Online Tool Store
Cohort Analysis Builder
Group users by signup period and lay out a retention cohort table, with period-over-period retention and the shape of the drop-off curve.
🔒 This tool runs entirely in your browser. Your files are never uploaded to a server.
Data & CSV
Cohort Analysis Builder
Frontend preview — no upload or external service.
Cohort table
Monthly cohorts: month 1 retention averages 42%, settling near 28% by month 3 — the March cohort is the outlier at 51%.
How the Cohort Analysis Builder works
- Provide signup dates and activity dates, one row per user event.
- Choose the cohort period; weekly suits fast products, monthly suits most.
- Read down a column to compare cohorts at the same age, not across a row.
The method
Retention for a cohort at period N is the share of that cohort's original members who were active in period N, always measured against the cohort's own starting size.
retention(N) = active members in period N / members at period 0
A cohort of 500 with 140 active in month 3 retains 28% — and comparing that with another cohort's month 3, not its month 1, is the whole point of the table.
FAQ
Why compare down the column rather than across the row?
Across a row you are watching one cohort age, which always declines. Down a column you compare different cohorts at the same age, which is what tells you whether the product is improving.
What counts as retained?
Whatever action means value in your product — a login is a weak signal, a completed core action is a strong one. Pick one definition and never change it mid-analysis.
Why does the newest cohort look bad?
It is usually incomplete: its later periods have not happened yet. Exclude partial periods or the curve will mislead you.
How we compare
| Feature | Online Tool Store | A spreadsheet | A BI platform |
|---|---|---|---|
| Cohort table laid out for you | ✓ | Pivot setup | Yes |
| Flags incomplete periods | ✓ | ✗ | Sometimes |
| Data stays in the browser | ✓ | ✓ | ✗ |
| No analytics subscription | ✓ | ✓ | ✗ |
Cohort Analysis Builder is opinionated about the reading: same-age comparison down the column, and partial periods marked so they cannot flatter you.