Subscription app pricing usually works best with two or three plans: an annual plan set as the default, plus a weekly or monthly option that makes the annual price look like a deal.
Pick a starting price near your category's median, then raise it with A/B tests, because tests in our database show apps often charge too little.
Judge each price on revenue per user, not conversion rate alone.
This guide covers the main app pricing models, how anchoring and default plans work, the rules for price tests on iOS and Google Play, localization, and what the price tests in our database show.
App pricing models: weekly, monthly, annual, lifetime and credits
Most subscription apps mix a few of these models.
Each one suits a different user and a different cash-flow need.
- Weekly: a low upfront price and a high cost over a year. Common in utility, AI and photo apps where people want one result fast. Churn is high, so weekly plans need steady new installs.
- Monthly: the familiar middle option. It often works best as an anchor that makes the annual plan look cheap.
- Annual: the most cash up front and the fewest renewal chances to lose someone. It's the default for most health, education and productivity apps.
- Lifetime: a one-time purchase. It brings in early cash and appeals to people who dislike subscriptions, but it caps what each customer is worth.
- Credits or consumables: packs of uses, common in AI apps with real per-use costs. They can pull buyers away from the subscription.
Benchmarks help you pick a starting point. RevenueCat's State of Subscription Apps 2026, based on more than 115,000 apps, puts the median annual plan at $34.80, up from $31.60 the year before.
The plan mix varies widely by category: the same report finds gaming sales are mostly weekly, while productivity apps sell mostly yearly plans.
Plan length interacts with the trial. AEZAKMI Group tested a PDF scanner's paywall and found monthly converted at 1.21%, weekly at 1.63% and weekly with a trial at 2.29%.
Pixelcut saw 16% more paying customers with a weekly plan instead of monthly (vendor-reported in a RevenueCat customer story).
Credits can compete with the subscription. Mojo hid its credit packs and sold only subscriptions, and saw +14% new subscription revenue (vendor-reported by Superwall).
For more on structuring the plan list, see our plans tests and the guide to in-app purchases.
Anchoring and default plans
People judge a price against whatever sits next to it.
A paywall's job is to put the plan you want to sell next to one that makes it look like a good deal.
Show the annual plan as a weekly or monthly price
Breaking the annual price into a smaller unit is one of the most repeatable wins we have.
Mojo tested showing its annual plan as a monthly price: +45% in Brazil, +26% in Mexico and +8% in the USA.
Stardust combined a discount badge with a weekly-equivalent price and got +50% conversion, +62.5% proceeds per user and a 30% higher yearly share.
Preu AI priced its annual plan against the monthly one and saw +80% proceeds per user and +22% trial conversion.
Rash ID moved to three plans with a best-offer badge and reported about +20% ARPU and over +50% conversion to paid. All three are vendor-reported Superwall case studies.
Order and default selection
Plan order matters too. Microsoft 365 put the pricier plan first, and both plans sold more, the cheaper personal plan most of all.
Don't remove the anchor entirely
Showing only the annual plan is a common next step, and the results are mixed.
Mojo's yearly-only paywall raised the yearly share by 15-20 percentage points.
Stardust's yearly-only paywall lost on both conversion and proceeds per user, and so did hiding other plans behind a "More Pricing" drawer.
The cheaper plan often does its job just by being visible.
Price testing in apps
Price tests in apps run on the paywall.
You show different user groups different products or prices through a paywall tool such as Superwall or RevenueCat, and compare revenue per user.
You can't show two prices for the same product, so each test price is its own product in App Store Connect or Google Play Console.
Store rules matter most once you change the price for people already paying.
Apple offers 800 price points in each currency across 175 storefronts, and lets you keep the current price for existing subscribers when you raise it.
If you raise prices for existing subscribers, Apple requires their consent when the increase is over 50% and over about $5 per period (or $50 a year for annual plans), when the person had another increase in the past 12 months, or in regions that require consent.
Otherwise Apple notifies subscribers 60 days ahead for annual plans, 27 days for monthly and 7 days for weekly.
Google Play keeps existing subscribers on their old price in a legacy price cohort until you move them.
Moving them is normally an opt-in increase: users get at least 30 days' notice and are canceled if they don't accept.
In some countries, developers in good standing can run one opt-out increase per base plan per year.
It's capped at the greater of 50% or 17 US cents a day.
The safest pattern is to test new prices on new users only, then decide later whether to move legacy subscribers.
Our broader guide to price testing covers the method outside of apps.
Localization
A price that works in the US can be far too high elsewhere.
RevenueCat's 2026 report puts the median annual plan at $39.99 in North America and about $18.32 in India and Southeast Asia.
The App Store and Google Play convert your base price automatically, but they don't adjust it for purchasing power.
Localizing usually means setting your own price per country, then translating the paywall copy and matching it to local habits.
Daily Horoscope localized its paywalls by country and reported a +183% trial start rate.
Photoroom saw +50% upsell screen conversion with localized upsell paywalls.
Both are vendor-reported, and both changed more than the price at once.
Mojo's monthly-anchor test shows the same paywall change can vary by market: the lift was more than five times larger in Brazil than in the USA.
Run price tests per region when you have enough traffic, or at least split results by country.
What the database says about app pricing
We collected every app test tagged with pricing. A few patterns stand out.
Raising prices often works
Most of the price-increase tests we have came out ahead on revenue per user:
- Mojo raised its yearly price 50% and saw only a minimal drop in free-to-paid conversion.
- BuyBye tested a 50% higher yearly price: conversions fell 19%, but proceeds per user rose 25% (vendor-reported by Superwall).
- PropGPT moved its weekly price from $6.99 to $9.99 and got +47.5% proceeds per user (vendor-reported).
- Planner 5D tested two higher tiers and reported +83% ARPU for the highest one (vendor-reported).
- LookLab AI paired a 3-day trial with a higher weekly price and went from 8.3% to 15.0% campaign conversion (vendor-reported).
This matches the wider benchmark. In RevenueCat's report, higher-priced hard-paywall apps converted downloads to paid at a 2.8% median by day 35, against 2.0% for mid-priced and 1.4% for low-priced apps.
That's a correlation across apps, not a test, but it shows a low price doesn't buy you more conversions.
Lower isn't always wrong
SellRaze lowered its annual price and got +10% ARPU (vendor-reported).
The lesson is to test in both directions.
Raising prices on existing customers costs more
TalkingParents raised its price alongside a new feature.
About a quarter of customers churned, but the company came out better financially overall.
That's a before-after comparison, not a controlled test, so treat it as one data point.
Discounts and offers
Discounts change the price people feel they're paying without changing the list price.
Flibbo's discounted first week brought +20% revenue (vendor-reported). For more, see our discount tests.
How to run a subscription price test
- Pick one variable. Test the price, the plan mix or the display, not all three at once. Many wins above came from bundled changes, which makes it hard to know what worked.
- Choose revenue per user as the main metric. Use proceeds per user or ARPU over a fixed window, such as 30 or 60 days. BuyBye's test lost 19% of conversions and still won.
- Create separate store products. Set up each test price as its own product in App Store Connect and Google Play Console, then assign them through your paywall tool.
- Test on new users only. Keep existing subscribers on their price so you don't trigger consent prompts or churn mid-test.
- Size the test before launch. Price tests need more data than design tests because revenue varies a lot from person to person. Use our sample size calculator.
- Wait for renewals. Trials and weekly plans settle within weeks, but annual plans take a year to show renewal rates. Watch refunds and trial cancellations as early warnings.
- Model the lifetime value. Plug each variant's price and retention into our subscription LTV calculator before you pick a winner.
- Roll out by region. Check whether the winner holds in each major market before applying it everywhere.
For paywall layouts that pair well with a new price, browse our paywall tests and the paywall guide.