skills/thought-leadership/references/content-and-evidence.md

# Content formats and evidence

## Select the structure

| Asset | Organizing question | Useful content |
|---|---|---|
| Point of view | What should this reader reconsider or do differently? | Thesis, supported observation, mechanism, counterargument or boundary, practical implication. |
| White paper | How does a consequential problem work, and what should be done? | Standalone summary, evidence, method as needed, explanation, options, recommendations, sources. |
| Case study | What happened in this case, why, and what transfers? | Verified context, challenge, intervention, measured results, limitations, transferable lesson. |
| Industry brief | What changed and what does it mean for participants? | Dated market facts, drivers, affected segments, uncertainty, outlook, decisions. |
| Research report | What did the investigation establish? | Question, design, sample, method, results, uncertainty, interpretation, limits, supporting data. |

Follow user or publisher constraints rather than default word ranges. A case may lead with the result, but should not hide what was measured or who did the work. A public-company example should use public evidence instead of implying a consulting relationship.

## Evidence record

For a substantial piece, maintain claim, source, locator, date, population, method, limitation, and publication eligibility. For a small piece, inline citations and notes may suffice. Check whether sources independently support the claim or repeat one original study. Separate association from causality and absolute change from percentage change.

For original research, disclose sampling, collection period, exclusions, definitions, missing data, analysis approach, and limitations. Do not claim representativeness or statistical significance without a basis. For engagement patterns, define which cases were included and avoid selecting only successes.

## Openings without invented data

A strong opening can state a documented tension, explain a specific decision, or identify an overlooked mechanism. Use a sourced statistic only when it supports the argument and its context is preserved.

Fictional example of framing, not a research finding: “The pilot shortened drafting time. The investment case still depends on what happens to review effort and the capacity released.” This identifies a testable business question without inventing engagement history or a percentage.

## Practical recommendations

Explain the action, where it applies, resources or dependencies, and what evidence would justify continuing or changing course. Distinguish low-cost diagnostic steps from a full transformation commitment. Do not force a numbered framework or a fixed count of recommendations onto every piece.

## Publication preparation

Use the specific review and consent requirements that apply. Preserve approved attribution and exact quote context. Do not assume an anonymized case is safe to publish if its details identify the client or the contract restricts use. Draft source or approval gaps visibly for the reviewer.

When coordinating contributors, establish one narrative owner and consistent definitions. Check for contradictions with other supplied firm material and identify whether a change is intentional. Do not create a new practice-wide approval process or publication registry for a one-off draft.