Cohorts and retention
The Cohorts page groups your paying customers by the calendar month of their first successful payment, then tracks how many of each month's group are still subscribed one, three, six and twelve months later.
The three hero cards
| Card | What it shows |
|---|---|
| 90-day retention | The share of customers from mature signup months still subscribed at the three-month mark, pooled across every connected provider. Pooling walks backward from the newest mature month until it has counted at least three cohorts and twenty pooled starting customers, reaching back at most 24 months. Below either threshold the card shows a placeholder and names the two thresholds instead of a number. |
| Average tenure | How long a typical customer stays before cancelling, worked out from your combined monthly churn rate. Shows a placeholder when nobody cancelled in the trailing 30 days. |
| Lifetime value | What an average customer pays over their whole time with you: average revenue per customer divided by combined churn, shown in your detected currency. |
See churn and retention metrics and customer metrics for how tenure and lifetime value are defined everywhere else in SSA.
The cohort retention heatmap
Each row is one calendar month of new paying customers, newest month first; the table keeps roughly the newest thirteen months on screen out of the 24 months the engine tracks underneath.
- Cohort: the signup month.
- Started: how many customers entered that month, summed across every connected provider. This count is frozen the first time it is written, so a later sync can never push a cell above 100%.
- M1, M3, M6, M12: the share of that month's starting customers still subscribed one, three, six and twelve months on.
Each filled cell is a shade of pink on a six-step scale: pale for under 50% retained, moving through darker steps at 60%, 70%, 80% and 90%, to a solid brand pink with white text at 90% or above. A cell has no shade yet, and shows a dashed placeholder instead, until a sync runs on or after the first day of that month's marker.
A customer belongs to a cohort by the calendar month of their first successful, unrefunded payment, provided they already had a subscription that started on or before that payment. They count as retained at a given marker as long as any of their subscriptions started on or before that marker and had not been cancelled by it; a paused, past due, unpaid or trialing subscription still counts as retained until it is actually cancelled.
Best and worst cohort callouts
Two cards appear under the heatmap once at least two cohorts have a three-month figure to compare: the highest-retained cohort in green, and the lowest-retained in amber, each naming the month, its three-month retention and how many customers started it. They are computed from the same data as the heatmap, with no extra request.
Why members are leaving
A horizontal bar chart of the top eight cancellation reasons recorded in the last 90 days, pooled across every connected provider, longest bar first. Cancellations with no reason recorded are left out of the chart entirely. Bar labels are the raw codes stored by whichever provider recorded them, rather than rewritten into friendlier text. When nothing was recorded in the window, the chart is replaced by a short message saying so.
Currency and when it updates
Lifetime value is formatted in your detected primary currency. Every figure on this page recomputes automatically as part of each sync; there is no button to recalculate it by hand.