Cohort Retention Calculator
100% private — runs on your device, never uploaded. Works offline once loaded.
Turn raw cohort numbers into a colour-coded retention heatmap and an average retention curve in seconds. Everything is computed in your browser, so your customer data never leaves your device.
What cohort retention analysis tells you
A cohort is a group of users who share a starting point — usually the month they signed up. Cohort retention analysis tracks what fraction of each cohort is still active after one period, two periods, and so on. Instead of one blended churn number that hides the truth, you get a grid: rows are cohorts, columns are the number of periods since acquisition, and each cell is the share of the original group that remains.
This structure is powerful because it separates the effect of when someone joined from how long ago they joined. A product that is genuinely improving will show newer cohorts retaining better at the same age than older ones — a signal you simply cannot see in a single aggregate metric.
The formula behind the heatmap
Retention for a cohort in period p is straightforward: divide the users still active in period p by the cohort's original size, then multiply by 100. The first number you paste on each line is that denominator, so period 0 is always 100% by definition.
- retention%(p) = active_users(p) ÷ cohort_size × 100
- The cohort_size is the first number in each row and never changes.
- The average curve for period p is the mean of retention%(p) across every cohort that has reached period p — cohorts too young to have that period are excluded.
Reading the colours and the curve
The heatmap shades each cell from red (low retention) through amber to green (high retention), so weak periods jump out visually. Look down a column to compare the same age across cohorts, and look across a row to watch a single cohort decay over time.
Healthy retention curves flatten out — they drop quickly at first and then settle onto a stable plateau, which represents your loyal core. A curve that keeps sliding toward zero with no flattening is a warning that the product has no durable base of habitual users.
When to use it and common caveats
Use cohort retention when you want to judge product-market fit, measure the impact of an onboarding change, or forecast the long-term value of newly acquired users. It is far more honest than a single monthly churn figure.
Be careful about a few things: keep your period definition consistent (all weekly or all monthly), remember that very recent cohorts have few data points and noisy percentages, and make sure 'active' means the same thing for every cohort. Small cohort sizes also make individual percentages volatile, so weight your conclusions toward the columns with the most data.
Frequently asked questions
What should the first number in each row be?
The size of the cohort at period 0 — the total number of users who started. Every later number in that row is measured against it, so getting this denominator right is essential.
Weekly or monthly cohorts — which is better?
Either works; the tool does not care about the unit. Weekly cohorts surface early onboarding drop-off faster, while monthly cohorts are steadier for longer-horizon retention. Just keep every row on the same cadence.
Why is my average row blank in later periods?
The average only covers cohorts that have reached that period. If only one young cohort has data there, the average reflects just that cohort; if none do, the cell is left empty.
Can I paste data exported from a spreadsheet?
Yes. Comma-separated or tab-separated values both work, one cohort per line. Blank lines and stray text are ignored.
How many periods can I chart?
As many as you paste. The retention curve automatically scales to the longest cohort, and the table adds a column for each period found in your data.
Does a flattening curve really matter?
Yes — the plateau is the single most important part of a retention curve. It represents the share of users who stick around indefinitely and is what ultimately drives compounding growth and lifetime value.
Advertisement