Help me structure a cohort analysis on {{behavior or metric — retention, revenue, feature usage, etc.}}.
Context:
- What I want to learn: {{the question}}
- Cohort dimension I'm considering: {{e.g., signup month, plan tier, acquisition channel}}
- Data I have: {{rough description}}
Output:
1. **The cohort definition** — what makes a cohort, and why this dimension vs alternatives?
2. **The metric** — exactly how to measure (formula, time windows, edge cases)
3. **The shape of the analysis** — table, heatmap, line chart? With what dimensions on each axis?
4. **What patterns to look for** — what's a "good" cohort curve vs "bad"?
5. **Common artifacts** — patterns that look real but are explained by data quality or definition issues
6. **What I should compare cohorts against** — peer cohorts, baseline, expectation
Then tell me: what would each plausible result MEAN for the business? If I see X, I should do Y.cohortsretentionanalysis