Product Metrics

Cohort Analysis

Cohort analysis groups users who share a common characteristic (typically the time period they signed up) and tracks their behavior over time — revealing retention patterns, churn timing, and product improvement impact.

In depth

A retention cohort table groups users by signup week/month, then tracks what percentage return in subsequent weeks/months.

What a healthy cohort looks like: - Week 0: 100% (signup week) - Week 1: 40–60% - Week 4: 20–30% - Week 12: 15–20% (and the curve flattens here)

If your retention curve never flattens — it keeps declining toward 0% — you don't have product-market fit. The flattening means you've found your 'retained core' who love the product.

Cohort analysis is more valuable than aggregate retention because it isolates the impact of product changes. If your Week 4 retention jumped from 15% to 25% for the cohort that signed up after you launched feature X — feature X drove retention.

Tools: Mixpanel (best for product analytics cohorts), Amplitude (enterprise-grade), PostHog (open source), Retention.ly, or build in SQL against your database.

Real example

Your January cohort showed 8% retained at Week 12. February cohort (after you redesigned onboarding) showed 19% at Week 12. That 11-point improvement tells you the onboarding redesign meaningfully improved retention — not visible in aggregate metrics.

Related terms

Ready to build your product?

Fixed price. 21-day delivery. Senior team.

Get a free scoping call →

Browse the glossary