, take rate, ad load/pricing, fulfillment, payment, and content costs. ### Leading indicators Track traffic/app engagement, conversion, merchant/seller adds, ad pricing, e-commerce spend, regulatory changes, and competitive acquisition costs. ### Accounting and comparability risk Gross-versus-net revenue, traffic acquisition costs, SBC, content capitalization, payments/credit, and international FX can distort margins. ### Valuation nuance Use DCF, EV/FCF, and segment SOTP. Reverse price into user growth, monetization, take rate, and mature margin. ### Thesis-break focus Breaks include platform disintermediation, regulation, user engagement decline, take-rate pressure, rising acquisition cost, or network effects weakening through multi-homing. ## Cybersecurity ### Modeling core Model ARR, new ARR, retention, platform/module adoption, billings, RPO, seats/endpoints/usage, gross margin, sales productivity, and SBC. ### Leading indicators Track breach environment, IT budgets, platform consolidation, sales hiring, renewal commentary, cloud workloads, channel activity, and competitor displacement. ### Accounting and comparability risk Contract timing, deferred revenue, capitalized commissions, SBC, M&A add-backs, and usage pricing changes can mask economics. ### Valuation nuance Use DCF, EV/FCF, and revenue/gross-profit multiples with mature margin and retention. Reverse DCF should test platform consolidation and duration. ### Thesis-break focus Breaks include retention decline, sales efficiency deterioration, platform commoditization, cloud-provider bundling, product failure/breach, or excessive dilution. ## Railroads and Logistics ### Modeling core Model volume by commodity/customer, revenue per unit, fuel surcharge, operating ratio, velocity, dwell, labor, equipment utilization, capex, and network capacity. ### Leading indicators Track carloads, intermodal volumes, industrial production, port activity, trucking rates, fuel, service metrics, labor agreements, and capex. ### Accounting and comparability risk Fuel surcharge timing, pension, casualty/environmental reserves, asset lives, and network-capex classification can distort margins/cash. ### Valuation nuance Use DCF, EV/EBIT, FCF yield, and replacement/network value context. Normalize service and volume through cycle. ### Thesis-break focus Breaks include service deterioration, labor inflation, modal share loss, regulatory constraints, network bottlenecks, or capex required to sustain service above modeled levels. ## Sector encyclopedia completion gate - The analyst uses the sector-specific modeling unit rather than a generic revenue-growth template. - Leading indicators are backtested against reported outcomes and definition-controlled. - The model includes the sector-specific accounting items most likely to distort comparisons. - Valuation reflects capital intensity, cycle, regulation, and the correct claimant level. - Thesis-break conditions are written in the language of the sector, not generic growth or margin thresholds. # MASTER LAB 12 - ZERO-TO-EXPERT TRAINING, CERTIFICATION, DELIBERATE PRACTICE, AND ANSWER RUBRICS > Purpose: provide a training path for a motivated beginner who does not yet possess professional accounting, modeling, valuation, or industry-research experience. Reading is not sufficient. Advancement requires reproducible work products, timed drills, model builds, written defenses, error reviews, and oral examination. The learner advances by demonstrated competency, not calendar time. ## 1. Competency levels | Level | Description | Minimum observable capability | | --- | --- | --- | | 0 - Foundation | Understands financial vocabulary and arithmetic | Can read the three statements, calculate growth/margins, and explain basic debit/credit mechanics | | 1 - Source analyst | Can retrieve and reconstruct facts | Can find filings, footnotes, exhibits, ownership forms, and reconcile reported historicals | | 2 - Accounting analyst | Can explain statement mechanics | Can build roll-forwards, normalize non-GAAP, and trace transactions through all three statements | | 3 - Model analyst | Can build integrated forecasts | Can build a three-statement model from raw sources with driver schedules and checks | | 4 - Valuation analyst | Can connect economics to value | Can build DCF, reverse DCF, relative valuation, SOTP, and special-case methods coherently | | 5 - Industry analyst | Can underwrite market structure | Can size markets, map bottlenecks, benchmark competitors, and identify leading indicators | | 6 - Research analyst | Can form and defend a thesis | Can write a falsifiable variant view, risk register, valuation range, and monitoring plan | | 7 - Senior analyst | Can challenge and improve other work | Can detect hidden assumptions, run IC review, diagnose errors, and update views without hindsight bias | ## 2. Phase 0: arithmetic, statements, and accounting language A beginner should not begin with DCF. First build fluency in percentages, growth rates, margins, weighted averages, present value, and statement vocabulary. The learner should be able to explain the difference between revenue, bookings, billings, cash collection, profit, EBITDA, operating cash flow, and free cash flow without relying on memorized slogans. ### Required drills - Calculate year-over-year and sequential growth from raw quarterly data. - Calculate gross, operating, pretax, and net margins. - Reconcile beginning and ending retained earnings. - Reconstruct change in cash from CFO, CFI, CFF, and FX. - Create ten journal entries and show their effect on all three statements. - Explain why depreciation lowers earnings but is added back in the indirect cash flow statement. - Explain why inventory growth can reduce cash even when it is not an income-statement expense in the same period. ### Pass standard Complete a closed-book 60-minute exercise with no unreconciled statement error and at least 90% correct arithmetic. Any error involving signs, units, or the accounting equation requires remediation before advancement. ## 3. Phase 1: source retrieval and filing navigation The learner must become faster at primary-source retrieval than at searching commentary. Practice locating the latest 10-K, 10-Q, 8-K, proxy, exhibits, debt documents, ownership forms, and amended filings. For each source, record date, period, accession/document ID, and the exact section used. ### Drill set 1. Find the issuer fiscal year-end and all reportable segments. 1. Find the revenue-recognition policy and identify the key judgment. 1. Find debt maturities, rates, and covenant references. 1. Find share-based compensation expense and outstanding awards. 1. Find acquisition purchase price and goodwill created. 1. Find related-party disclosures and customer concentration. 1. Find auditor critical audit matters and material weaknesses if any. 1. Compare the current risk-factor section with the prior year and identify three substantive changes. ### Pass standard A reviewer selects any five material facts from the learner source pack. The learner must open the exact supporting source within two minutes per fact and explain the definition and period without relying on a secondary summary. ## 4. Phase 2: historical reconstruction Build five years of income statement, balance sheet, cash flow, segment, share count, and core KPI history. The learner should not forecast yet. The objective is to learn where data live and how definitions change. - Tie annual statements to audited filings. - Reconcile quarterly periods to annual totals. - Preserve reported and normalized series separately. - Create segment reconciliation including eliminations. - Create acquisition and discontinued-operation timeline. - Calculate working-capital days, ROIC, FCF, dilution, and leverage from sourced components. - Write a one-page historical economics summary using only evidence from the reconstruction. ### Pass standard Historical statements must balance and source checks must be complete. A senior reviewer should be able to reproduce each derived KPI from raw components. No normalization may replace the reported number without a reversible bridge. ## 5. Phase 3: accounting mastery Use Master Lab 01. The learner must rebuild at least six material footnote roll-forwards and write transaction-level explanations. The required set should include revenue/contract balances, PP&E, debt, stock compensation, taxes, and one company-specific area such as pensions, warranties, reserves, or acquisitions. ### Accounting oral exam - If receivables rise $20 with no other change, what happens to CFO? - Why can a deferred-revenue increase boost CFO before revenue is recognized? - How can useful-life extension improve earnings without improving cash? - Why is SBC both a noncash expense and an economic dilution issue? - How can supplier finance flatter CFO? - Why can releasing a tax valuation allowance raise GAAP earnings without equivalent current cash? - When can acquired-intangible amortization be economically relevant even though it is noncash? The learner passes only if answers explain mechanics and economic interpretation, not memorized accounting labels. ## 6. Phase 4: first integrated model Use Master Lab 02 to build a five-year forecast from raw history. The training company should include at least two revenue drivers, working capital, capex/depreciation, debt/interest, taxes, SBC, and share repurchases or issuance. The model must have base, bull, bear, stress, and thesis-break states. ### Model-break test 1. Reviewer changes a volume driver by 10%. Learner explains every statement line affected. 1. Reviewer delays capex by one year. Model updates depreciation, capacity, cash, and debt. 1. Reviewer increases DSO by 15 days. Model updates cash and financing. 1. Reviewer increases interest rates. Debt and liquidity update. 1. Reviewer changes share price used for buybacks/options. Per-share value and dilution respond. 1. Reviewer disables one scenario switch and verifies historical periods remain unchanged. Any hard-coded forecast result that prevents the causal update is a failure. ## 7. Phase 5: valuation mastery Build DCF, reverse DCF, multiples, and SOTP or another sector-appropriate valuation. The learner must explain why each method is appropriate and reconcile differences. The DCF must include terminal reinvestment/ROIC consistency, dilution, and enterprise-to-equity bridge. ### Valuation oral exam - Why can two companies with the same EBITDA deserve different EV/EBITDA multiples? - What happens to terminal value if long-run growth rises but the required reinvestment also rises? - Why can a low P/E be expensive for a cyclical company? - When should debt not be subtracted in an industrial-style EV bridge because the business is a financial institution? - How does reverse DCF change the research question? - Why does a higher DCF value sometimes increase option dilution? ### Pass standard The learner must reproduce the valuation from operating assumptions, explain every bridge item, and state the market-implied operating path. A point target without scenario range and reverse expectations does not pass. ## 8. Phase 6: industry research Select one sector and build the value chain without beginning from the target stock. Include customers, suppliers, substitutes, regulators, standards, capital providers, physical bottlenecks, and profit pools. Build bottom-up market size and supply capacity. Identify at least three leading indicators and backtest them against reported outcomes. ### Pass standard The learner must defend why the identified profit pool should persist, what capital response could erode it, and which technology or regulatory change could move value to another layer. A TAM slide copied from an investor presentation is an automatic failure. ## 9. Phase 7: management, governance, and capital allocation Create a five-year management chronology, compensation map, acquisition ledger, and capital-return history. Recalculate at least one management incentive metric using an economic definition. Evaluate buybacks based on price and net dilution rather than total dollars. ### Pass standard The learner must distinguish process from outcome and identify one management decision that appeared rational ex ante but had a poor outcome, or vice versa. This tests resistance to hindsight bias. ## 10. Phase 8: forensic accounting Use Master Lab 04 on a company with at least three forensic issues. Build a benign and adverse interpretation for each, then identify discriminating evidence. Do not select a company solely because it is controversial. The objective is disciplined uncertainty, not accusation. ### Pass standard The learner may not use a fraud-screen score as the conclusion. The final memo must show source evidence, cash and balance-sheet reconciliation, model impact, and the evidence required to resolve each open issue. ## 11. Phase 9: full initiation Produce the complete deliverable pack: source log, history, industry map, management chronology, model, valuation, reverse expectations, risk register, memo, monitoring dashboard, and decision journal. The analyst should be able to hand the package to another reviewer who has not followed the project and still have the work reproduced. ## 12. Timed earnings simulation Use a historical earnings release that the learner has not studied. Freeze all information after the release date. Give the learner the prior model, then provide the release, filing, presentation, and transcript on a timed schedule. Require a factual update, model revision, thesis update, and two-page note within a realistic research-team window. ### Scoring | Dimension | Weight | Failure condition | | --- | --- | --- | | Source accuracy | 20% | Material value or period unsupported | | Model mechanics | 20% | Statements fail or forecast overwritten incorrectly | | Interpretation | 20% | Timing issue mistaken for structural change or vice versa | | Valuation / risk | 15% | Material bridge or liquidity effect omitted | | Communication | 15% | Decision maker cannot identify what changed and why | | Process / audit trail | 10% | Changes cannot be reproduced | ## 13. Research writing drills - Write a 100-word business model explanation without jargon. - Write a one-sentence falsifiable thesis. - Write a two-page initiation summary. - Write the strongest short thesis against your own long thesis, or vice versa. - Write a post-earnings update that separates fact changes from assumption changes. - Turn a 10-page industry section into one chart and five decision-relevant bullets. - Rewrite a paragraph so every sentence is labeled fact, calculation, claim, estimate, or judgment. ## 14. Error journal Every learner maintains an error journal. For each error, record date, task, error type, why the error passed existing controls, financial impact, corrective action, and regression test. Repeated error classes are more important than isolated mistakes. | Error class | Example remediation | | --- | --- | | Source/version | Add amendment and period check before import | | Units/sign | Add automated scale/sign validation | | Accounting mechanics | Rebuild transaction journal before model entry | | Forecast assumption | Add leading indicator or cohort analysis | | Valuation | Add terminal consistency and bridge checklist | | Behavioral | Precommit contrary evidence and red-team review | | Communication | Require decision-summary template and source labels | ## 15. Senior-review rubric | Dimension | 99-level behavior | | --- | --- | | Evidence | Primary-source claims reproducible; definitions and versions controlled | | Accounting | Can reconstruct material transactions, estimates, and statement effects | | Modeling | Driver-based integrated model with stable scenarios and visible checks | | Valuation | Multiple appropriate frameworks with reverse expectations and consistent reinvestment | | Industry | Bottom-up market, supply, profit pool, competition, cycle, and technology analysis | | Management | Evidence-based execution, incentive, governance, and capital-allocation assessment | | Forensics | Contradictions investigated with benign/adverse hypotheses and quantified impact | | Risk | Liquidity, thesis-break, and transmission mechanisms explicit | | Communication | Concise decision layer backed by full audit trail | | Learning | Post-mortems change the process; prior views are preserved rather than rewritten | ## 16. Graduation examination The final examination should use a public company not previously covered by the learner. The learner receives only the ticker/name and a deadline. Over several days, the learner must produce a complete initiation and defend it orally. The reviewer deliberately introduces counterevidence and changes one source document late in the process to test version control and updating discipline. 1. Day 1: source pack, business model, historical reconstruction, open-question map. 1. Day 2: industry/value chain, competitor benchmark, management chronology, forensic scan. 1. Day 3: integrated model and scenarios. 1. Day 4: valuation, reverse DCF, risk/liquidity, monitoring. 1. Day 5: two-page decision memo and full oral defense. Passing requires no critical source or model error, a coherent causal thesis, transparent uncertainty, and the ability to change the view when contrary evidence is introduced. A polished presentation cannot compensate for unsupported facts or broken mechanics. ## 17. Continuing mastery after graduation Expertise decays if the analyst stops testing it. Maintain quarterly accounting drills, annual model rebuilds from scratch, sector rotations, post-mortems, and periodic blind initiations. Track the personal error taxonomy. The goal is not to eliminate all error, which is impossible, but to reduce avoidable error and detect it earlier. - One filing reconstruction drill per quarter. - One model built from a blank workbook at least annually. - One sector outside the analyst comfort zone each year. - One red-team memo against a current thesis each quarter. - One audit of AI/automation outputs after every significant system change. - One annual review of source standards, accounting/regulatory changes, and internal research procedures. ## 18. Training completion gate - Advancement is based on work product, not reading completion. - The learner can reproduce facts and calculations under time pressure. - The learner can build a model without a template containing hidden logic. - The learner can explain accounting mechanics and valuation orally. - The learner can defend a thesis and update it when contradictory evidence arrives. - The learner maintains an error journal and shows process changes from repeated mistakes. - A senior reviewer can hand the learner an unfamiliar public company and receive a reproducible initiation without supplying the analytical answer. # MASTER LAB 13 - MULTI-SECTOR WORKED CASEBOOK WITH ANSWER KEYS > Purpose: force transfer of the framework across business models. Each fictional case contains enough data to form a preliminary view but also contains ambiguity. The correct answer is not a preselected stock rating. The correct answer is a reproducible economic diagnosis, model structure, valuation logic, and list of evidence required to resolve uncertainty. ## CASE A - AURORA CLOUD: SUBSCRIPTION SOFTWARE WITH DECELERATING RETENTION Aurora Cloud sells workflow software to mid-market and enterprise customers. The company reports 25% revenue growth and 22% free-cash-flow margin. Management emphasizes strong RPO growth and AI-related product demand. However, net retention has declined and stock-based compensation remains high. | Metric | FY1 | FY2 | FY3 | | --- | --- | --- | --- | | Revenue ($m) | 800 | 1,000 | 1,250 | | Ending ARR ($m) | 900 | 1,120 | 1,360 | | Net retention | 121% | 114% | 106% | | Gross retention | 94% | 92% | 89% | | RPO ($m) | 1,300 | 1,620 | 2,060 | | RPO growth | - | 25% | 27% | | SBC ($m) | 150 | 205 | 275 | | FCF ($m) | 120 | 190 | 275 | | Diluted shares (m) | 100 | 105 | 111 | | Sales & marketing / revenue | 42% | 40% | 39% | ### Analyst tasks 1. Reconcile revenue growth with ARR, NRR, gross retention, and RPO. 1. Identify at least three explanations for RPO growth remaining strong while retention falls. 1. Build a five-year ARR driver model with beginning ARR, new ARR, expansion, contraction, churn, and price. 1. Reconstruct FCF both before and after considering dilution economics. 1. Reverse the current valuation under two possible growth-duration assumptions. 1. Define the earliest thesis-break evidence. ### Answer key and reasoning Revenue growth remains strong because recognized revenue is a lagging result of prior bookings and contracts. Falling NRR and gross retention suggest the installed base is becoming less self-propelling. RPO can still grow if contract duration lengthens, enterprise mix rises, large multiyear renewals are signed, or new bookings offset weaker cohort economics. Therefore RPO growth alone does not resolve the retention concern. The model should forecast ending ARR from beginning ARR plus new ARR plus expansion minus contraction and churn. Revenue should then be derived from average in-period ARR and recognition timing rather than simply from ending ARR. A base case might assume NRR stabilizes near 105% while gross retention recovers modestly. A bear case might assume gross retention falls to 86% and NRR below 100%, forcing higher new-logo spending to sustain growth. Reported FCF is real cash, but SBC creates dilution. Compare cumulative FCF with cumulative increase in fully diluted shares and repurchase cash needed to offset issuance. If dilution is 5% to 6% annually, a per-share valuation must reflect that path. Do not subtract SBC mechanically twice. Either include compensation in operating economics and model dilution, or use another internally consistent approach. The reverse valuation should solve for the combination of growth duration and mature FCF margin implied by market price. The key question is whether falling retention is consistent with the required long-run growth. The most useful near-term evidence is cohort retention, renewal rate, new ARR productivity, RPO duration, and sales efficiency. A thesis break could be defined as NRR below 100% for two consecutive periods with gross retention below 88% and no evidence that a temporary product migration explains the decline. ## CASE B - VECTOR SILICON: AI ACCELERATOR SUPPLIER WITH SCARCITY MARGINS Vector Silicon designs accelerators and sells systems through cloud and enterprise partners. Revenue has tripled in two years because advanced-packaging and memory constraints limited industry supply. Gross margin expanded sharply. Two hyperscalers represent 58% of revenue and are developing custom silicon. | Metric | FY1 | FY2 | FY3 | | --- | --- | --- | --- | | Revenue ($bn) | 3.0 | 5.4 | 9.2 | | Accelerator units (000s) | 100 | 170 | 260 | | Average system ASP ($000) | 30.0 | 31.8 | 35.4 | | Gross margin | 52% | 61% | 68% | | R&D ($bn) | 0.8 | 1.1 | 1.7 | | Inventory ($bn) | 0.7 | 1.4 | 2.6 | | Purchase commitments ($bn) | 1.0 | 3.2 | 6.5 | | Top-two customer share | 44% | 51% | 58% | | Capex ($bn) | 0.2 | 0.3 | 0.5 | ### Analyst tasks 1. Decompose growth into units, ASP/content, and supply availability. 1. Estimate how much margin is attributable to structural differentiation versus scarcity. 1. Map foundry, advanced packaging, HBM, power, and customer deployment constraints. 1. Build an industry capacity response through three years. 1. Model a custom-silicon downside scenario. 1. Reverse the market value into required unit growth, ASP, and mature margin. ### Answer key and reasoning Revenue growth reflects both units and increasing ASP/content. The gross-margin jump is too large to assume fully structural without examining supply scarcity, product mix, and customer urgency. Build a normalized gross-margin bridge using product cost, yield, memory/packaging cost, pricing, and utilization. A base case can preserve a premium for software and performance while fading scarcity pricing as supply expands. The industry model should not stop at accelerator TAM. Track advanced packaging, HBM, foundry wafers, power delivery, networking, and data-center commissioning. If customer deployment capacity becomes the bottleneck, chip supply growth can exceed usable rack deployments and inventory can build. Rising purchase commitments increase both upside capacity and downside working-capital risk. Custom silicon should be modeled by customer and workload rather than as an arbitrary market-share haircut. Ask which workloads can move, what software switching cost exists, how performance per dollar compares, and whether hyperscalers are willing to diversify. A downside case might combine lower external accelerator share, price normalization, and lower gross margin while total compute demand remains strong. The reverse DCF should reveal whether current price requires both sustained 30%+ unit growth and margins near scarcity peaks. If so, the thesis is sensitive to normalization even if the AI end market remains healthy. Thesis-break evidence could be customer orders shifting materially to internal silicon while advanced-packaging lead times normalize and inventory rises faster than shipments. ## CASE C - TITAN INDUSTRIAL: CYCLICAL MACHINERY AT PEAK BACKLOG Titan manufactures heavy industrial equipment and aftermarket parts. Orders surged after a multiyear capital-spending recovery. Backlog is at a record, price realization is strong, and EBITDA margin is 18%, above the prior-cycle peak of 16%. Dealer inventory is also rising. | Metric | Trough | Mid-cycle | Current | | --- | --- | --- | --- | | Revenue ($bn) | 4.0 | 5.5 | 7.2 | | EBITDA margin | 8% | 13% | 18% | | Backlog ($bn) | 2.2 | 3.5 | 6.8 | | Book-to-bill | 0.75x | 1.00x | 0.92x | | Dealer inventory days | 55 | 70 | 105 | | Capex / revenue | 3.0% | 3.5% | 5.2% | | Aftermarket share of revenue | 24% | 27% | 29% | | Net debt / EBITDA | 3.5x | 2.0x | 0.8x | ### Analyst tasks 1. Separate backlog from future sustainable demand. 1. Estimate mid-cycle revenue, utilization, and margin. 1. Analyze whether record margin is price-cost, mix, utilization, or structural aftermarket improvement. 1. Model dealer destocking and book-to-bill below one. 1. Value the company on current and normalized earnings and explain the multiple inversion risk. 1. Assess whether current growth capex earns attractive incremental returns. ### Answer key and reasoning Backlog supports near-term revenue but does not guarantee cycle durability. Book-to-bill below one and dealer inventory rising are early caution signals. Reconstruct backlog additions, cancellations, price, and duration. Determine whether delivery delays inflated backlog and whether dealers are carrying inventory because of prior scarcity rather than end-demand growth. Normalize margin by separating aftermarket mix, realized pricing, material/freight normalization, and utilization. If aftermarket genuinely rose from 24% to 29% and carries better economics, some margin improvement may be structural. However, utilization and scarcity pricing should fade in a mid-cycle case. A reasonable normalized margin might sit above the old 13% mid-cycle level but below 18%, subject to evidence. The low current EV/EBITDA multiple can be misleading because the denominator is peak-cycle EBITDA. Compare enterprise value with mid-cycle EBITDA and FCF. In the bear case, orders fall, backlog converts, dealer inventories are reduced, utilization falls, working capital initially releases cash, then earnings decline. The balance sheet is stronger than in the last cycle, which can reduce permanent impairment even if earnings fall sharply. Growth capex should be evaluated on incremental units and normalized margins, not current peak pricing. If new capacity comes online after the demand peak, returns can disappoint. The key thesis-break question is whether aftermarket and structural share gains are enough to keep normalized returns above the cost of capital after utilization normalizes. ## CASE D - HARBOR BANK: DEPOSIT FRANCHISE, CREDIT NORMALIZATION, AND SECURITIES MARKS Harbor Bank has a strong regional deposit franchise and a large fixed-rate securities portfolio accumulated when rates were low. Reported book value fell because of unrealized losses. Deposit costs are rising, but credit losses remain low. Management argues that the securities marks will accrete back if held to maturity. | Metric | Year 1 | Year 2 | Current | | --- | --- | --- | --- | | Average loans ($bn) | 20 | 22 | 23 | | Average deposits ($bn) | 25 | 27 | 26 | | Deposit cost | 0.4% | 1.2% | 2.6% | | Loan yield | 4.5% | 5.3% | 6.2% | | NIM | 3.6% | 3.4% | 3.0% | | Net charge-offs / loans | 0.20% | 0.25% | 0.35% | | Allowance / loans | 1.20% | 1.25% | 1.30% | | Tangible common equity ($bn) | 2.8 | 2.9 | 2.5 | | AOCI securities loss ($bn) | 0.1 | 0.3 | 0.8 | ### Analyst tasks 1. Build NIM from average asset yields and funding costs. 1. Estimate deposit beta and deposit attrition under higher-rate scenarios. 1. Stress credit losses by loan type and compare allowance coverage. 1. Analyze economic versus accounting capital for securities marks. 1. Estimate normalized ROE and a residual-income or P/TBV valuation range. 1. Define a liquidity stress involving deposit outflows and asset sales. ### Answer key and reasoning The central issue is not whether the unrealized securities loss eventually accretes. The issue is whether the bank can avoid selling assets at a loss while funding deposits and meeting liquidity needs. Model cash, securities liquidity, borrowing capacity, deposit outflow, collateral, and regulatory capital. A mark that is economically temporary can become realized under liquidity stress. NIM compression reflects deposit repricing faster than asset yields. Estimate cumulative deposit beta and the portion of noninterest-bearing or low-cost deposits that can migrate. Loan repricing and securities maturities can restore margin over time, but the timing matters. A reverse scenario should test how long NIM can remain near 3% before normalized ROE falls below the cost of equity. Credit remains benign in the reported data, but reserve adequacy should be tested by portfolio. Use delinquency, criticized loans, charge-offs, loan-to-value, borrower cash flow, and macro sensitivity. Do not use the low current charge-off rate as the stressed loss assumption. Valuation should connect P/TBV to sustainable ROE, growth, capital needs, and cost of equity. If normalized ROE is only 8% against a 10% required return, a premium to tangible book is hard to justify without growth or franchise value. If deposit costs normalize and ROE returns to 13%, a higher multiple can be economically supported. The thesis break is a deposit/liquidity or credit state that forces capital raising or permanently lowers franchise returns. ## CASE E - CROSS-CASE SYNTHESIS The four cases demonstrate why one template cannot produce expert analysis. Aurora Cloud is driven by retention, sales efficiency, dilution, and duration. Vector Silicon is driven by physical supply, customer concentration, architecture, and scarcity normalization. Titan Industrial is driven by cycle, backlog quality, dealer inventory, and mid-cycle margin. Harbor Bank is driven by funding, credit, capital, and asset-liability management. The analytical operating system is common, but the economic model must change with the business. ### Final casebook exercise Select a fifth company in a sector not represented above. Write the five variables that determine its value, the five primary sources needed to underwrite those variables, the three accounting issues most likely to distort the numbers, the two leading indicators with the shortest causal path, and one thesis-break condition. If those cannot be stated before detailed modeling, the analyst does not yet understand the business well enough to build the model. ## Casebook completion gate - The learner can change modeling architecture across sectors rather than forcing one template. - Each case separates reported facts from assumptions and alternative explanations. - Valuation methods match claimant and business economics. - Cash, capital intensity, and dilution are modeled where economically relevant. - Each case has a concrete thesis-break mechanism and evidence plan. - The answer key explains reasoning and mechanics rather than providing a stock recommendation or one predetermined conclusion. # MASTER LAB 14 - FORMULA, MODELING, AND ANALYTICAL REFERENCE ENCYCLOPEDIA > Purpose: provide a fast desk reference for frequently used calculations while preventing formula worship. Every metric must be defined for the specific company, period, accounting treatment, and analytical question. A formula is only as good as the inputs and definitions supplied to it. ## 1. Growth and margin | Metric | Construction | Key caution | | --- | --- | --- | | Growth | Current / Prior - 1 | Reconcile acquisitions, FX, accounting presentation, and period length. | | CAGR | (Ending / Beginning)^(1/years) - 1 | Use actual elapsed years and avoid hiding path volatility. | | Gross margin | Gross profit / Revenue | Gross-versus-net revenue and depreciation classification can change comparability. | | Operating margin | Operating income / Revenue | Define whether restructuring, SBC, or amortization is included. | | Incremental margin | Change in operating profit / Change in revenue | Use comparable scope and period; negative revenue change can make interpretation unintuitive. | | Contribution margin | Revenue - variable costs, divided by revenue if expressed as percent | Requires true variable-cost identification, not accounting line labels. | ## 2. Working capital and cash conversion | Metric | Construction | Key caution | | --- | --- | --- | | DSO | Average receivables / Revenue x Days | Use credit sales if materially different and adjust for factoring where needed. | | Inventory days | Average inventory / COGS x Days | Forward COGS can be more useful in fast-changing businesses. | | DPO | Average payables / COGS or purchases x Days | Use denominator matching the payable and identify supplier finance. | | Cash conversion cycle | DSO + Inventory days - DPO | Not meaningful for all business models; deferred revenue can be a major source of float. | | Working-capital investment | Change in operating current assets - Change in operating current liabilities | Remove financing items and acquisition/FX effects when material. | | CFO conversion | CFO / Net income or operating profit measure | Interpret through business model and noncash charges, not as a standalone score. | ## 3. Returns on capital | Metric | Construction | Key caution | | --- | --- | --- | | ROIC | NOPAT / Average invested operating capital | Define operating cash, goodwill, leases, and capitalized intangibles consistently. | | Incremental ROIC | Change in NOPAT / Change in invested capital | Use multi-year windows when investment and returns are lumpy. | | ROE | Net income / Average common equity | Leverage and reserve/accounting differences can dominate. | | ROA | Profit measure / Average assets | Choose profit measure consistent with asset claimant. | | Reinvestment rate | Operating reinvestment / NOPAT | Intangible investment may be expensed and need analytical adjustment. | | Growth identity | Approx. reinvestment rate x incremental return | Steady-state relationship, not a mechanical short-term identity. | ## 4. Free cash flow | Metric | Construction | Key caution | | --- | --- | --- | | FCFF | NOPAT + noncash operating charges - operating reinvestment | Keep financing flows out and capture all growth/maintenance investment. | | FCFE | Net income + noncash charges - capex - working-capital investment + net borrowing +/- other equity cash items | Best when debt policy is modeled explicitly. | | Simple FCF | CFO - capex | Useful shorthand but can be distorted by working-capital timing, supplier finance, leases, and acquisition spending. | | Owner-oriented FCF | Normalized operating cash after maintenance investment and recurring economic costs | Requires judgment; preserve bridge to reported cash flow. | ## 5. Capital structure and credit | Metric | Construction | Key caution | | --- | --- | --- | | Net debt | Debt and debt-like claims - excess cash/non-operating liquid assets | Define lease/pension treatment and minimum cash. | | Net leverage | Net debt / normalized EBITDA or another relevant cash proxy | EBITDA add-backs and cycle normalization can change result sharply. | | Interest coverage | EBIT or EBITDA / cash or accounting interest | Match numerator to fixed-charge obligations and include leases where relevant. | | Fixed-charge coverage | Cash earnings before fixed charges / interest + lease + other fixed charges | Definition varies; state exactly. | | Covenant headroom | Covenant threshold minus contractual metric | Use the legal debt agreement definition, not headline leverage. | | Liquidity runway | Usable cash + committed capacity + cash generation - mandatory uses | Model timing and restrictions, not only annual totals. | ## 6. Equity and dilution | Metric | Construction | Key caution | | --- | --- | --- | | Market capitalization | Current share price x period-end shares | Not weighted-average shares used for EPS. | | Enterprise value | Equity value + debt and debt-like claims + minority/preferred claims - excess cash/non-operating assets | Bridge items depend on accounting and chosen operating metric. | | Net dilution | Change in shares adjusted for capital raises/acquisitions as analytically appropriate | Repurchases can mask gross employee issuance. | | Buyback yield | Repurchase cash / beginning or average equity value | Does not indicate value creation without repurchase price and dilution. | | Treasury-stock option dilution | In-the-money option shares less shares theoretically repurchased with exercise proceeds | Valuation-dependent and simplified for some awards. | ## 7. Valuation | Metric | Construction | Key caution | | --- | --- | --- | | EV/Revenue | Enterprise value / Revenue | Needs gross margin, opex, capital intensity, dilution, and mature economics. | | EV/EBITDA | Enterprise value / EBITDA | Can hide capex and working capital; normalize leases and cycle. | | EV/EBIT | Enterprise value / EBIT | More sensitive to depreciation policy but often better for asset intensity. | | P/E | Equity value / Net income | Affected by leverage, tax, non-operating gains, and buybacks. | | FCF yield | Free cash flow / Equity value | Define FCF and normalize working capital/capex. | | Terminal value | FCFF next period / (WACC - g) | Requires WACC > g and steady-state cash flow/reinvestment consistency. | | WACC | E/V x CoE + D/V x after-tax CoD + other capital weights | Use market weights and risk-consistent inputs. | | Residual income | Book value + PV of future (ROE - CoE) x beginning book | Useful when book capital is economically meaningful. | ## 8. SaaS and recurring revenue | Metric | Construction | Key caution | | --- | --- | --- | | ARR | Annualized recurring contracted or run-rate revenue under issuer definition | Definitions vary on usage, services, FX, acquisitions, and contract start. | | Gross retention | Beginning cohort recurring revenue retained before expansion / beginning cohort recurring revenue | Logo and revenue retention are different. | | Net retention | Beginning cohort recurring revenue after churn/contraction/expansion / beginning cohort revenue | Can remain above 100% while gross churn worsens. | | CAC payback | Customer acquisition cost / annualized new-customer gross profit | Allocation and timing of sales/marketing cost matter. | | LTV | Expected customer contribution profit over life discounted appropriately | Highly sensitive to churn, expansion, margin, and acquisition cost. | | Magic-number style sales efficiency | Change in recurring revenue or ARR annualized / prior sales and marketing spend | Useful only with consistent definitions and sales-cycle timing. | ## 9. Semiconductors and manufacturing | Metric | Construction | Key caution | | --- | --- | --- | | Utilization | Actual effective output / effective capacity | Capacity must reflect yield, downtime, mix, and qualified equipment. | | Yield | Good output / total processed output | Definition varies by wafer, die, module, or system. | | ASP | Revenue / units | Mix can dominate apparent price change. | | Inventory days | Inventory / COGS x days | Channel inventory outside the company also matters. | | Book-to-bill | Orders / billings or shipments under defined period | Definitions and cancellation risk vary. | | Content per system | Supplier revenue opportunity / end system | Avoid double counting channel markup and components. | ## 10. Banks | Metric | Construction | Key caution | | --- | --- | --- | | NIM | Net interest income / average earning assets | Tax-equivalent adjustments and balance mix vary. | | Deposit beta | Change in deposit rate / change in market reference rate | Depends on time window, product mix, and cycle. | | NCO ratio | Net charge-offs / average loans | Lagging credit measure. | | Allowance coverage | Allowance / loans or nonperforming loans | Portfolio risk and accounting regime matter. | | CET1 ratio | Regulatory CET1 capital / risk-weighted assets | Use current regulatory definition and transition rules. | | P/TBV | Price / tangible book value per share | Interpret through sustainable ROE, cost of equity, growth, and asset quality. | ## 11. Insurance | Metric | Construction | Key caution | | --- | --- | --- | | Loss ratio | Incurred losses / earned premiums | Prior-year reserve development can change the reported ratio. | | Expense ratio | Underwriting expenses / premiums under chosen definition | Acquisition-cost treatment varies. | | Combined ratio | Loss ratio + expense ratio | Below 100% usually indicates underwriting profit before investment income under common P&C definitions. | | Reserve development | Current estimate of prior accident-year losses versus prior estimate | Favorable development can support earnings but may not repeat. | | Float | Policyholder funds held before claim payment, conceptually | Economic value depends on underwriting cost and duration. | ## 12. Resources and projects | Metric | Construction | Key caution | | --- | --- | --- | | AISC | Sector-defined all-in sustaining cost per unit | Definition is non-GAAP and varies; reconcile to cash spending. | | Reserve life | Recoverable reserves / annual production | Production and reserve classification change. | | Capacity factor | Actual output / theoretical maximum output | Weather, outage, curtailment, and resource quality matter. | | Project IRR | Discount rate that sets project NPV to zero | Can hide scale, timing, reinvestment, and terminal assumptions. | | Project NPV | PV of after-tax project cash flows - initial/incremental capital | Use realistic construction, ramp, closure, and financing assumptions. | ## 13. Analytical unit controls - Record currency and whether amounts are nominal or real. - Record scale: units, thousands, millions, billions. - Record fiscal period and days in period. - Record whether values are point-in-time, average, or flow. - Record whether ratios use beginning, ending, or average balances. - Record whether metrics are GAAP/IFRS, regulatory, company-defined, or analyst-defined. - Record treatment of acquisitions, FX, discontinued operations, and recasts. - Record whether per-share metrics use basic, diluted weighted-average, or period-end fully diluted shares. ## 14. Formula completion gate A formula may be used in published research only when the analyst can identify every input, unit, period, definition, and reconciliation. If the formula is mathematically correct but the denominator does not match the economic exposure, the output is wrong. When in doubt, reconstruct the metric from source components and compare it with the issuer definition before using it in a model or peer table.