App retention is the share of users who keep opening your app, or keep paying for it, a set number of days after they first install or subscribe.
To improve it, get new users to a real result in the first session, give them a reason to come back each day, and catch subscribers before and after they cancel.
In our test database, the retention gains came from onboarding built around the user's goals, honest trial screens, and well-timed offers to lapsed or about-to-leave subscribers.
This guide covers the metrics to watch, how to read benchmarks, and the five levers that move retention: onboarding, push and email, habit features, cancellation flows and win-back.
It closes with what the real tests in our app retention test library show.
Retention metrics: D1, D7, D30 and renewal rate
Subscription apps need two kinds of retention numbers. One tracks usage. The other tracks payment.
- D1, D7 and D30 retention: the percent of an install cohort that opens the app on day 1, day 7 and day 30. D1 shows whether the first session worked. D7 shows whether a habit is starting. D30 shows whether the app keeps delivering value.
- Trial-to-paid rate: the share of trial starters who become paying subscribers. A trial that converts poorly is a retention problem that shows up before the first charge.
- Renewal rate: the share of paying subscribers who renew at the end of each billing period. Track it per plan length, because weekly, monthly and annual plans behave very differently.
- Churn rate: the inverse of retention over a period. Split it into voluntary churn (the user cancels) and involuntary churn (a payment fails).
Always measure by cohort: users who installed or subscribed in the same week or month.
A blended monthly number hides whether newer users stay better or worse than older ones.
You can work out your numbers with our retention rate calculator and churn rate calculator, and see how they feed revenue in our guide to subscription metrics.
Retention benchmarks: what good looks like
Usage retention drops fast for almost every app.
Braze, citing AppsFlyer data via Business of Apps, says Day 30 retention runs from about 1% to 15% across categories, with news apps near the top and gaming and social apps near the bottom.
Paid retention depends heavily on plan length. RevenueCat's State of Subscription Apps 2025 reports year-one retention of 44.1% for annual plans, 17.0% for monthly plans and 3.4% for weekly plans.
The same report says the first month brings the biggest cancellation spike, often above 30% across categories.
An older RevenueCat analysis of more than 10,000 apps found a median first renewal rate of 56% for monthly plans and 27% for annual plans after their first year.
Compare yourself to your category and plan mix, not to one global number.
Our app retention rate benchmarks and subscription churn rate benchmarks collect more of these figures.
Onboarding's role in retention
Most retention is won or lost in the first session.
If a new user doesn't see the app solve their problem quickly, D1 retention falls and no later reminder brings them back.
Goal-based onboarding is the pattern with the clearest retention data in our database.
ASL Bloom, working with the agency Applica, rebuilt onboarding around the user's goals and paired it with a two-screen paywall.
The app reported +15% trial conversion, +12% ARPU and +18% 30-day retention.
Because the onboarding and paywall shipped together, the retention gain can't be credited to one change alone.
Applica's Five Minute Journal redesign, based on user research, reported +20% ARPU, about +10% trial starts and about +14% trial conversions.
More trial starters converting to paid is the first renewal you can measure.
Our app onboarding guide and the onboarding quiz tests cover the flow in detail.
Onboarding can also set expectations for later. Symmetry tested a Free vs Pro comparison table and saw +1% conversion and +5% paid user activation.
Users who understand what they paid for are more likely to use it, and users who use it are more likely to renew.
Push notifications and email
Push and email bring back users who meant to return but forgot.
They can't fix an app that doesn't deliver value, and too many messages push users to turn notifications off or uninstall.
A few rules keep messages useful:
- Ask for push permission after a win. On iOS and on Android 13 and later, users have to opt in. Asking after the user has seen value tends to work better than asking on the first screen. See our push opt-in rate benchmarks.
- Tie each message to the user's goal. A reminder about the habit the user chose in onboarding beats a generic "we miss you."
- Use email for things push can't carry. Weekly progress summaries, receipts, trial-ending reminders and win-back offers work well by email.
- Remind people before the trial ends. It feels like it will cost conversions, but it cuts refunds and angry reviews, and it builds trust that helps renewals.
In-app messages are a close relative. Lose It! moved its offers into in-app messages and reported about 10x more money from them.
That is a monetization result, but it shows how much timing and placement change response. Our push notifications guide covers the details.
Habit features: streaks, reminders and progress
Habit features give users a reason to open the app on days when no notification arrives. The common ones are:
- Streaks: a count of consecutive days of use. They work best when missing a day isn't a total loss, for example with a streak freeze or a grace day.
- Scheduled reminders: a time the user picks themselves, often set during onboarding.
- Progress views: charts, levels or milestones that show the user how far they've come.
- Content that changes: a daily lesson, workout or prompt so there is always something new.
Our database doesn't yet hold a clean A/B test of a streak feature, so we won't quote a number for one.
If you test a streak, measure D7 and D30 retention and renewal rate, not just daily opens.
A streak can raise opens while users feel guilt rather than value, and that shows up later as cancellations.
A trial is a short habit test too.
In one test, a wellness app's 14-day trial, run by Daphne Tideman, brought slightly more trial starts but fewer paid conversions than a 7-day trial.
Zumba saw +8% install-to-paid when it moved from 14 to 7 days.
A longer trial doesn't build a stronger habit by default. See our free trial tests for more.
Cancellation flows and win-back
Subscribers leave in two ways: they cancel, or their payment fails. Each needs its own fix.
Involuntary churn: failed payments
Turn on the store tools first, since they cost almost nothing.
Apple's Billing Grace Period keeps a subscriber's access for 3, 16 or 28 days (up to 6 days on weekly plans) while Apple keeps trying to collect payment.
On Google Play, a grace period of 3 to 30 days is followed by an account hold of up to 30 days.
Google says Play now tailors grace period length to each subscriber and offers an in-app messaging API that prompts users to fix a declined payment.
Voluntary churn: the cancel moment
On both stores, the cancel button lives in the store's subscription settings, not in your app. You can still shape the moment:
- Retention offers in the Play cancel flow: Google lets developers show a discount directly in the Play Store cancellation flow, or offer a cheaper plan instead.
- Cancellation reasons: Google Play shows a cancellation survey, and developers can read the answers. Users who leave over price are better targets for a discount than users who left for other reasons.
- Pause instead of cancel: Google Play lets subscribers pause for a set period, which suits seasonal apps.
- In-app "manage subscription" screens: a page you control can offer a downgrade, a pause or a discount before sending the user to the store.
Don't make cancelling hard. BoldVoice tested a discount offer that couldn't be closed and a hard paywall with no close button.
Both raised trial starts, and in both, refunds rose by more than the gain.
Win-back: lapsed subscribers
Apple's win-back offers give lapsed subscribers a free or discounted period.
You set how long the customer must have paid (1 to 24 months, or 3 to 5 years) and how long ago they lapsed (1 to 24 months).
Offers show on the Manage Subscriptions page, and in your app and on the App Store from iOS 18. Google Play has its own winback offers shown in the Play Store.
Expect a small pool. RevenueCat's 2025 report found that within 12 months, 9.37% of churned weekly subscribers and 4.58% of churned annual subscribers reactivated.
Even so, a good win-back offer is cheap to run.
The win-back tests in our database show the formats.
What the tests show
These are the retention and churn results in our database, with the numbers each team reported.
Most come from vendor case studies, so treat them as vendor-reported unless noted.
- Personal win-back offers work. Unscripted sent lapsed users a personal reactivation trial offer and saw over +240% trial starts (vendor-reported, Superwall case study).
- Offers before the trial ends can pay. Vahe Baghdasaryan's "end your trial now" discount offer averaged +15–17% ARPU across tests. It locks in the payment before the trial-end cancellation spike.
- A longer, cheap paid trial lifted later revenue. In health and fitness apps, a $1.99 paid 30-day trial raised D90 ARPU from $2.23 to $2.89 (+30%) versus monthly with no trial.
- Honest trial screens cut cancellations. Blinkist's trial timeline paywall, reported by Marcus Burke, showed users when they'd be charged. Trial opt-in rose significantly and trial cancellations fell 4%. Stardust's timeline test raised conversion 13.6% with trial-to-paid unchanged, so the extra trials didn't churn faster.
- A guarantee can beat a trial on lifetime value. Teaching.com replaced its trial with a money-back guarantee: conversion dropped slightly, but revenue and LTV rose.
- Some churn is worth accepting. When TalkingParents raised its price alongside a new feature, about a quarter of customers churned, but the business came out better financially overall (before-after, not an A/B test).
- Forcing users in backfires. BoldVoice's unclosable screens raised trial starts but raised refunds by more.
The pattern across these tests: retention gains come from honesty and timing, not pressure.
Show people what they get and when they'll pay, reach them when they're about to leave or have just left, and judge every test on revenue over 30 to 90 days, not on the first conversion.
To see what each extra month of retention is worth, run your numbers through the subscription LTV calculator.