Cohort Retention Calculator
A cohort retention calculator turns raw user counts into a retention triangle: f
A cohort retention calculator turns raw user counts into a retention triangle: f
How to use Cohort Retention Calculator
- Enter one cohort per row: an optional label, the starting user count, then the active users in each later period, separated by commas.
- Read the retention triangle to see each cohort's percentage retained per period, with the weighted average in the bottom row.
- Check the per-period retention stats below the table, then use Copy table to paste the result into a doc or spreadsheet.
About Cohort Retention Calculator
A cohort retention calculator turns raw user counts into a retention triangle: for each cohort (usually a sign-up month or week), it shows what percentage of the original users were still active in each later period. Enter one row per cohort — an optional label, the starting number of users in period 0, then how many remained active in each following period. The tool computes retention as active users divided by that cohort's own starting size, so every row is measured against its own period-0 baseline.
Under the triangle it reports a weighted average retention curve: for each period it sums the still-active users across every cohort that reached that period and divides by the sum of their starting sizes. Weighted averaging gives larger cohorts proportionally more influence, which is usually what you want for a blended retention picture; a plain per-cohort average can read differently when cohort sizes vary a lot. The tool does not project or curve-fit future retention — it only reports the numbers you enter.
Everything runs locally in your browser. Your cohort numbers are never uploaded, stored, or sent anywhere, and there is no sign-up. You can paste data straight from a spreadsheet (commas, tabs, or semicolons all work) and copy the finished table back out.
Frequently asked questions
- What numbers do I put in each row?
- Start each row with an optional label (like a month name), then the number of users in the cohort at period 0, then how many of those same users were still active in period 1, period 2, and so on. For example: Jan, 1200, 540, 384 means 1,200 users joined in January, 540 were active the next period, and 384 the period after.
- How is retention percentage calculated?
- Each cell is that period's active-user count divided by the cohort's own period-0 starting size, times 100. So a cohort that started with 1,200 users and had 540 active in period 1 shows 45.0% retention for period 1.
- What is the weighted average retention shown below the table?
- For each period it adds up the still-active users across every cohort that reached that period, then divides by the sum of those cohorts' starting sizes. This weights bigger cohorts more heavily, giving a blended retention rate rather than a plain average of the individual percentages.
- Can rows have different lengths?
- Yes. Newer cohorts naturally have fewer periods of data. The calculator lines everything up into a triangle and only averages a period across the cohorts that actually reached it, so partial rows are handled correctly.
- Does it predict future retention?
- No. It only reports and averages the counts you enter. It does not fit a curve, extrapolate, or forecast churn for periods you haven't provided data for.
- Is my data private?
- Yes. All parsing and math happen in your browser with JavaScript. Nothing is uploaded or saved, and there is no account or sign-up.