Why Cohort Retention Rate matters
Improving curves suggest newer cohorts are staying active longer. A drop at a particular age can identify the moment when value, habit, or communication weakens.
- Business question
- How many people from the same starting group are still active after the same amount of time?
- Teams that use it
- Product, growth, customer success, lifecycle marketing, and analytics teams.
- Decisions it supports
- Onboarding changes, lifecycle programmes, product releases, acquisition quality, and retention experiments.
Cohort Retention Rate formula
(Cohort Retained Users ÷ Cohort Total Users) × 100
Formula components
- Cohort total users
- Distinct users in the cohort at its defined start.
- Cohort retained users
- Members of that same cohort who satisfy the active rule at the measured age.
- Cohort definition
- The shared event or attribute used to group users, such as first purchase month.
- Cohort age
- Elapsed time since the cohort began, such as week 4 or month 3.
How to calculate Cohort Retention Rate
- Define the cohort entry event, active event, and time interval.
- Count distinct users when the cohort begins.
- At the chosen cohort age, count only original members who meet the active rule.
- Divide retained members by original cohort size and multiply by 100.
Cohort Retention Rate example
A January sign-up cohort contains 1,200 users. In the third month after sign-up, 510 of those original users complete the defined active event.
- Cohort total users = 1,200.
- Month-3 retained users = 510.
- Cohort Retention Rate = 510 ÷ 1,200 × 100 = 42.5%.
Month-3 Cohort Retention Rate is 42.5%.
Just under half of the January cohort met the activity rule in its third month.
How to interpret the result
Compare cohorts at the same age. January at month 6 cannot be compared fairly with June at month 1. The shape of the full retention curve is often more informative than one point.
Expected retention varies by use frequency, customer type, acquisition channel, season, cohort definition, and activity rule. Document each of those choices.
Common mistakes and limitations
- Comparing different cohort ages
- Older cohorts have had more time to lose users.
- Letting users enter twice
- A person should not be duplicated inside the same cohort.
- Changing the active event
- A login, purchase, and core action represent different levels of retention.
- Small-cohort noise
- Rates from small groups can move sharply when only a few users change status.