What a cohort is
A cohort is a group of users who started at the same time — for example, everyone who installed in the first week of March. Cohort analysis follows each group separately, so you can compare users who arrived before a change with users who arrived after it.
Example: a retention table
| Install week | Users | Week 1 | Week 2 | Week 4 |
|---|---|---|---|---|
| 3 Mar | 1,200 | 38% | 27% | 19% |
| 10 Mar | 1,350 | 37% | 26% | 18% |
| 17 Mar (new onboarding) | 1,280 | 45% | 33% | 24% |
| 24 Mar | 1,410 | 46% | 34% | — |
- Read across a row to see how one cohort drops off.
- Read down a column to compare cohorts. Here, the new onboarding from 17 March clearly retains better.
An average retention number for all of March would have hidden this.
What to analyze by cohort
- Retention and churn
- Revenue per user, ARPPU and lifetime value
- Conversion to paid or to a key action
- Cohorts by source: organic vs paid, or by campaign — see user acquisition
Tips
- Pick the cohort size by volume: daily for big apps, weekly or monthly for small ones.
- Mark releases and campaigns on the timeline, so you know which change a cohort saw.
- Compare cohorts with what users wrote in reviews in the same weeks — a drop in retention and a run of one-star app reviews usually have the same cause.
The free retention rate calculator works out retention for one cohort; more in app retention strategies.