Cohort Retention Calculator

Cohort retention groups users by when they joined and tracks how many are still active in each later period. Period 0 is the group's starting size, so it reads 100 percent. A cohort that starts with 100 users and keeps 60 by period 2 has 60 percent retention there. Reading cohorts side by side shows whether newer groups stick around better than older ones.

For each cohort, retention at a period is the users still active divided by the cohort's starting size, multiplied by 100. So 60 of an original 100 users is 60 percent. Because the denominator is always the starting size, every row begins at 100 percent and you can compare any later period directly against the start.

The average curve weights each cohort by its starting size rather than treating every cohort equally, so a cohort of 1,000 users counts more than one of 10. Each period only averages over cohorts that are old enough to have data there, which stops a brand new cohort from distorting longer-horizon retention. The churn curve is simply 100 minus average retention at each period.

Yes. Older cohorts naturally have more periods of history than recent ones, so you can leave newer cohorts shorter. The heatmap leaves missing cells blank, and the average retention for a given period only includes cohorts that have actually reached it.

It can if a cohort wins back lapsed users or reactivates dormant accounts, since retained users in a later period can exceed an earlier count. The calculator allows this so the numbers stay faithful to your data, and the matching churn value simply turns negative for that period.

No. Everything runs entirely in your browser, so your cohort names, user counts, and retention figures never leave your device or get stored on a server.