A Day 30 retention rate of 4 per cent sounds like failure. It is the median. In MWM's April 2026 benchmark, roughly 96 in every 100 installs are not opening the app a month later, and that is a normal app, not a broken one.
So the useful question is not "is my Day 30 high?" It is "high compared to what?" Your category, your platform mix, the retention method behind the number, and your own Day 7 all change the answer. Our post on what a good Day 7 retention rate is covers the first week. This one picks up where it stops: the month mark, where the users who formed a habit separate from the ones who were only curious.
What is Day 30 retention rate?
Day 30 retention rate is the percentage of an install cohort that opens the app on the 30th day after install. That is the classic retention definition, and it is how nearly every public benchmark is built.
Day 30 retention = (cohort installs that opened the app on Day 30) / (cohort size) × 100
Three things change the number before your product does:
- Method. Classic counts exactly Day 30. Cumulative counts any open from Day 1 to Day 30, so it reads far higher. Rolling retention counts Day 30 or later and keeps climbing as the cohort ages. The full comparison is in cumulative vs classic vs rolling retention.
- Cohort maturity. An install cannot have a Day 30 until it is 30 days old. Any cohort from the last month is incomplete, so its D30 column reads low or empty.
- "30-day retention" is ambiguous. Some tools use it to mean "active at any point in the first 30 days", which is cumulative. Check the definition before you compare.
Rule: benchmark classic against classic. Comparing a cumulative or rolling D30 with a published classic figure flatters you every time.
Day 30 benchmarks: the overall picture
MWM's April 2026 medians across its app catalogue give the clearest single reference curve:
| Day | Median app | Top 10% of apps |
|---|---|---|
| Day 1 | around 27% | not published |
| Day 7 | around 9% | not published |
| Day 30 | around 4% | around 11% |
Source: MWM, Retention benchmarks ("State of April 2026" medians).
What this tells you:
- A good Day 30 retention rate is one at or above the median for your category, platform and method. Around 4 per cent is the cross-category median, not a target. Clearing 10 per cent puts an app in roughly the top tenth.
- iOS retains better than Android. Business of Apps' 2026 averages, as cited by Phiture, put Day 30 at around 4 per cent on iOS and around 2 on Android. An Android-heavy app (most Indian apps are) should expect to sit below a blended figure.
- Region matters less than category. MWM's country medians for Q3 2025 sit at around 4 per cent on Day 30 in every major market it tracks, India included, while its category medians range from about 3 to 6.
Day 30 retention by app category
Category is where the real bar sits. This table uses AppsFlyer data as compiled by Business of Apps and published by Braze in 2026, rounded to whole numbers:
| Category | iOS Day 30 | Android Day 30 |
|---|---|---|
| News | around 15% | around 10% |
| Finance | around 7% | around 3% |
| Transportation | around 5% | around 4% |
| Shopping | around 5% | around 4% |
| Health and fitness | around 4% | around 3% |
| Entertainment | around 4% | around 3% |
| Gaming | around 4% | around 2% |
| Travel | around 3% | around 3% |
| Social media | around 3% | around 2% |
| Education | around 3% | around 2% |
Two things stand out when you set this next to MWM's category medians (US, Q3 2025):
- Day 1 leaders are not Day 30 leaders. In MWM's data, games lead Day 1 at around 37 per cent, yet finish last of its six categories at Day 30, around 3. In the AppsFlyer cut, gaming also sits near the bottom by Day 30, while news and finance lead. Fast first sessions do not guarantee a habit. Recurring need (a daily news cycle, a salary credit, a bill) does.
- The two datasets disagree on ranking. Both put most categories between 2 and 6 per cent at Day 30, but MWM has social and communication apps on top at around 6, while the AppsFlyer table has social media near the bottom at 2 to 3. Different app samples, category definitions and methods produce different answers (the Braze table does not state its method). Pick one source per report and stay with it. Mixing them turns a methodology gap into a fake performance gap.
How much should you lose between Day 7 and Day 30?
On MWM's 2026 medians, Day 30 retention is roughly 40 per cent of Day 7 retention, and about a seventh of Day 1. The curve goes from around 9 per cent at Day 7 to around 4 at Day 30. Across its six categories the ratio runs from about 0.3 (games) to about 0.5 (lifestyle and productivity).
That makes the D30/D7 ratio a quick diagnostic you can run on your own cohorts, whatever your category:
- Under 0.3: users try the app for a week, then leave. Onboarding is not the problem; the reason to come back in weeks two to four is missing.
- Around 0.4 to 0.5: typical decay. Your levers are the same as for Day 7.
- 0.6 or above: the retention curve is flattening early. You have a sticky core, and the job is to get more users into it.
These bands are a rough rule of thumb around published medians, not a benchmark in themselves. The ratio is useful because it strips out category: a news app and a game can both be compared on how much of week one they keep. Where the curve finally flattens is covered in how to read a mobile retention curve.
Flip it round and the churn rate between Day 7 and Day 30 is the number to watch. Losing around three in five of your week-one users is normal. Losing three in four or more is a product signal, unless you are a game.
What actually moves Day 30
Day 1 is won in the first session. Day 30 is won in weeks two to four, with different levers:
- A recurring trigger. The categories that hold at Day 30 have a reason to return that the user did not choose: news, balances, bills, bookings. If your app lacks one, build it (a weekly summary, a streak, a renewal, new content on a fixed cadence).
- A second core action between Day 7 and Day 14. The first action proves value. The second one, such as a second order, a repeat deposit or a saved routine, is usually what separates a habit from a trial.
- Re-engagement inside the window, not after it. A message on Day 10 that opens the exact screen a user left is cheaper than winning them back on Day 45. Track it as its own campaign so it does not inflate organic retention, as set out in attribution for re-engagement campaigns.
- The channel mix you bought. Across retention audits, a blended number that looks weak usually turns out to be one broad or incentivised source dragging a healthy base. Those sources often look acceptable at Day 7 and collapse by Day 30, which is why the next section matters.
How to measure Day 30 retention by channel
Blended D30 hides which spend built your retained base. The full set-up is in how to measure app retention by channel; for Day 30 specifically:
- Use classic D30 per source, so each channel compares with the benchmarks above.
- Only read matured cohorts. Install dates at least 30 days old, and stamp the "as of" date on every figure.
- Pool weeks for small sources. A source with a few dozen installs a week swings wildly; group cohorts until each one has a few hundred.
- Compare the D30/D7 ratio per channel, not only the D30 level. Two sources at the same Day 7 can split sharply by Day 30.
- Separate reattributed users from first-time installs so a re-engaged cohort does not flatter a channel.
How to validate this in your MMP: check that Day 30 is reported per campaign, not only app-wide, and that the method is labelled. Linkrunner reports classic and cumulative retention at Day 1, 7, 14 and 30 per campaign, with rolling up to Day 7, as set out in the retention documentation. The same fields come through the Campaign Reporting API as classic_retention and retention, so no spreadsheet join of installs and opens is needed. Campaign density is where this pays off: Jumbo Gaming tracks more than $1M in revenue across 200+ campaigns on Linkrunner, a scale at which app-wide retention says little about which campaigns to keep.
Frequently asked questions
What is a good Day 30 retention rate? Around 4 per cent is the median in MWM's April 2026 data, and the top tenth of apps clear about 11. Category matters more: news apps often clear 10 per cent, while gaming, social and education apps commonly sit at 2 to 4 per cent in the AppsFlyer table.
Is 30-day retention the same as Day 30 retention? Not always. Day 30 retention usually means classic retention, an open on exactly Day 30. Some tools use "30-day retention" for any open within the first 30 days, which is cumulative and reads much higher.
Why is my Day 30 retention empty for recent cohorts? An install only has a Day 30 once it is 30 days old. Cohorts from the last month are incomplete, so wait for them to mature before you compare.
Should I compare iOS and Android Day 30 separately? Yes. Business of Apps' 2026 averages put iOS at around 4 per cent on Day 30 against about 2 on Android, and the category table shows the biggest gaps in finance and gaming. A blended figure moves whenever your platform mix moves.
Can I benchmark rolling retention at Day 30? Not against public benchmarks, which are classic. Rolling retention keeps rising as a cohort ages, so it has no fixed Day 30 value to compare.
Benchmark the month, then fix the weeks
A good Day 30 retention rate beats the classic median for your category and platform, and keeps at least 40 per cent of your Day 7 users. Match the method, wait for cohorts to mature, read the D30/D7 ratio as well as the level, and split it by channel before you decide whether you have a product problem or a spend problem.
If you want classic Day 30 retention per campaign next to installs and revenue, without stitching exports together, book a demo with Linkrunner. Then run the five-step channel check above on your last three matured cohorts.