Retention Analysis
Retention analysis is the method of measuring whether, and how often, users return to a product after a starting action such as signing up. It groups users by when they started, tracks the share of each group that comes back in later periods and plots the results as curves. This shows how quickly users are lost, where usage levels off and which changes or behaviors improve it.
How Retention Analysis Works
Every retention analysis needs three definitions:
- A starting event that puts a user into a group, such as sign-up or first purchase.
- A return event that counts as coming back. Any activity is the loosest option. A value action, such as creating a report, is stricter and more meaningful.
- A time unit, such as days, weeks or months, matched to how often the product is naturally used.
Users are grouped into cohorts by start date, and the results appear as a table with one row per cohort and one column per period. Analytics tools offer different ways of counting a return. Amplitude's documentation describes three: N-day retention counts users who return on a specific day, unbounded retention counts users who return on that day or any later day, and bracket retention counts returns within custom windows such as days 1 to 7 and days 8 to 14. The choice changes the numbers a lot, so it should be stated with every result.
The curve shape matters more than any single point. Most products lose many users early. A curve that flattens shows a group of users who have found lasting value. A curve that keeps declining toward zero shows the product is not holding anyone.
Why Retention Analysis Matters
Retention is one of the clearest signals of product value, and retention analysis is how teams measure it reliably. Comparing cohorts shows whether the product is getting better: if users who joined after an onboarding change retain better than earlier cohorts, the change likely helped. Splitting cohorts by behavior shows which early actions go with users staying, which gives teams hypotheses to test, though correlation alone does not prove cause.
It also explains numbers that look fine in aggregate. Total active users can grow while each new cohort retains worse than the last, hidden by strong acquisition. Retention analysis exposes that before it shows up as rising churn. For how retention cohorts fit into judging product-market fit, see measuring product-market fit with AARRR metrics and retention cohorts.
Retention Analysis Example
A habit-tracking app runs weekly retention using "logged a habit" as the return event. Users who signed up in January show 45 percent active in week one, 28 percent in week four and a curve that flattens near 22 percent by week eight. After the team adds a reminder setup step to onboarding, the March cohorts flatten near 27 percent. Splitting March users by whether they set a reminder shows a large gap between the two groups, so the team makes reminder setup more prominent and keeps watching the next cohorts.
Retention Analysis vs. Retention and Cohort Analysis
These three terms are often used interchangeably. Retention is the outcome: the share of users who keep using the product. Cohort analysis is the broader technique of grouping users by a shared start or trait and comparing any metric over time, including revenue or feature use. Retention analysis applies that cohort approach specifically to return behavior.