---
title: "What State of Paywalls taught me about the tests subscription apps skip"
description: "Paywall optimization data from 78,000+ paywalls: which A/B tests win, why redesigns lose, and five tests most subscription apps rarely run."
url: "https://adapty.io/blog/paywall-optimization-tests/"
language: "en"
slug: "paywall-optimization-tests"
category: "grow-your-app"
date: "2026-11-01T16:19:17.573Z"
date_modified: "2026-11-01T18:54:31.790Z"
author: "Victoria Kharlan"
---

# What State of Paywalls taught me about the tests subscription apps skip

## TL;DR

- Top-app paywalls look close to identical: two plan cards, annual pre-selected, a "Start free trial" button. Polishing that screen rarely pays off: visual or copy tests raise LTV in 35% of cases.
- Offer tests win far more often. Localization, trial, and plan-duration tests raise LTV in 59–62% of cases.
- Most apps show every user the same paywall, once. 2% of headlines reflect what the user said in onboarding, and 4% of apps try again with a second offer after the user closes the paywall.
- With ten tests a year, nine should go to timing, the offer, and personalization. Visuals get one.

I spent the last month inside the [State of Paywalls 2026](https://adapty.io/state-of-paywalls/) data: 78,000+ paywalls from 30,000+ App Store apps, plus 100 paywalls from top subscription apps that we took apart screen by screen through Mobbin.

The first thing that stands out in the data is how alike the paywalls look. The second is the gap between test types: visual and copy tests have the worst win rate, and offer tests win almost twice as often. Personalizing the paywall with what the user said in onboarding barely shows up.

That points to a different testing order. The tests that win most often in our data change things the user never notices as design: when the paywall appears, which plans it offers, what it costs in their country, and what happens after they close it.

Each test below comes with the number behind it, a live paywall that does it where one exists, and a test setup.

## What does almost every top-app paywall look like?

Almost every top-app paywall shows two plans as cards, pre-selects annual, and sells a free trial on the button. In our sample of 100 iOS paywalls, 62% show two plans, 73% use cards, and 67% pre-select annual.

![Most top-app paywalls share one layout: two plan cards, annual pre-selected, a free-trial button and a cancel-anytime line](https://adapty.io/uploads/paywall-anatomy.png)
*The default paywall in our sample of 100 top iOS apps: two plan cards with annual pre-selected, a free-trial button, and a line about cancelling. Only 27% add social proof. Source: Mobbin iOS screenshots, analyzed by Adapty, 2026.*

The copy sells the trial. 66% of paywalls offer one, 56% of buttons lead with it, and 36% say "Start free trial" word for word. Only 9% of buttons use "Subscribe," "Buy now," or "Purchase." The typical headline is five words long, and most sell access, a trial, or an outcome. 2% are personalized.

![Paywall headlines mostly sell access, a free trial, or an outcome; only 2% are personalized](https://adapty.io/uploads/paywall-headlines.png)
*Most paywall headlines sell access (18%), a free trial (18%), or an outcome (17%). Only 2% use what the app knows about the user. Source: Mobbin iOS screenshots, analyzed by Adapty, 2026, 100 paywalls.*

Plenty of persuasion tools go unused. 54% of paywalls have no image, 27% show social proof, and 4% express the annual price per week.

![Most paywalls show the full annual price; only 4% break it down per week](https://adapty.io/uploads/annual-price-display.png)
*62% of paywalls show the full annual price, 24% break it down per month, and 4% per week. Savings usually come as a "Save X%" badge (41%) rather than a crossed-out price (16%). Source: Mobbin iOS screenshots, analyzed by Adapty, 2026, 100 paywalls.*

If your paywall matches this description, you're running the market default. The report calls this setup a benchmark, and it works best as the starting line for testing. It shows what most teams ship and says nothing about what converts in your app.

One caveat on this data: the Mobbin analysis describes patterns and doesn't measure their effect. The fact that 73% of paywalls use cards is no evidence that cards convert better.

## Why do so many paywall redesigns lose?

A redesign changes how the offer looks, but the user still decides on the same price, plan, and trial. In Adapty's A/B test data ([State of In-App Subscriptions 2026](https://adapty.io/state-of-in-app-subscriptions-report/)), 35% of visual or copy tests increased LTV and 31% improved conversion.

![Offer changes win more A/B tests than visual changes: localization raises LTV in 62% of tests, visual or text changes in 35%](https://adapty.io/uploads/test-win-rates.png)
*Localization, trial, and plan-duration tests raise LTV in 59–62% of cases. Visual or text-only changes do it in 35%, the lowest of any test type. Source: Adapty, State of In-App Subscriptions 2026, share of A/B tests that increased each metric.*

Adding more products to the paywall wins 53% of the time, about a coin flip.

Visual tests can still pay off. They win less often, but when they win, the lift is real: animation and stronger visuals raised ARPU by 16–34% in winning tests, and social proof lifted conversion by 10–30%. They just shouldn't take the first slots.

## Which paywall tests win most often?

Offer tests win most often. 62% of localization tests increased LTV, followed by trial tests at 60% and plan-duration tests at 59%. Timing comes before all of them, because it decides how many users see the offer at their most ready moment.

| Order | What to test | What the data shows |
| --- | --- | --- |
| 1 | Timing and offer: placement, localization, trial, plan duration, price | Onboarding paywalls convert almost 4× better; 59–62% of offer tests increased LTV |
| 2 | Plan choice: annual pre-selected with the trial on it | +25–40% ARPU in winning tests |
| 3 | Underused tests: personalized headline, targeting, trial timeline, second offer, weekly framing | +13–35% ARPU from personalized onboarding, +12–36% from trial explainers |
| 4 | Visuals and copy | 35% of tests increased LTV |

### Test when the paywall shows up

93% of first purchases happen on the day of install, and 60% of buyers subscribe on the first paywall view. Onboarding paywalls convert almost 4× better than later ones: 0.85% of views against 0.23%, or $170 per 1,000 views against $40.

![60% of buyers purchase on the first paywall view; in Health & Fitness only 40% do](https://adapty.io/uploads/view-to-purchase.png)
*60% of buyers purchase on the first paywall view, and 84% of those who need more views still buy within 24 hours of the first one. Health & Fitness (40%) and Photo & Video (44%) need the most exposure. Source: Adapty, State of Paywalls 2026, share of buyers by paywall view, first 90 days after install.*

Later paywalls still matter. 7 in 10 paywall views happen after the first session, and subscriptions started there bring 38% of revenue. In Lifestyle and Utilities, 48% of revenue starts after the first session, slightly more than during onboarding. Monthly plans skew late too: 47% of monthly revenue starts after the first session, compared with 38% for weekly and 32% for annual.

![Onboarding brings 53% of subscription revenue overall, but in Lifestyle and Utilities later paywalls bring 48%](https://adapty.io/uploads/revenue-start.png)
*Onboarding paywalls start 53% of subscription revenue overall and 66% in Graphics & Design. In Lifestyle and Utilities, later paywalls bring 48%, slightly more than onboarding. Source: Adapty, State of Paywalls 2026, share of subscription revenue by where the subscription started.*

### Localize the offer, not just the language

The median European app converts 1.38% of paywall viewers, against 2.75% in the US. The gap is moving, though: since FY24, Europe's view → paid conversion grew 18% while the US fell 14%. Category rankings shift by region too: Education ranks #3 in the US and #7 in APAC, LATAM, and MEA.

![Paywall conversion rose in every region except the US between FY24 and FY26](https://adapty.io/uploads/conversion-by-region.png)
*Median view → paid conversion rose in Europe (+18%), MEA (+13%), APAC (+11%), and LATAM (+5%) from FY24 to FY26, while the US fell 14%. The US still leads at 2.75%, against 1.38% in Europe. Source: Adapty, State of Paywalls 2026, all apps, all plans; FY runs October–September.*

![Category rankings shift by region: Education converts 3.28% in the US but 0.89% in APAC](https://adapty.io/uploads/conversion-region-category.png)
*Median view → paid conversion by region and category. Education ranks #3 in the US at 3.28% and drops to #7 in APAC (0.89%). Health & Fitness in MEA converts 0.45%, the lowest cell in the table. Source: Adapty, State of Paywalls 2026, all plans, within 30 days.*

Prices move in ways most teams don't expect. Europe is the most expensive region, at $8.06/week, $11.55/month, and $40.88/year. Annual prices fall from $39.99 in the US to $30.11 in LATAM, while weekly prices stay close to $6.99 everywhere.

![Weekly prices sit near $6.99 in most regions; Europe is the most expensive at $8.06, and Health & Fitness is the cheapest category](https://adapty.io/uploads/weekly-price-region.png)
*The median weekly plan costs $6.99 in the US and LATAM, $6.76–$6.80 in APAC and MEA, and $8.06 in Europe. Health & Fitness is the cheapest category in every region. Source: Adapty, State of Paywalls 2026, median full price in USD; European prices include VAT.*

### Match trial length to the plan

79% of trials last 1–3 days. In our data, 3-day trials perform better on weekly plans and 7-day trials perform better on monthly and annual plans. Longer trials also see fewer day-one cancellations across all three plan lengths. Education and Health & Fitness are the only categories where most trials run 4–7 days, and they also have the fewest early cancellations and the highest trial → paid rates.

![Most trials last 1–3 days, but Education and Health & Fitness lean toward 4–7-day trials](https://adapty.io/uploads/trial-length.png)
*79% of trial starts last 1–3 days. Education (63%) and Health & Fitness (55%) are the only categories where most trials run 4–7 days. Source: Adapty, State of Paywalls 2026, share of trial starts by trial length.*

### Pre-select annual and put the trial on it

The biggest single lift in our winning tests came from this layout change: pre-selecting annual and attaching the trial to it raised ARPU by 25–40%.

![Annual-forward layouts delivered the biggest lift in winning A/B tests: +25–40% ARPU](https://adapty.io/uploads/biggest-lifts.png)
*Annual-forward layouts lifted ARPU by 25–40% in winning tests, followed by trial explainer screens (+12–36%) and onboarding personalization (+13–35%). Ranges come from winning tests only, and results vary by app. Source: Adapty, State of Paywalls 2026.*

[Cal AI](https://mobbin.com/screens/b730224e-cacd-450a-98b4-74cb2be349e7) does this. The annual plan sits on top, pre-selected, under a "3 DAYS FREE" banner with "Start for free & save 75%," and the monthly plan below carries no trial. [Liven](https://mobbin.com/screens/a612cffd-939d-471f-9b89-bf734058ab1a) makes the opposite bet: a weekly plan with a 3-day trial at $7.99/week comes pre-selected, with annual and lifetime underneath.

![Two opposite bets on plan choice: Cal AI pre-selects annual with a 3-day trial, Liven pre-selects weekly with a 3-day trial](https://adapty.io/uploads/calai-vs-liven.png)
*Left: Cal AI pre-selects annual and puts the 3-day trial on it; monthly carries no trial. Right: Liven pre-selects a weekly plan with a 3-day trial, with annual and lifetime below. Screenshots via Mobbin.*

Buyers tend to stay with whatever is pre-selected: 90% keep weekly when it's the default, and 46% keep annual. Paywalls with annual pre-selected convert about as well as those with weekly.

![Buyers stick with a pre-selected weekly plan 90% of the time, against 46% for annual](https://adapty.io/uploads/preselect-by-category.png)
*Across Adapty apps, 43% of paywalls pre-select annual and 37% weekly. Health & Fitness leans hardest on annual (63%), Utilities on weekly (51%). When weekly is pre-selected, 90% of buyers keep it; for annual, 46% do. Source: Adapty, State of Paywalls 2026, share of paywalls by pre-selected plan and resulting purchases.*

### Judge price tests on LTV

Price tests trade conversion for value. 28% improve conversion, while 46% increase LTV, and a price increase that wins raised ARPU by 15–25%. If you judge a price test on conversion, you'll [throw away winners](https://adapty.io/blog/paywall-ab-testing-mistakes/).

Lifetime works as a supporting plan rather than a standalone one. 14% of apps offer it, and it brings 1% of total subscription revenue, or 5.6% in the typical app that sells it. Placed next to annual, though, a lifetime plan lifted revenue by 10–15% in winning tests.

![14% of apps offer a lifetime plan, most often in Health & Fitness (20%); it typically costs 1.3× the annual price](https://adapty.io/uploads/lifetime-by-category.png)
*14% of apps sell a lifetime plan, from 20% in Health & Fitness to 10% in Photo & Video. It typically costs 1.3× an annual plan: about the same as a year in Lifestyle and Entertainment, and 1.7× in Graphics & Design. Source: Adapty, State of Paywalls 2026.*

## Which paywall tests does almost nobody run?

Our data points to five tests with real upside that few apps run. The apps mentioned below show the pattern on their paywalls. Their test results aren't public, so treat them as examples of the setup.

### 1. Put the onboarding answer in the headline

[Yazio's paywall](https://mobbin.com/screens/d444f04c-d1b9-4872-a056-a0dff266f600) opens with "We'll help you lose weight to hit 70.0 kg by April 7!" The target weight and the date both come from onboarding. [BitePal](https://mobbin.com/screens/de10057f-86df-48b6-bae9-8f3bc8eb854a) writes in first person: "I'll help you reach 65 kg by 27 Dec." [Runna](https://mobbin.com/screens/3fa764cd-6481-433f-bfc3-63c15c5f09da) greets the runner by name and sells "your personalized training plan."

![Three paywalls with headlines built from onboarding answers: Yazio and BitePal show the user's target weight and date, Runna greets the user by name](https://adapty.io/uploads/personalized-headlines.png)
*Left to right: Yazio and BitePal put the user's target weight and date into the headline; Runna greets the user by name and sells a personalized plan. Screenshots via Mobbin.*

2% of paywall headlines do this. In winning tests, personalized onboarding raised ARPU by 13–35%.

Health & Fitness outside the US is the clearest place to start. The category has the lowest trial start rate in every non-US region: 1.46% of paywall viewers in LATAM and MEA, 2.93% in Europe. Once users start a trial, though, Health & Fitness converts more trials than any other category, from 19.8% in MEA to 33.3% in the US. The drop happens at the paywall, which is where a personalized headline works. Health & Fitness buyers also need more views: only 40% buy on the first one, and about a quarter need five or more. Each repeat view is another chance to show the user their own goal.

![Health & Fitness has the lowest trial start rate in every region outside the US, down to 1.46% in LATAM and MEA](https://adapty.io/uploads/trial-start-region-category.png)
*Median paywall view → trial start rate. Health & Fitness ranks last in every non-US region: 2.93% in Europe, 2.20% in APAC, 1.46% in LATAM and MEA, against 5.73% in the US. Source: Adapty, State of Paywalls 2026, apps that offer a trial.*

### 2. Target by who the user is and where they came from

One in four apps personalizes paywalls at all, and most of those split by context: 61% by country or language, 44% by app version, 43% by platform. 15% use their own user attributes, and 6% use acquisition source. Among mid-size apps, 69% of the top 10% by revenue per view personalize, compared with 42% of the rest.

![Most paywall personalization targets country, version, or device; only 6% of apps target by acquisition source](https://adapty.io/uploads/personalization-audiences.png)
*Among apps that target paywalls by audience, 61% split by country or language, 44% by app version, and 43% by platform. Signals about the user lag far behind: subscription status (19%), custom attributes (15%), install date (7%), and acquisition source (6%). Source: Adapty, State of Paywalls 2026.*

The easiest first segment is lapsed subscribers. 56% of returning subscribers who take an offer are still paying three months later, compared with about 1 in 10 new users who start with one.

### 3. Show the trial as a timeline

[Vocabulary](https://mobbin.com/screens/32238af5-552f-4ac3-bede-16981f4c23d3) gives the trial its own screen, "How your free trial works": the trial starts today, a reminder comes on March 26, membership begins on March 27. A reminder toggle sits right above the button. [TIDE](https://mobbin.com/screens/4946c946-896e-433d-a459-e031e0e011b6) states the date outright: "You'll be charged on 02/15, cancel anytime before." [Copilot](https://mobbin.com/screens/b6ba77c0-7d8f-423e-8453-9db2a97f8035) builds the same timeline around a one-month trial.

![Three paywalls that show the trial as a timeline: Vocabulary, TIDE, and Copilot list today, the reminder, and the charge date](https://adapty.io/uploads/trial-timelines.png)
*Left to right: Vocabulary, TIDE, and Copilot each lay out the trial as steps: access today, a reminder before it ends, and the date the charge happens. Screenshots via Mobbin.*

Only 24% of trial paywalls show a timeline. Most stop at a single line: 61% carry a note about cancelling, and 68% of those say "Cancel anytime." In winning tests, explaining how the trial works raised ARPU by 12–36%.

![Trial paywalls lean on reassurance: 74% add it near the CTA, but only 24% show a trial timeline](https://adapty.io/uploads/design-with-trial.png)
*Paywalls with a free trial add reassurance near the button (74% vs 37% without a trial) and pre-select annual more often (78% vs 48%). Only 24% show a trial timeline. Paywalls without a trial work harder on value: images (56%), feature lists (67%), and "Save X%" badges (56%). Source: Mobbin iOS screenshots, analyzed by Adapty, 2026.*

The timeline targets the first hour. 32% of trials are canceled within an hour and 44% within 24 hours. In Europe, day-one cancellations reach 55%, compared with 40% in the US and 34% in MEA. Health & Fitness and Education have the fewest early cancellations and the highest trial-to-paid rates.

### 4. Make a second offer after the close

[Sunlitt](https://mobbin.com/screens/0c960695-73aa-4a20-ac28-86648ddf5246) follows a closed paywall with "Save 50% on your first year" and the line "Close this page, and it's gone forever." [Beside](https://mobbin.com/screens/6beb136d-47f6-4ded-bc0e-c9e4a6bc4ed0) offers "33% off forever" on monthly. [Vocabulary](https://mobbin.com/screens/d5dd533b-c0d6-4eb0-ae96-e7fc4d471046) goes to 63% off.

![Three second-chance paywalls: Sunlitt offers 50% off the first year, Beside 33% off forever, Vocabulary 63% off](https://adapty.io/uploads/second-offers.png)
*Left to right: Sunlitt (50% off the first year), Beside (33% off monthly, forever), and Vocabulary (63% off annual). All three frame the offer as one-time. Screenshots via Mobbin.*

96% of apps show nothing after the first paywall closes. The 4% that show a second one get it to convert at 0.88× the rate of the main paywall, and in 45% of those apps it beats the main paywall.

![Only 4% of apps show a second-chance paywall; it converts at 0.88× the main paywall with a typical 38% discount](https://adapty.io/uploads/second-chance-offer.png)
*4% of apps show a second paywall right after the first one is closed. It converts at 0.88× the rate of the main paywall. The typical discount is 38%, ranging from 23% in the lower quarter of apps to 50% in the upper quarter. Source: Adapty, State of Paywalls 2026.*

Timing supports showing it right away. 40% of buyers need more than one paywall view, and 84% of buyers who purchase on the second view or later do it within 24 hours of the first view. Only 4% of buyers in the typical app come back on another day to purchase.

The depth of the discount decides whether the test pays off. The typical second offer is 38% off, and a quarter of apps go to 50% or more. Subscribers who start at up to 30% off bring in 90% of full-price revenue over six months. At 31–60% off, that drops to 70%, and above 60% to 46%. At Vocabulary's 63%, a subscriber from the second offer brings in less than half of a full-price one.

![Deeper first-payment discounts cut six-month revenue: $90 per $100 at up to 30% off, $46 above 60% off](https://adapty.io/uploads/discount-revenue.png)
*Subscribers who start at up to 30% off bring in 90% of full-price revenue over six months; above 60% off, just 46%. The same offer keeps 56% of returning subscribers paying after three months, against about 1 in 10 new users. Source: Adapty, State of Paywalls 2026, six-month revenue per subscriber indexed to full price = $100.*

### 5. Show the annual price per week

[Runna](https://mobbin.com/screens/3fa764cd-6481-433f-bfc3-63c15c5f09da) lists the annual plan at S$109.98/year with S$2.11/week next to it, and does the same for monthly (S$4.60/week). The comparison takes one glance. [TIDE](https://mobbin.com/screens/be7bd970-cfc3-4196-a415-9029461af0df) puts its yearly plan at $1.06/week next to a $4.99/week weekly plan, and [Fabric](https://mobbin.com/screens/667e5ecf-e086-4bc6-843a-1a129f8e82b4) prices all three plans per week.

![Three paywalls that show the annual price per week: Runna, TIDE, and Fabric](https://adapty.io/uploads/weekly-framing.png)
*Left to right: Runna shows the annual plan as S$2.11/week, TIDE shows yearly at $1.06/week against $4.99/week for weekly, and Fabric prices weekly, monthly, and annual plans all per week. Screenshots via Mobbin.*

62% of paywalls show the full annual price, 24% break it down by month, and 4% show it per week. Meanwhile, weekly plans bring 62% of all subscription revenue. Your users already think in weekly prices.

![Weekly plans bring 62% of subscription revenue, from 50% in Europe to 76% in LATAM](https://adapty.io/uploads/revenue-plan-region.png)
*Weekly plans bring 62% of subscription revenue overall, against 21% for annual and 12% for monthly. The share ranges from 50% in Europe, where annual and monthly matter more, to 76% in LATAM and 70% in MEA. Source: Adapty, State of Paywalls 2026, Oct 2025–Sep 2026.*

The math is simple. The typical app charges $6.99/week or $39.99/year. Per week, the annual plan comes to $0.77, about 9× cheaper than the weekly plan. On most paywalls, the buyer has to work that out alone.

### Why so few apps run these tests

Each of these tests needs the paywall to read data the app already has and to change without a release. In Adapty, [segments](https://adapty.io/docs/segments) can filter by custom attributes, attribution source and campaign, and subscription state. The [Flow & Paywall Builder](https://adapty.io/flow-builder/) turns onboarding answers into variables for paywall copy, pricing, and offers, and you ship changes without an app update. For a no-code walkthrough, see [how to personalize onboarding and paywalls](https://adapty.io/blog/how-to-personalize-onboarding-and-paywalls-in-your-mobile-app/).

## How many paywall tests should you run a year?

Among apps with 500K+ monthly paywall views, the top quarter runs 21+ tests a year, and the top 10% runs 55+, more than one a week. In the 50K–500K range, the top quarter runs 9+, and the top 10% runs 24+. The most active apps also keep more paywalls live: 72% of apps above 500K views use five or more, against 38% below 50K.

![The most active large apps run 55+ paywall tests a year; mid-size apps in the top quarter still run 9+](https://adapty.io/uploads/test-cadence.png)
*A/B tests started per app over 12 months. Among apps with 500K+ monthly paywall views, the top quarter runs 21+ and the top 10% runs 55+. In the 50K–500K range, the top quarter runs 9+ and the top 10% runs 24+. Source: Adapty, State of Paywalls 2026.*

Smaller apps get fewer tests, because each one needs [around 200 subscriptions per variant](https://adapty.io/blog/paywall-ab-testing-mistakes/) before you can trust the result. A ten-test year could look like this:

- Five on timing and the offer: placement, localization, trial length per plan, annual-first, and a price test judged on LTV.
- Four on the underused tests: personalized headline, targeting by source or subscription status, trial timeline, and a second offer.
- One on the visuals.

---

The short version of the report: test when the paywall shows up and what it offers before you test how it looks. The top performers in our data keep going past the default: 69% of the top 10% of mid-size apps personalize, compared with 42% of the rest, and the most active large apps start a new test every week.

Adapty's Flow & Paywall Builder lets you build those variants, target them by onboarding answers or ad source, and A/B test them without an app release. [Book a demo](https://adapty.io/schedule-demo/).

To see where your current paywall stands first, read the full [State of Paywalls 2026 report](https://adapty.io/state-of-paywalls/) or add it as a [Claude skill](https://github.com/vkharlan-adapt/state-of-paywalls-2026-skill): it compares your setup with the report's benchmarks and suggests your next test.

## FAQ

### What is a good paywall conversion rate?

The median app converts 2.75% of paywall viewers to paid within 30 days in the US and 1.38% in Europe. The top 10% start at 3.30% in MEA and 5.71% in Europe. Compare yourself with the top apps in your region instead of the US median.

### When should you show the paywall?

Show your strongest paywall in onboarding: 93% of first purchases happen on the day of install, and onboarding paywalls convert almost 4× better than later ones. Plan at least one later placement too, since paywalls after the first session still bring 38% of subscription revenue.

### Should you pre-select the annual plan?

67% of top-app paywalls already do. In Adapty data, paywalls with annual pre-selected convert about as well as those with weekly, and annual-first layouts lifted ARPU by 25–40% in winning tests. If most of your revenue comes from weekly plans, test before you switch.

### Is a hard paywall better than a soft one?

Within the same app, hard paywalls get 4.7% of viewers to start a trial, against 3.2% for soft ones. Soft paywalls drive more direct purchases: 0.75% against 0.50%. Counting trials and purchases together, hard paywalls perform better in 57% of apps.

### How long should a free trial be?

Match the trial to the plan: 3 days for weekly, 7 days for monthly and annual. Longer trials also see fewer day-one cancellations.

### Do you need a lot of traffic to A/B test paywalls?

You need enough to reach about 200 subscriptions per variant. In the top quarter of apps with 50K–500K monthly paywall views, teams still run 9+ tests a year.