~/learn/Product Metrics · Engagement, retention & churn
Cohort retention curves
Group users by when they joined and read the shape of their retention curve.
why this matters
Retention curves are how interviewers expect you to talk about product-market fit.
A cohort is a group of users who joined in the same period (e.g. everyone who signed up in March). A retention curve plots what share of a cohort is still active on day 1, 7, 30, 90… Grouping by cohort stops a flood of new users from hiding the fact that older users are leaving.
The shape matters more than any single number. A curve that keeps falling to zero means the product never found lasting value. A curve that flattens (say at 25%) means a core group keeps using it — product-market fit. A 'smile' curve that dips then rises means lapsed users come back, often thanks to new features or network effects.
Three curve shapes
100% ┐
│╲
│ ╲___________ flattens → retained core (good)
│ ╲
│ ╲______╱‾‾ smile → users return (great)
│ ╲
0 └─────╲______ to zero → no lasting value (bad)
day 1 30 90Day-30 retention = users from cohort active on day 30 / users in cohort
Interview trap
Comparing a 2-week-old cohort's 'retention so far' to a 6-month-old cohort's is unfair. Compare cohorts at the same age (day 30 vs day 30).
Pro tip
If newer cohorts' curves sit higher than older ones, the product is improving. That's one of the strongest charts you can show in a case answer.
key takeaways
- Group users by sign-up period.
- A flattening curve means a retained core.
- Compare cohorts at the same age.
check your understanding
A retention curve keeps falling toward 0% over 90 days. What does it suggest?