Free Trial Length and Conversion: 3, 7 or 14 Days?

Longer trials convert better in every public dataset, with 17–32 day trials reaching a 42.5% median trial-to-paid rate against 25.5% for trials of four days or fewer in 2026. Most apps still choose short trials, because conversion is only one side of a trade that also involves refunds, cost to serve and how fast you learn.

The benchmarks by trial length

RevenueCat's 2026 report, built on more than 115,000 apps, groups trials into four buckets. Trials of four days or fewer converted at a 25.5% median, 5–9 day trials at 37.4%, and 17–32 day trials at 42.5% ¹. The 2025 edition had the long bucket even higher at 45.7% ². Adapty's 2026 report, from a different sample of about 16,000 apps, puts the global average trial-to-paid at 25.6%, with Health & Fitness highest at 35.0% and Entertainment lowest at 19.1% ⁵.

Plan type changes the picture. RevenueCat's September 2026 analysis splits conversion by plan: weekly plans convert at 22.3% with trials of four days or fewer and 24.3% with 5–9 day trials; monthly plans go from 39.6% to 45.9% across the same buckets; annual plans climb from 24% at four days or fewer to 44.6% at 17–32 days ³. First renewal moves the same way: for annual plans it rises from 18.3% to 47.5% between the shortest and longest buckets ³.

Two cautions. The buckets compare different apps, not one app with different trial lengths, and RevenueCat notes the correlation is partly survivorship: a user still inside a long trial is more likely to pay because they have not left yet ³. And the medians are global. Travel converted at 43.5% and Photo & Video at 22.2% in 2026, and North America at 34.2% versus 15.2% in India and South-East Asia ¹.

Why most apps still pick 3 days

Despite the numbers, 46.5% of apps ran trials of four days or fewer in 2026, up from 42.1% ¹. Among the 100 apps with the most trial starts, 81 to 100 of the weekly-plan apps in each category used trials of four days or fewer ³. The reasons are practical. A 3-day trial produces a conversion signal within a week, which matters when you are buying installs and need a return-on-ad-spend read, and it limits the cost of serving features such as AI inference to people who never pay ³.

The cost is a cancellation cliff. For 3-day trials, 55.4% of all cancellations happen on day 0 and 84% by the end of day 1; the day-0 share falls to 39.8% for 7-day, 35.7% for 14-day and 31.1% for 30-day trials ^1,4. After day 2, apps with 14-day and 30-day trials see daily churn fall below 10% of cancellations ¹². Superwall attributes the instant cancel to a fear of forgetting, which is easy to act on because cancelling on iOS takes a few taps ⁹.

Trial-to-paid is not the metric that pays the bills

Trial-to-paid flatters apps that make trials hard to start. The 2025 report shows median trial start rates of 6.2% of installs, with the top 10% of apps at 20.3% ². Multiply through: a 40% trial conversion on a 6% start rate is a 2.4% install-to-paid funnel, while 25% on 20% is 5%. The fairer comparison is download-to-paid at day 35, where the 2026 global median is 2.0% ¹. Adapty's numbers tell the same story from the other end: an average install-to-trial rate of 10.9% and trial-to-paid of 25.6% ⁵.

Retention matters more than either. Adapty found trial subscribers retain 1.4–1.7 times better than people who bought without a trial ⁵, and RevenueCat reports first-renewal rates of 35.9% for weekly, 49.5% for monthly and 26.6% for annual plans sold without a trial, generally below trial-led cohorts of the same plan ³.

Introductory offers beyond the free trial

Apple supports three introductory offer types: a free trial, pay-as-you-go (a discounted price for several periods) and pay-up-front (one discounted payment for a set duration). Free trials can be 3 days, 1 or 2 weeks, 1, 2, 3 or 6 months, or 1 year; pay-as-you-go runs from one to twelve periods; each user can redeem one introductory offer per subscription group, and you can have one current and one future offer per storefront ⁶. Google Play's free-trial phases run from 3 days to 3 years, introductory prices must be at or below the base-plan price, and discounts can be spread over 1 to 52 billing periods ⁸.

Usage is lopsided across stores: introductory offers appeared in 13.5% of App Store transactions but 5.9% on Google Play in 2025 ². A paid intro price can outperform a trial when the audience is price-sensitive rather than trust-sensitive. One July 2025 case on the RevenueCat blog replaced a 3-day trial with steep weekly and monthly intro discounts: install-to-pay rose from about 3% to around 12%, SKAN cost per acquisition dropped from $105 to $17, and 30-day LTV fell from $55 to $10, so the app traded value per user for volume ¹⁰. The author's own warning is that discounting always lowers LTV, and that offers are not helpful for every business ¹⁰. Remember that free-trial days do not count towards Apple's one-year threshold for the 15% commission, while paid intro periods do ¹³.

Win-back offers for lapsed subscribers

Winning people back is harder than the dashboards suggest. In the 2026 data, only 5% of lapsed annual subscribers returned within a year, against 20% of monthly subscribers; Productivity apps reached 36.1% monthly reactivation ^4,12. Month one accounts for 35% of annual cancellations, with a second spike of 9–14% in month twelve ⁴.

Apple's win-back offers, available on iOS 18 and later, target people by how long they paid (1 to 24 months, or 3, 4 or 5 years) and how long ago they lapsed (1 to 24 months); they appear in the App Store, in the app via a system sheet, in subscription settings and through a direct link, with up to five active offers per storefront per subscription ⁷. Google Play has no dedicated win-back type; the documented route is a developer-determined offer, for example 50% off for three months, shown to users your app identifies as lapsed ⁸. RevenueCat's playbook adds non-price tactics: show what the user will lose before expiry, point them at unused premium features, send a recap of what they made while subscribed, and hold discounts until the typical renewal cycle has passed so you do not train current subscribers to cancel for a deal ¹¹.

How to measure trial performance

Measure by install-week cohort, not calendar month, and let every cohort age past the trial and the first billing retry window before comparing. A trial-length A/B test must run until both arms reach first renewal, otherwise the shorter trial always looks better because its revenue arrives sooner. Superwall's fixes for the day-0 cliff are a reminder before billing that names the date and price, surfacing core value during onboarding, and reacting to early cancellations with an email or in-app offer while the trial is still active ⁹.

How this shows up in the charts

Trial length sets the lag between a download spike and a revenue spike. An app with a 3-day trial that climbs the iOS Top Free chart will show up on Top Grossing within a week if conversion holds; a 14-day trial pushes that by two weeks and a 30-day trial by a month, which is why some apps appear on the movers list twice. Apps that lean on monthly plans with high reactivation stay on the grossing chart longer than annual-heavy apps whose lapsed users almost never return. On Google Play, where intro offers are used less than half as often, the gap between the Android free and grossing charts tends to be wider for the same app.

Summary

• Longer trials convert and renew better in every dataset; short trials win on speed, cost and refund exposure ^1,3.
• Half of 3-day-trial cancellations happen on day 0; a billing reminder and clear terms are the first fix ^1,9.
• Judge trials by download-to-paid and first renewal, not trial-to-paid alone ^1,2.
• Paid intro offers can replace trials for price-sensitive audiences, at a cost to LTV ¹⁰.
• Annual lapsers rarely come back; use Apple's win-back offers and Google's developer-determined offers early ^4,7,8.