Average metrics lie. "Our Day 30 retention is 15%" tells you nothing about whether things are improving or declining. Cohort analysis reveals the truth by grouping users who share a common characteristic—typically install date—and tracking their behavior over time.
Why Cohort Analysis Matters
Cohort analysis answers critical questions:
- Is our product improving? (Compare recent vs older cohorts)
- Which acquisition channels bring quality users?
- How long until users become profitable?
- When do users typically churn?
Retention Cohorts
The most common cohort analysis tracks retention—what percentage of users from each cohort remain active on Days 1, 7, 14, 30, etc.
Reading the Heatmap
In a retention cohort table, rows are cohorts (install weeks), columns are days since install. If recent rows are greener (higher retention), your product is improving. If older rows are greener, you have a problem.
LTV Cohorts
Track cumulative revenue per user over time by cohort. This reveals:
- How quickly users monetize
- Whether newer users are more or less valuable
- When cohorts reach profitability vs CPI
Segmented Cohorts
Beyond install date, segment cohorts by:
- Acquisition source: Compare Facebook vs Google vs organic
- Geography: US users vs LATAM vs APAC
- First action: Users who completed tutorial vs skipped
- Device: iOS vs Android, old vs new devices
Common Cohort Analysis Mistakes
- Too short window: Don't judge cohorts before they mature
- Ignoring sample size: Small cohorts have high variance
- Missing segmentation: Averages hide segment differences
- Not acting on insights: Analysis without action is wasted
Visualize Your Cohorts
ClicksFlyer's cohort analysis tools help you track retention and LTV by acquisition source, see trends over time, and identify which users deliver the most value.