Retention rate formula
Retention is always measured for a cohort: the users who installed (or first opened) the app on the same day.
Day-N retention = users from the cohort who are active on day N ÷ users in the cohort × 100
Day 0 is the install day. "Active" usually means opening the app, but you can define it more strictly, for example completing a workout or sending a message. Whatever you choose, keep it the same over time, or the numbers stop being comparable.
Day 1, day 7 and day 30 retention
| Metric | What it measures | What it tells you |
|---|---|---|
| Day 1 | Share of a cohort that comes back the day after install | Whether onboarding and the first session delivered value |
| Day 7 | Share active one week after install | Whether the app became part of a weekly routine |
| Day 30 | Share active one month after install | Whether you have a lasting product; drives lifetime value |
Day 1 is mostly about the first impression. Day 30 is about the product itself: a great onboarding can lift day 1, but only real, repeated value holds day 30.
Worked example: calculating 30-day retention
A meditation app gets 2,000 installs on 1 March. That day's cohort comes back as follows:
| Day | Active users from the 1 March cohort | Retention |
|---|---|---|
| Day 0 | 2,000 | 100% |
| Day 1 | 760 | 38% |
| Day 7 | 420 | 21% |
| Day 30 | 180 | 9% |
30-day retention = 180 ÷ 2,000 × 100 = 9%. In practice you average the same metric over many daily cohorts (for example, all installs in March) so one unusual day does not distort the picture.
Classic, rolling and range retention
Analytics tools use three definitions, and they give very different numbers for the same app:
- Classic (day-N) retention: active exactly on day N. The strictest version and the one most benchmarks use.
- Rolling (unbounded) retention: active on day N or any day after it. Always higher than classic retention; useful for apps people open irregularly.
- Range (bracket) retention: active at least once within a window, for example days 7–13. A good middle ground for weekly-use apps.
When you compare your numbers with someone else's, check which definition they used first.
How to read a retention curve
Plot retention for every day from 0 to 30 or 90 and you get a curve that drops steeply at first. The shape matters more than any single point:
- The curve flattens: a core group of users stays. That plateau is your real audience.
- The curve keeps falling towards zero: users try the app but nobody sticks. Fix the product before you buy more installs.
- Newer cohorts sit above older ones: your changes are working.
Average app retention rate (benchmarks)
Adjust’s global benchmarks put the average app retention at roughly:
| Day | Average retention |
|---|---|
| Day 1 | 26% |
| Day 7 | 13% |
| Day 30 | 7% |
By category, day-1 retention ranges from about 20% (e-commerce) to 29% (games), and day-30 retention from about 5% (travel) to 9% (fintech). Benchmarks differ between sources and years, so treat them as a rough scale, not a target.
What is a good retention rate?
It depends heavily on the category. Games and social apps live on daily use; a tax or travel app may be opened a few times a year and still be a success. Industry reports are useful for a rough sense of scale, but the more reliable test is your own cohorts over time: is this month's day-30 retention better than last month's?
How to improve app retention
- Shorten the path to the first success. The first session should end with the user having done the thing they installed the app for.
- Ask for permissions in context, not all at launch. A notification prompt makes more sense after the user has a reason to be notified.
- Use push notifications sparingly and only with something useful to say.
- Fix what drives people away. App reviews often name the exact reason users leave: a missing feature, a confusing paywall, a crash on a specific device.
- Measure by cohort after every release, so you can see which change moved the curve.
Retention vs. churn
Churn rate is the opposite view: the share of users who stopped using the app in a period. For a single cohort and period, churn ≈ 100% − retention. Daily engagement among the users you keep is measured with DAU/MAU and session length.