A cohort is a group of users who share a defining event within the same time window — most commonly, users who first signed up or made their first purchase in the same week or month. Cohort analysis tracks what happens to those users over time, revealing patterns that aggregate metrics hide.
Why Cohort Analysis Matters
Aggregate metrics lie by averaging across groups with very different histories. If your product is growing, new users are always diluting your retention numbers. A 60% day-30 retention rate could be trending up or down — you can't tell from the aggregate alone.
Cohort analysis isolates these groups so you can answer:
- Are users acquired this month retaining better or worse than last month's cohort?
- Did a product change in January improve or hurt long-term engagement?
- Which acquisition channels produce the highest-LTV cohorts?
Reading a Cohort Retention Table
A standard cohort retention table shows each cohort's retention rate at each time interval:
| Cohort (signup month) | Week 1 | Week 2 | Week 4 | Week 8 |
|---|---|---|---|---|
| Jan 2024 | 68% | 45% | 32% | 24% |
| Feb 2024 | 71% | 48% | 35% | 27% |
| Mar 2024 | 74% | 52% | 39% | — |
Each row is a cohort. Each column shows what fraction of that cohort was still active at that point. Improving numbers left-to-right within a row are good; improving numbers top-to-bottom within a column indicate your product is improving over time.
Cohort Analysis in CRO
Cohort analysis connects experimentation to long-term outcomes:
- A/B test follow-up — Did users who converted via a winning variant retain better or worse than control users over 90 days?
- Acquisition quality — Is a new marketing channel sending users who convert quickly but churn fast?
- Onboarding changes — Do users who experienced the new onboarding flow in March have better 30-day retention than the February cohort?
Cohort analysis is especially valuable for subscription products where conversion is just the start of the revenue relationship.