Cohorts
The Cohorts page: monthly subscriber retention and cumulative proceeds per subscriber from Apple's subscriber report, with the exact cohort and retention rules.
The Cohorts page (/cohorts) shows a retention triangle: rows are cohort months, columns are months since the cohort month. It is built from Apple's SUBSCRIBER report and honors the app filter.
Views
| Toggle | URL | Cells |
|---|---|---|
| Retention % | ?view=retention (default) | Share of the cohort active that month |
| Proceeds per subscriber | ?view=ltv | Cumulative USD proceeds per cohort subscriber (the LTV curve) |
Cells are shaded green in proportion to the value (retention against 100%, proceeds against the deepest value on the page). A · is a month with no value; blank cells are months the cohort has not reached yet. Columns are labelled M0, M1, and so on; rows run oldest first.
Rules
cohort month = month of the subscriber's earliest observed event in that app
active (month i) = distinct subscribers with any charge, renewal or refund event in that month
cohort size = active count in M0
retention % = round(active / size × 100)
proceeds per sub = round(Σ proceeds from M0 through month i / size, 2)Apple's report leaves the purchase date blank, which is why the cohort month is the first observed event. Proceeds are converted to USD at each event's own date.
Yearly plans only produce an event at renewal, so mid-year months read low for annual products.
Subscriber IDs are used only to fold events into counts; they never appear in the page or any API response.
Empty state
"No cohort data yet." Cohorts build up as daily syncs ingest Apple's subscriber-level report. Give it a sync or two.
Last updated on
Attribution
The Attribution page: installs by Apple source type, top referrers, and Apple Ads spend, installs, cost per install and ROAS on proceeds, with every formula.
Apps
The Apps list and app detail page: revenue tiles, top products, Apple Ads, TestFlight, ingest URL, S2S audit, forwarding, nickname, history import and refunds.