Each install day, tracked: how many were still active N days later.

Apple's own measurement, not ours. Each row is one day's new installs (Devices = that day's cohort). Each column N shows the share of that cohort active again N days after installing. Read any row left to right: it barely decays.

How to read: Jul 3 row, column 7 = of the 95 devices installed on Jul 3, 77.9% opened the app again 7 days later. Blank = window not yet elapsed (or no activity that day in a tiny cohort). Brighter = higher. 0% 85%

Opt-in sample: Apple reports only devices whose owners enabled "Share With App Developers" in iOS settings — industry-typical opt-in is ~20–30% of users. The Devices column is therefore a fraction of each day's real installs (actual install counts run ~3–5× higher). The sample is Apple-selected, not ours — absolute counts are a sample; the retention shape is the signal.

The signal: the Average row holds between 32% and 45% from day 1 through day 23 — a flat line where casual games typically decay toward single digits. Small-cohort rows (Jul 7–8, Jul 21–26: 3–13 devices) are noisy by nature; Jul 21–26 installs were additionally suppressed by the store-availability pause, so those rows are thin — the pattern carries in the large cohorts (Jul 3–4, Jul 14–16, Jul 27+).

Source: Apple App Store Connect · Daily Retention (opt-in devices only) · captured Aug 4, 2026 · companion to iSpotly's Numbers