one-line definition
Cohort analysis groups users by when they signed up and tracks their behavior over time, revealing whether your product is improving for newer users.
formula: Group users by signup week or month. Track a key metric (retention, revenue, activation) for each cohort over time.
tl;dr
Aggregate metrics hide trends. A 70% monthly retention looks stable, but cohort analysis might show January signups retain at 80% while March signups retain at 55%. That tells you something broke in March — aggregate numbers never would.
Simple definition
Cohort analysis takes your users, groups them by when they joined (usually the week or month they signed up), and then tracks how each group behaves over time. Instead of asking "what's our retention rate?", you ask "is retention getting better or worse for each new batch of users?" This is the difference between a snapshot and a trend. For solo founders, cohort analysis is how you know whether your product changes are actually helping.
How to calculate it
Pick a time-based grouping (weekly or monthly signup cohorts). For each cohort, track a metric at fixed intervals. For retention, you'd measure: of users who signed up in Week 1, how many were active in Week 2, Week 3, Week 4, and so on.
Method: Group users by signup week or month. Track a key metric (retention, revenue, activation) for each cohort over time. Display as a table where rows are cohorts and columns are time periods.
Most analytics tools (Mixpanel, Amplitude, PostHog) build cohort tables automatically. In a spreadsheet, create a row per signup month and a column per month-since-signup.
Example
You launch a new onboarding flow on March 1. Your February cohort (old onboarding) shows 60% week-1 retention and 35% week-4 retention. Your March cohort (new onboarding) shows 72% week-1 retention and 48% week-4 retention. That's a clear improvement. But your aggregate retention for March still looks similar because old cohorts are mixed in. Without cohort analysis, you'd miss that your change worked. With it, you know to keep iterating on onboarding.
Related reading
Related terms
- Retention Cohort
- Churn Rate
- Activation Rate
FAQ
How many users do I need before cohort analysis is useful?+
You need at least 30-50 users per cohort for patterns to emerge. Below that, individual behavior dominates. Use weekly cohorts if your signups are small to get usable group sizes faster.
What is the most important metric to track by cohort?+
Retention at week 4 or month 1. If newer cohorts retain better than older ones, your product changes are working. If retention is declining cohort over cohort, something is degrading — usually onboarding or expectation-setting.