Mobile game retention benchmarks: what to compare first

Retention benchmarks can tell a game studio where to look. They cannot tell you why players leave. Use them to find a question, then use your own player cohorts to answer it.

This guide shows how to read Day 1, Day 7 and Day 28 results. It also shows when a broad benchmark can mislead your team.

What the published figures actually say

GameAnalytics' 2025 Mobile Gaming Benchmarks report studies 11,600 games across nine regions, two mobile platforms and 16 genres. Its data covers 2024. The title is the report edition, not the collection year.

In that report, games in the top 25% reached roughly 26% to 28% Day 1 retention. Median Day 7 retention ranged from 3.42% to 3.94%. The report says 75% of games did not clear 3% at Day 28.

These are different groups and measures. Do not read the three figures as one game's retention curve. A top-quartile Day 1 result beside a median Day 7 result creates a false story.

Day 1 asks whether players return after their first session. Day 7 asks whether the first week gives them a reason to continue. Day 28 tests whether that reason lasts. The report uses Day 28, so do not silently relabel its last figure Day 30.

Match the benchmark to your game

A benchmark is useful only when its comparison group resembles your players. Start with these cuts:

GameAnalytics found that top-quartile iOS games reached about 31% to 33% Day 1 retention. Android games in the same quartile reached about 25% to 27%. That does not mean changing platform improves a game. It means a platform-mixed target can hide a real difference in player mix.

Use a genre-specific median when you can. Still label the source, year and sample. A report drawn from one analytics platform reflects the games that use that platform. It is a reference, not a universal standard.

Build a comparable internal cohort

First define a return. An app open is easy to count, but it may include players who quit before playing. For a game, consider a completed level, match, quest or other meaningful action. Keep the action and time zone fixed during a comparison.

Then group players by install date. Count how many eligible installers perform your return action on Day 1, Day 7 and Day 28. Wait until every player in the group has had the full chance to reach that day. Never compare a complete older cohort with a newer cohort whose window is still open.

For example, suppose 1,000 players install on the same day. If 240 complete another level on Day 1, that cohort's meaningful Day 1 return rate is 24%. This is an illustrative calculation, not an industry figure. Do not compare it directly with a benchmark based on app opens without noting the different return rule.

Record a small table with install date, platform, genre or game mode, acquisition source, version, eligible installers and returners. Keep the denominator visible. Exclude team tests and duplicate accounts where possible.

Turn a gap into a product question

If Day 1 lags a fitting peer group, inspect the first session. Can players reach the first useful action? Do they understand the next move? Look at tutorial exits, first failures and time to first win.

If Day 1 is sound but Day 7 falls, inspect what happens after the first win. Are goals clear? Does progression stall? Does the next session offer anything new? A larger welcome reward may increase claims without improving play.

If later return falls, examine content pacing and the long-term goal. Separate players who never came back from players who returned but stopped progressing. These are different problems.

Do not redesign from one small cohort. Watch a few sessions around the break and read matching player feedback. Choose one change, define a meaningful return result, and keep guardrails for opt-outs or shallow sessions. Review only after a full return window.

A one-page benchmark worksheet

1. Write the benchmark source, collection year, genre, region, platform and return definition.

2. Choose one complete install cohort and record its size.

3. Calculate your Day 1, Day 7 and Day 28 rates with the same return action.

4. Mark which outside figures are genuinely comparable.

5. Find the earliest meaningful gap, then inspect player behavior at that point.

6. State one test and the result that would justify keeping it.

A benchmark is a prompt for diagnosis. Your own cohort is the test of whether a change helped.

Related guide: Mobile game user retention strategies shows how to connect a weak return point to a small product test.

If your team needs help checking the data and choosing that test, tell Interactive Catalyst where players leave.