Definition
Cohort analysis groups users by a shared starting point, usually the week or month they signed up, and follows each group over time. Because every user in a cohort started at the same time, cohorts show whether behaviour is improving, which totals hide.
The retention table
The most common form is a retention table: one row per cohort, one column per period since the start, and in each cell the share of the cohort that was active in that period.
| Cohort | Week 0 | Week 1 | Week 2 | Week 4 |
|---|---|---|---|---|
| 1 Sep | 100% | 42% | 33% | 27% |
| 8 Sep | 100% | 46% | 37% | 30% |
Reading down a column shows whether newer cohorts retain better.
Other cohorts
Cohorts can also be defined by acquisition source, first plan or first feature used. Comparing them shows which starts lead to lasting use.
How to measure it
You need a start event and a return event with a user ID. LogZuno's retention charts build the table from them; see product analytics.