Attribution Model Reconciliation

Reconcile {{model_pair}} attribution models:

1. Model descriptions:
   - Definition of each model (last-click, first-click, linear, time-decay, multi-touch, etc.)
   - Key assumptions and limitations

2. Comparison metrics:
   - Revenue (or conversions) by channel
   - % disagreement between models
   - Channels where models diverge most

3. Root cause analysis:
   - Long vs short customer journeys
   - Multi-channel vs single-channel users
   - Seasonality / promotional period effects

4. Business impact:
   - How would budget allocation change with each model?
   - Which model better reflects true business causality?

5. Recommendation:
   - Which model should we adopt for decision-making?
   - Hybrid or blended approach?
   - Data quality assumptions to validate

6. Next steps:
   - How to pressure-test recommendations with controlled experiments (e.g., turning off a channel)?

Deliverables: comparison table + waterfall chart + recommendation memo.