Mobile App Retention Benchmarks by Category (2025–2026)

Most mobile apps lose three quarters of their new users within a day and almost all of them within a month, which makes retention the metric that separates chart regulars from one-week wonders. Here are the published 2025–2026 numbers, with their sources and dates, and a frank note on how much to trust each.

Why the numbers disagree

Before reading any table, know that benchmark publishers measure different populations with different methods. GameAnalytics pools more than 16,000 games with at least 1,000 monthly active users and uses classic (exact-day) retention ¹. Business of Apps and Phiture compile figures from attribution vendors such as AppsFlyer and Adjust, whose data skews toward paid installs at larger advertisers ^3,4. Apple and Google report your own app's retention from opted-in or modeled usage data and compare it with percentile bands, not averages ^6,7. A "26% D1" from one source and a "22% D1" from another can both be right.

Cross-category benchmarks for apps

The most recent compiled app figures, with platform splits, come from Business of Apps (updated July 2026, citing AppsFlyer) and Phiture (updated March 2026):

Phiture does not attribute each row to a specific vendor or data year; it cites a shared pool of Statista (H1 2024), AppsFlyer, Adjust, Business of Apps and Sendbird, so individual category figures are best treated as indicative unconfirmed ⁴. Business of Apps' qualitative ranking agrees with the table: news and magazine apps retain best, while generative AI and photo and video apps retain worst ³.

Adjust's retention handbook gives a global D1 of 26%, with North America lowest at 23% and gaming the strongest vertical at up to 28–29% D1; the handbook does not state the data year, so treat the vintage as unverified ⁹.

Benchmarks for games

Games have by far the richest open data, because GameAnalytics publishes percentile bands rather than a single average. The figures below cover calendar-year 2025 and were published in early 2026 ¹:

Those medians are lower than the app tables above because the pool includes thousands of small and indie titles. The previous edition (2024 data, published February 2025) showed the same shape: median D7 of 3.4–3.9%, down from 4–5% in 2023, and three quarters of games under 3% at D28 ². In other words, game retention has been drifting down for two consecutive years.

Genre matters. GameAnalytics' 2025 edition found arcade games lead on D1 but fade fast, while board, card, puzzle and casino games hold up best in the medium and long term; its 2026 edition withheld genre tables while it reworks its genre taxonomy ^1,2. Business of Apps reports RPG and sports games retaining best on iOS, and puzzle and RPG on Android ³.

iOS versus Android

Every source that splits by platform shows iOS ahead. Business of Apps puts the gap at about 5 points on D1 and 1.5 points on D30 for apps overall ³. For games, GameAnalytics' 2025 edition reported top-quartile D1 of 31–33% on iOS against 25–27% on Android, even though Android users logged more sessions per day ². The usual explanations are device-quality differences, higher iOS acquisition costs that filter for more committed installs, and a larger share of pre-installs and low-end devices on Android. None of these is proven in the reports themselves.

The subscription data adds a twist: RevenueCat found that nearly 31% of Google Play subscription cancellations in 2025 were involuntary billing failures, versus 14% on the App Store, so a chunk of the Android "churn" gap is payment plumbing rather than disengagement ⁵.

Games versus apps

Set side by side, a median game retains roughly 22% at D1 and under 1% at D30, while a typical app retains 20–25% at D1 and 4–5% at D30 ^1,3. Games win the first session and lose the month. Sensor Tower's State of Mobile 2026 shows the commercial consequence: in 2025, in-app purchase revenue from non-game apps passed games for the first time, growing 21% year over year against 1.3% for games, even though games still drew nearly $82 billion ⁸. Long-tail retention is where subscription apps make their money.

What good looks like

A workable rule of thumb from the tables above:

For games, use the GameAnalytics percentiles instead: a D7 of 7% already puts you in the top quarter, and 25% is top-1% territory ¹. For subscription apps, pair retention with RevenueCat's conversion medians: 2.0% download-to-paid at day 35 globally, 10.7% for hard-paywall apps, and trial-to-paid between 25% (games) and 43.5% (travel) in 2025 ⁵.

The best check is the one inside the stores. App Store Connect benchmarks show your D1, D7 and D28 retention against the 25th, 50th and 75th percentile of your peer category ⁶; Play Console's "Compare to peers" tab does the same, scale-normalised, for day 2 through day 28 retention ⁷. Being above the 75th percentile in your own category means more than beating any global average.

How this shows up in the charts

Chart position is a flow metric: Top Free ranks reward downloads over a short window, and a burst of paid installs can lift an app regardless of retention. Retention decides what happens next. An app with a 25% D1 and 1% D30 needs to keep buying the same volume every week to hold its rank, while an app at 8% D30 accumulates a base of returning users who generate ratings, word of mouth and re-engagement that compounds into organic downloads. Neither Apple nor Google publishes how engagement feeds its ranking algorithms, so any claim that retention directly boosts rank is unverified. What the published data does support is indirect: the categories with the strongest D30 retention (news, finance, marketplaces) are also the ones where the same apps occupy the top of their category charts year after year, and the fastest-fading category in Business of Apps' ranking, generative AI, is also the one where Sensor Tower recorded downloads doubling in 2025 while sessions grew even faster, a sign of a market still sorting keepers from churners ^3,8.