Sources

RevenueCat — State of Subscription Apps 2025 (2025)

https://www.revenuecat.com/state-of-subscription-apps-2025

Sample. Aggregated across apps using RevenueCat's SDK; the report states figures as medians and percentiles across its customer base.

Caveat. RevenueCat sells subscription infrastructure, so the sample is apps that chose that tooling — better-resourced than the App Store as a whole. Medians and percentiles are reported; exact app counts per figure are not.

• Download → trial start:4.3%(median, low-priced apps)
• Download → trial start:9.8%(median, high-priced apps)
• Trials that start on install day:82%(share of all trial starts)
• Trial → paid:48.7%(median, Travel — the highest category)
• Trial → paid:43.8%(median, Media & Entertainment)
• Trial → paid by trial length:45.7%(median for trials of 17–32 days — the best-converting length)
• Share of trials offered for 5–9 days:52%(in 2024, up from 48.5% in 2023)
• Download → paid at day 35:12.1%(median, hard paywall)
• Download → paid at day 35:2.2%(median, freemium)
• Download → paid at day 35:2.66%(median, high-priced apps)
• Download → paid at day 35:1.49%(median, low-priced apps)
• Refund rate:5.8%(hard-paywall apps)
• Refund rate:3.4%(freemium apps)
• Refund rate:4.71%(Health & Fitness)
• Refund rate:4.86%(Education)
• Annual subscriptions cancelled in the first month:~30%
• Renewal rate at year 1:3.4%(median, weekly plans)
• Renewal rate at year 1:17.0%(median, monthly plans)
• Still subscribed after one year:up to 36.0%(apps with cheap annual plans)
• Still subscribed after one year:6.7%(high-priced monthly plans)
• Year-1 retention, annual plans:53.7%(low-priced annual)
• Year-1 retention, annual plans:48.3%(high-priced annual)
• Year-1 retention, monthly plans:22.5%(low-priced monthly)
• Year-1 retention, monthly plans:12.2%(high-priced monthly)
• Realised year-1 LTV per new subscriber:$55.21(median, high-priced apps)
• Realised year-1 LTV per new subscriber:$8.08(median, low-priced apps — roughly a seventh)
• Share of subscriptions that are annual:67%(Health & Fitness)
• Share of subscriptions that are annual:65%(Travel)
• Share of subscriptions that are weekly:78%(Gaming — the outlier)
• Reliance on weekly plans:36% · 35%(India/South-East Asia · Middle East & Africa — highest regions)
• Median monthly price:$29.99(Business category)
• Median prices year over year:flat(P75 and P90 prices rose slightly across most durations)

Unverified attribution (retired) — Mobile onboarding benchmarks (2026)

https://uxcam.com/blog/

Sample. Unverified; no traceable primary study located.

Caveat. Retired September 2026. Do not quote these onboarding numbers or treat them as evidence.

• Onboarding benchmark:Not verified(Retired: no traceable primary study. Do not use as evidence.)
• Onboarding benchmark:Not verified(Retired: no traceable primary study. Do not use as evidence.)
• Onboarding benchmark:Not verified(Retired: no traceable primary study. Do not use as evidence.)
• Onboarding benchmark:Not verified(Retired: no traceable primary study. Do not use as evidence.)

Superwall — Multi-page onboarding paywalls convert 37% better than single-page (2026)

https://superwall.com/blog/new-postmulti-page-onboarding-paywalls-convert-37-better-than-single-page-heres-why

Sample. Just over 40 million onboarding-placement paywall opens, February to May 2026; paywalls with fewer than 50 opens or zero transactions excluded.

Caveat. Observational, not randomised: apps that build multi-page paywalls are self-selected and likely better at everything else too. Superwall sells paywall testing. The metric is trials or purchases divided by paywall views at onboarding placements.

• Onboarding paywall conversion, multi-page vs single-page:12.41% vs 9.07%(trials or purchases ÷ paywall views; +37% relative)
• Share of onboarding paywall opens that are multi-page:24%

Superwall — How many products should you offer on your paywall (2022)

https://superwall.com/blog/how-many-products-should-you-offer-on-your-paywall

Sample. 15 apps, each with more than 10,000 paywall views and at least two tests with different product counts; ~32M interactions in total.

Caveat. Fifteen apps is a small sample for a distribution claim, the page's own headline statistics do not reconcile arithmetically, and it is four years old. Directionally consistent with practitioner experience; do not treat the percentages as precise.

• Conversion, two products vs one:+61%(relative; 15 apps)
• Conversion, three products vs two:+44%(relative; 15 apps)

Superwall — Transaction-abandon paywalls: an 18-company case study (2024)

https://superwall.com/blog/17-revenue-boost-with-transaction-abandon-paywalls-a-case-study

Sample. 18 companies, each running the pattern for at least two weeks.

Caveat. Not an A/B test: the 'control' and 'variant' are different populations (all installs versus users who abandoned a purchase). The 17% is a share of revenue attributed to the pattern, not measured incremental lift.

• Revenue attributed to transaction-abandon paywalls:17%(share, not lift)
• Paywalled users who abandon a started purchase:~20%
• Revenue attributed to transaction-abandon paywalls:17%(share of revenue, not measured incremental lift; 18 companies)
• Refund rate, abandon-offer group vs control:3.3% vs 6.8%(different populations, not an A/B test)
• Share of paywalled users who abandon a started purchase:~20%

Adjust — ATT opt-in rates 2025 (2025)

https://www.adjust.com/blog/att-opt-in-rates-2025/

Sample. Adjust's measured apps, Q2 2025; the rate is defined as opt-in among users who were shown the prompt.

Caveat. Vendor of attribution; sample size not stated. The same metric with an 'all devices' denominator (Singular, 2024) comes out around 14%, so the denominator is the whole story with ATT figures.

• ATT opt-in among users shown the prompt:35%(Q2 2025 industry average; 34% in 2023, 34.5% in 2024)

Apple — Human Interface Guidelines — Onboarding; App Review Guidelines §3.1.2 (Subscriptions) (2026)

https://developer.apple.com/design/human-interface-guidelines/onboarding

Sample. Not a study. Apple's published design guidance and the review rules a paywall must satisfy.

RevenueCat — State of Subscription Apps 2026 (2026)

https://www.revenuecat.com/state-of-subscription-apps-2026-business/

Sample. Over 115,000 apps, more than $16 billion in revenue and more than a billion transactions, across iOS, Android and web; figures are 2025 data reported as medians and quartiles. Apps below a minimum install or revenue threshold are excluded.

Caveat. RevenueCat sells subscription infrastructure, so the sample is apps that chose that tooling. Every comparison is observational: 'longer trials convert better' compares apps that chose long trials with apps that chose short ones, and RevenueCat's own analysts say on the record that this is probably correlation. Per-figure app counts are not published.

• Download → paid at day 35:10.7%(median, hard-paywall apps; was 12.1% in 2025)
• Download → paid at day 35:2.1%(median, freemium apps; unchanged from 2025)
• Revenue per install at day 60:$3.09(median, hard-paywall apps)
• Revenue per install at day 60:$0.38(median, freemium apps)
• Yearly subscribers still subscribed after one year:27%(median, hard-paywall apps (26.8%; P90 54%))
• Yearly subscribers still subscribed after one year:28%(median, freemium apps (27.7%; P90 58%))
• Freemium conversions occurring 6+ weeks after download:23%
• Trial → paid:42.5%(median, trials of 17–32 days; top quartile above 59.4%)
• Trial → paid:37.4%(median, trials of 5–9 days; top quartile above 52.8%)
• Trial → paid:25.5%(median, trials of 4 days or less; top quartile above 38.5%)
• Share of 3-day-trial cancellations occurring on day 0:55.4%(~51% the previous year)
• Share of 3-day-trial cancellations occurring by day 1:84%
• Share of 30-day-trial cancellations occurring on day 0:31.1%(7-day trials 39.8%; 14-day 35.7%)
• Share of apps showing exactly two plans on the paywall:41–60%(range across categories; Health & Fitness highest at 60%)
• Revenue per install at day 14, by the app's most-sold plan duration:$0.36 vs $0.18(yearly-led apps vs monthly-led apps; weekly-led $0.19)
• Share of subscriptions that are annual:68%(Health & Fitness; Gaming is 82% weekly and Productivity 77% monthly)
• Median monthly subscription price:$9.99(North America — the highest region)
• Median monthly subscription price:$3.75(India and South-East Asia — the lowest region)
• IN/SEA prices as a share of top-tier market prices:46–54%(across all plan durations)
• Share of all annual cancellations occurring in month 1:35%
• Share of App Store subscription cancellations caused by billing failure:14%(15.1% the previous year)
• Share of Google Play subscription cancellations caused by billing failure:31%(28.2% the previous year)

RevenueCat — The State of Subscription Apps in 10 minutes: benchmarks for 2026 (2026)

https://www.revenuecat.com/blog/growth/subscription-app-trends-benchmarks-2026

Sample. Same dataset as rc26; this page is where the year-over-year deltas are stated.

• Share of apps offering trials of 4 days or less:46.5%(up from 42.1% the previous year)
• Share of annual subscriptions cancelled within year one:~72%(~56% the previous year; first-year annual retention fell 31% → 28%)

RevenueCat — How to build a reactivation strategy that actually works (2026)

https://www.revenuecat.com/blog/growth/app-reactivation-strategy-how-to

Sample. Practitioner essay quoting State of Subscription Apps 2026 for the reactivation rates.

Caveat. The reactivation rates are from the report; the recommended timing windows (7/21/45 days for monthly churners, 30/90/180 for annual) are the author's experience and are not measured anywhere in the report.

• Churned monthly subscribers reactivating within 12 months:18–24%(20.1% overall, up from 13.7% in 2025; Productivity highest at 36.1%)
• Churned weekly subscribers reactivating within 12 months:7–10%
• Churned annual subscribers reactivating within 12 months:4–6%(4.9–5.9% across every geography and 4.4–5.6% across every price tier — a structural ceiling, not a tuning problem)

Adapty — State of In-App Subscriptions 2026 (2026)

https://adapty.io/state-of-in-app-subscriptions/

Sample. 16,000+ apps, $3 billion in subscription revenue, 500 million transaction events and 105,000 paywalls; primarily App Store data, 2024–2025.

Caveat. Adapty sells paywall A/B testing and remote paywall building, and the report's central recommendation is to run more experiments. The experiment win rates are computed over tests Adapty customers chose to run, so test types that teams attempt only when they already suspect a win will look better. The '18.7x revenue for apps with 50+ experiments' figure is selection, not causation — large apps run more tests — and is not used as evidence in any rule.

• Lifetime value per subscriber, hard vs soft paywall:+21%(hard paywalls higher — directionally opposed to rc26 on conversion)
• Conversion, soft vs hard paywall:~+50%(soft paywalls higher)
• Share of all trial starts occurring on install day:89.4%(Entertainment 94.5%, Photo & Video 91.2%, Education the outlier at 71%)
• Share of all subscription app revenue from weekly plans:55.5%(up from 43.3% two years earlier)
• Share of new subscription apps that never reach $1,000 in total revenue:57.7%

Adapty — State of in-app subscriptions 2026: key findings on app monetization (2026)

https://adapty.io/blog/mobile-app-monetization-2026/

Sample. Same dataset as adapty26; this page carries the experiment win-rate table and the LTV curves.

• Lifetime value of a weekly-plus-trial cohort, day 0 → day 380:$7.40 → $54.50(+636%)
• Lifetime value of an annual-plus-trial cohort, day 0 → day 380:$42.08 → $49.92(+18.6%)
• Share of paywall experiments improving lifetime value, localisation tests:62.3%(highest win rate of any test type)
• Share of paywall experiments improving lifetime value, trial-structure tests:59.6%
• Share of paywall experiments improving lifetime value, plan-duration tests:58.7%
• Share of paywall experiments improving lifetime value, plan-count tests:57.1%
• Share of paywall experiments improving lifetime value, price tests:45.5%(and only 28.3% improve conversion)
• Share of paywall experiments improving lifetime value, visual or copy-only tests:34.6%(lowest win rate of any test type)

Superwall — Cutting GRAVL's onboarding paywall from three plans to two (2026)

https://www.linkedin.com/posts/superwall_activity-7460335261592240128-JYAD

Sample. One app, one A/B test at the onboarding placement, plus the same test run at the post-onboarding placement.

Caveat. A single app, posted by the vendor on LinkedIn, with no sample size or confidence interval. It is cited only for the qualitative finding — the same change helped at one placement and not at another — never as a size of effect to expect.

• Onboarding paywall cut from three plans to two:+36% conversion, +58% ARPU, +11pp yearly share(one app; the identical test at the post-onboarding placement did not win)
• Conversion lift from referencing the user's onboarding answer on the paywall:10–15%(Superwall bootcamp tests; no sample, no method published — treat as an anecdote)

Superwall — Mojo: removing AI credit packs from the paywall (2026)

https://superwall.com/case-studies/mojo

Sample. One app, one A/B test, results clear within a week.

Caveat. Single-app case study published by the vendor whose tool ran the test.

• New subscription revenue after removing credit packs from the paywall:+14%(one app, one week to significance)

Apple — Auto-renewable subscriptions — offers, billing grace period, Family Sharing, pricing (2026)

https://developer.apple.com/app-store/subscriptions/

Sample. Not a study. Platform behaviour and configuration limits as Apple documents them.

• Introductory offers redeemable per subscription group:one, per customer, for life(types: free trial, pay-as-you-go, pay-up-front)
• Promotional offers configurable per subscription:up to 10
• Billing grace period durations offered by Apple:3, 16 or 28 days(enabled in App Store Connect; subscriber keeps access while Apple retries)
• Period Apple keeps retrying a failed subscription renewal:60 days(since iOS 16.4 the system shows a payment-update sheet in the app on launch)
• Family members who can share one auto-renewable subscription:up to 5(enabling Family Sharing on a subscription cannot be undone; check transaction ownershipType)
• Subscription price points available per currency:800(plus 100 higher price points on request; one scheduled future price change at a time per territory)
• Free-trial durations Apple allows:3 days; 1 or 2 weeks; 1, 2, 3 or 6 months; 1 year(so the practical choice is 3, 7, 14 or 30 days)
• App Store storefronts and currencies prices are equalized across:175 storefronts, 44 currencies(auto-adjustments for exchange rate and tax are shown at least 14 days ahead; a manually priced storefront stops auto-adjusting)

Apple — Retention Messaging API; WWDC26 session 309; App Store Retention Messaging report (2026)

https://developer.apple.com/documentation/retentionmessaging

Sample. Not a study. The cancel-sheet surface, its configuration paths, and the analytics Apple reports for it.

• Server response budget for real-time Retention Messaging:700 ms(in production; a configured default message per product and locale is the required fallback)

Apple — Reducing involuntary subscriber churn; Tech Talk 111386 (Reduce involuntary subscriber churn) (2026)

https://developer.apple.com/documentation/storekit/reducing-involuntary-subscriber-churn

Sample. Apple's own platform-wide data, stated in a developer Tech Talk without a date range or denominator.

Caveat. Apple has an interest in developers enabling grace periods. The 40/75/90% recovery curve and the 'nearly 80 million subscriptions' figure are stated without a period, a sample size or a comparison group.

• Customers who correct their billing details, by days elapsed:40% by day 3, 75% by day 16, 90% by day 28(Apple's stated justification for the three grace-period lengths it offers)
• Involuntarily churned subscriptions Apple recovered in a year for developers with billing grace period enabled:nearly 80 million
• Customers who cancel a subscription with at least two days of paid service remaining:over 90%

Apple — Set up win-back offers; Supporting win-back offers in your app (2026)

https://developer.apple.com/help/app-store-connect/manage-subscriptions/set-up-win-back-offers

Sample. Not a study. App Store Connect limits and StoreKit eligibility rules.

• Win-back offer limits:350 per subscription, 5 active per storefront, 3-day minimum run(eligible = expired and auto-renew off; anyone in grace or billing retry is excluded; in-app sheet needs iOS 18)

Apple — Offer in-app events (2026)

https://developer.apple.com/help/app-store-connect/offer-in-app-events/offer-in-app-events/

Sample. Not a study. App Store Connect limits.

• In-app events limits:10 published at once, 15 approved held, 15 minutes to 31 days(discoverable up to 14 days before the start; 10 may have overlapping start times)

Duolingo — Improving the streak; How streaks keep Duolingo learners committed (2020)

https://blog.duolingo.com/improving-the-streak/

Sample. Duolingo's own A/B tests on its own learner base; results reported as percentage changes with no sample sizes and no confidence intervals.

Caveat. Real randomised experiments, but self-reported by a company whose product is the streak, and unauditable without n. Duolingo has never attributed a quantified share of its DAU growth to the streak in any shareholder letter; claims of that shape are over-readings.

• Day-14 retention change from decoupling the streak from the daily goal:+3.3%(same test: DAU +1%, share of daily learners on a streak +10.5% overall and +19% for new learners, over 20 days; no n or interval published)
• Day-7 retention change from the Streak Wager:+14%(largest of day 1 / 7 / 14, all reported statistically significant; no n published)

Airship — 2025 Push Notification Benchmarks; How push notifications impact mobile app retention rates (2025)

https://growth.airship.com/rs/313-QPJ-195/images/Airship-2025-Push-Notification-Benchmarks-EN.pdf

Sample. Benchmarks: apps with at least 1,000 active users and 1,000 cumulative pushes a month, 13 verticals, January–December 2024. The retention paper: 63 million users across 1,500 apps, install cohort September 2016 tracked to December 2016, published 2019.

Caveat. A push vendor measuring apps that bought a push vendor. The '3x retention' headline is a decade-old cohort and is contradicted by Airship's own segment tables, where opted-out users retain better than opted-in users receiving nothing. Airship and Batch report materially different iOS opt-in rates for overlapping periods.

• iOS push notification opt-in rate:49.4% median(90th percentile 74.1%, 10th 27.1%; flat year on year. A second push vendor (Batch) puts iOS at 56% for an overlapping period — the two disagree materially)
• Retention of users receiving any push vs none:+190% — do not use(September 2016 install cohort tracked to December 2016. In the same vendor's segment data, opted-OUT users retain 36–77% better than opted-in users sent nothing, which is the selection effect the headline hides)

Bidargaddi et al., JMIR mHealth and uHealth — To prompt or not to prompt? A microrandomized trial of time-varying push notifications (2018)

https://mhealth.jmir.org/2018/11/e10123/

Sample. 1,255 users of one workplace well-being app over 89 days, micro-randomised.

Caveat. Peer-reviewed, randomised and disinterested — and one app in one domain. Treat +3.9% as the honest ceiling for a single notification, not as a benchmark for a whole programme.

• Likelihood of engaging within 24 hours when sent a tailored push, randomised:+3.9%(RR 1.039, 95% CI 1.01–1.08; 1,255 users, one app, 89 days; weekends +8.7%, weekdays not significant)

Dropbox; Airbnb — Dropbox Startup Lessons Learned (Houston, 2010); Hacking word of mouth: making referrals work for Airbnb (2014) (2014)

https://medium.com/airbnb-engineering/hacking-word-of-mouth-making-referrals-work-for-airbnb-46468e7790a6

Sample. Each company's own reporting of its own referral programme at the time.

Caveat. Primary but old, and both are web products rather than subscription apps. The famous 'Dropbox grew 3900% in 15 months' is third-party arithmetic on a 40x slide, not a Dropbox claim, and Houston never attributed that growth wholly to referrals.

• Share of daily signups from the referral programme, and its lift:35%, +60% permanent lift(plus 20% from shared-folder virality; 2.8M direct invites in a trailing 30 days; web product, 2010)
• Share of bookings attributable to referrals:over 25% in some markets(referral signups and bookings rose over 300% per day after the relaunch — that is 300% of the programme's own output, not of total bookings; marketplace, 2014)

Google / Android Developers Blog — Gratitude saw 25% higher retention for widget users (2026)

https://developer.android.com/blog/posts/gratitude-saw-25-higher-retention-for-widget-users

Sample. One Android app with 6M+ downloads, self-reported, no control group.

Caveat. Android, not iOS; Google promoting its own framework; and widget adopters are self-selected as an app's most engaged users, so the comparison is between two very different populations. Cited to show that nothing better exists.

• Retention of widget users vs non-widget users:+25%(one Android app, self-reported, not randomised; widget adoption 10% of DAU. No equivalent iOS figure exists from anyone)

AppTweak — ASO trends & benchmarks report 2025 (2025)

https://www.apptweak.com/en/aso-blog/aso-app-store-trends-benchmarks-report

Sample. The top 1,000 apps and the top 1,000 games in the US, across the App Store and Google Play, January–December 2024, compared with 2023.

Caveat. AppTweak sells ASO tooling. The ratings figure is a correlation across apps — apps that moved from 3.6 to 4.2 stars also usually improved other things at the same time — not a measured effect of a rating change.

• Conversion rate of apps whose rating improved from 3.6 to 4.2 stars:nearly 60% higher(correlation across the top 1,000 US apps, not a controlled rating change)
• Share of App Store featured apps rated 4.0 or higher:90%
• Share of top apps using custom product pages:31% of apps, 26% of games
• Conversion lift from custom product pages in Apple Ads:+5.9% apps, +3.5% games, up to +8.6%(highest lift on generic-keyword campaigns)
• Share of top App Store apps that updated screenshots at least twice in a year:49%(top apps refresh 2–4 times a year)

Shotlingo — App Store screenshot conversion rate benchmarks 2026 (2026)

https://shotlingo.com/blog/app-store-screenshot-conversion-benchmarks-2026/

Sample. 200 indie apps that opted in to share App Store Connect analytics; each with at least 500 product page views a month; English (US) storefront; January–May 2026; CVR = first-time downloads ÷ product page views.

Caveat. Small, self-selected, and published by a screenshot-generation vendor. Apps with Apple Search Ads running were included. The method is at least stated, which most ASO figures are not; read the screenshot lift as the top of a plausible range.

• Median App Store conversion rate, benefit-headline first screenshot vs bare device render:39.8% vs 31.3%(+27% relative; 200 opt-in indie apps; overall median 31.2%)

Apple — Custom product pages; Product page optimization (2026)

https://developer.apple.com/app-store/custom-product-pages/

Sample. Not a study. App Store Connect limits.

• Custom product pages per app:70(each with its own URL and its own conversion and proceeds reporting in App Analytics)
• Product Page Optimization treatments per test, and test duration:3 treatments, 90 days(icons, screenshots and app previews; results reported with a confidence level)