Update to Toffu

{
  "description": "Explain marketing attribution with Toffu: attribution models, windows, incrementality tests and media mix models. Use when the user asks why platforms report different conversions, which attribution model to trust, or how to measure a channel's real impact.",
  "included_files": [],
  "name": "attribution-modeling",
  "skill_md_contents": "---\nname: attribution-modeling\ndescription: Explain marketing attribution with Toffu: attribution models, windows, incrementality tests and media mix models. Use when the user asks why platforms report different conversions, which attribution model to trust, or how to measure a channel's real impact.\n---\n\nUse the Toffu tools. Never apply a change the user has not confirmed.\n\n1. Call `get_account_context`. If a platform has several accounts, ask which one. Pass the ids as `account_ids` to `query_campaign_performance`, and as `account_id` to the other tools.\n2. Read data with `query_campaign_performance` (daily rows: `time_increment` \"1\"), `list_campaigns`, `list_ad_sets` and `list_ads`.\n3. To change something, call `propose_update_campaign`, `propose_update_ad_set` or `propose_update_ad`, and show the preview. Call `apply_change` only after the user confirms that change.\n\n# Marketing Attribution Methodology\n\n## Attribution Model Comparison\n\n### Rule-Based Models\n- **Last-click**: all credit to the final touchpoint. Simple but biases toward bottom-funnel channels.\n- **First-click**: all credit to the first touchpoint. Highlights acquisition channels but ignores nurturing.\n- **Linear**: equal credit to all touchpoints. Fair but undifferentiating.\n- **Time-decay**: more credit to touchpoints closer to conversion. Good for short sales cycles.\n- **Position-based (U-shaped)**: 40% first, 40% last, 20% distributed across middle. Balances acquisition and closing.\n\n### Algorithmic Models\n- **Data-driven attribution** (GA4, Google Ads): uses machine learning on your data to distribute credit.\n- Requires sufficient conversion volume (typically 300+ conversions/month for reliability).\n- More accurate than rule-based but less transparent — treat as a black box.\n\n## Incrementality Testing\n- Attribution models measure correlation. Incrementality measures causation.\n- **Geo-holdout tests**: suppress a channel in test markets. Compare conversion rates to control markets.\n- **Ghost ads / PSA tests**: serve a public service ad to the control group instead of your ad.\n- **Conversion lift studies**: platform-native (Meta, Google). Randomized control/exposed groups.\n- Run incrementality tests quarterly on your largest spend channels.\n\n## Media Mix Modeling (MMM)\n- Regression-based approach using aggregate data (spend, impressions, external factors like seasonality).\n- Doesn't require user-level tracking — works in a cookieless world.\n- Best for budget allocation decisions across channels at the strategic level.\n- Limitations: needs 2+ years of data, slow to reflect rapid changes, expensive to build in-house.\n\n## Cross-Device Tracking\n- Users convert across devices. Mobile discovery -> desktop purchase is common.\n- Use Google Signals (GA4) for cross-device reporting — requires users to be signed into Google.\n- Supplement with CRM data: email-based identity resolution across touchpoints.\n\n## Offline Conversion Import\n- Close the loop between online clicks and offline outcomes (phone calls, in-store purchases, closed deals).\n- Import offline conversions to Google Ads and Meta via their respective APIs.\n- Match on GCLID (Google) or FBCLID (Meta). Automate the upload pipeline.\n- Offline data dramatically improves algorithmic bidding for lead-gen businesses.\n\n## Attribution Window Selection\n- Shorter windows (7-day click) are more conservative but may undercount longer journeys.\n- Longer windows (28-day click) capture more conversions but include more noise.\n- Match the window to your typical sales cycle. B2C impulse: 7-day. B2B consideration: 28-90 day.\n"
}