What Is a Good App Paywall Conversion Rate?

Ask what a good paywall conversion rate is and you will get answers between 1% and 50%, each one sourced to a real study and none of them wrong. The spread exists because three completely different measurements share the same name, and the benchmark you compare yourself against is usually built on a denominator your analytics dashboard is not using.

Sorting that out first is worth more than any single number. Adapty’s State of In-App Subscriptions 2026, built on $3 billion in revenue across more than 16,000 apps, measures paywall views converting to payment. RevenueCat’s State of Subscription Apps 2026, covering more than 115,000 apps and $16 billion in tracked revenue, measures installs converting to paid. Both are credible and they answer different questions.

What counts as a paywall conversion rate?

A paywall conversion rate divides payers by one of three different denominators, and naming which one you mean is the whole job. Paywall view to payment divides payers by the number of people who saw a paywall screen. Install to paid, also written as download-to-paid, divides payers by everyone who installed. Trial to paid divides payers by everyone who started a trial.

The numbers those produce differ by roughly a factor of ten from end to end, so the choice of denominator swamps the difference between a good paywall and a bad one.

Metric	What it divides by	2026 median	Source
Paywall view to payment	everyone who saw a paywall	4.85% soft, 3.34% hard	Adapty
Install to paid, day 35	everyone who installed	2.6% iOS, 0.9% Android	RevenueCat
Trial to paid	everyone who started a trial	25.6% average	Adapty

A team reading 3% in a paywall tool and comparing it against the 25.6% trial-to-paid average reported by Adapty will conclude the paywall is catastrophic when it is close to the median for its own denominator.

Which one you should optimise for depends on what you control. Paywall view to payment is the right number for paywall work, because it isolates the screen. Trial to paid is the right number for pricing and trial length.

Install to paid is the number for acquisition. It is the only one denominated in the same unit as a cost per install, which makes it the only one that tells you whether a campaign pays back.

Report all three and the diagnosis usually writes itself. Low install-to-paid with healthy paywall-view-to-payment means most users never reach the paywall, which is an onboarding problem. Healthy install-to-paid with weak paywall-view-to-payment means the screen itself is losing people who arrived ready to buy.

What exactly is the denominator when a tool reports a paywall conversion rate?

In the tools, the denominator is paywall views, and a view is an impression rather than a person. One user who opens the paywall on Monday, backs out, and opens it again on Thursday is two views in the denominator and one human being.

Adapty says so directly. Per Adapty’s paywall metrics documentation, a user who visits the paywall twice is recorded as two views, and the dashboard carries two separate rates for that reason.

Conversion rate to purchases divides purchases by total views. Unique conversion rate divides purchases by unique views, so 10 purchases against 100 unique views reads as 10%, per the worked example in those same docs.

The same page notes that views refresh every few minutes while other metrics update in real time, which is enough to make a rate look wrong for a quarter of an hour after a release.

Superwall counts the same way under a different name. Its analytics pages describe paywall views as the total number of times a paywall was successfully delivered and presented to a user, an exposure count rather than a headcount, and its initial conversion rate is the share of brand new installs that convert on their first paywall view, per Superwall’s conversion stats documentation.

RevenueCat uses a third convention again. Per its paywall LTV chart documentation, customers are cohorted by their first paywall impression and revenue is credited from conversions in the three days after that impression, and only impressions from paywalls built with RevenueCat Paywalls are tracked, so a custom-coded screen produces nothing in that chart.

So if you fire your own event, a paywall_reached or a paywall_shown, you have already picked the impression denominator, because the event fires every time the screen appears rather than every time a new person sees it. Deduplicate to the user and you have the other rate.

Both rates are worth reporting. Per the two Adapty formulas above, a month with 1,000 people reaching the paywall, 1,600 impressions between them because some came back, and 48 purchases reads as 3.0% on views and 4.8% on unique views. The gap between those two numbers is your repeat rate, and it moves when you change where the paywall sits rather than what it says.

Use the per-view rate when you are comparing yourself against a published benchmark, because the published ones are built on views. Use the per-user rate when you are deciding whether the screen works, because a person who saw it four times and never bought did not reject it four times.

How do Adapty, Superwall and RevenueCat measure paywalls differently?

Adapty, Superwall and RevenueCat measure paywalls differently because each one defines the paywall event in its own way, which is why the same app wired into two of them reports two conversion rates. The table below is what each vendor’s own documentation says, read in October 2026.

Tool	What it counts in the denominator	The headline rate it shows	Where it stops counting
Adapty	every paywall view, with a parallel count of unique views	conversion rate on views, plus a unique conversion rate on unique views	views lag the other metrics by a few minutes
Superwall	every paywall successfully delivered and presented	paywall conversion after opening a paywall, plus an initial conversion rate on fresh installs at their first view	its pricing counts a conversion as a purchase made through a Superwall paywall, per its pricing FAQ
RevenueCat	first paywall impression per customer, as a cohort	conversions in the three days after that first impression, as paywall LTV	impressions from paywalls not built in RevenueCat Paywalls

Three of those differences bite in practice. A RevenueCat paywall chart on a custom-coded screen is empty rather than low, per the scope note in its own documentation, which is worth checking before concluding a paywall is failing.

A Superwall initial conversion rate is not comparable to an Adapty conversion rate, because one is denominated in fresh installs and the other in impressions. And a three-day attribution window, which is what RevenueCat’s paywall LTV chart credits, will under-count an app whose buyers take a week to decide.

None of that makes one vendor’s number more honest than another’s. It does mean a migration between them produces an apparent change in conversion rate on a product that did not change, and the only safe comparison across a switch is a metric you compute yourself from your own events on a fixed definition.

What is a good paywall view to payment rate?

A good paywall view to payment rate sits between 3% and 5%, which is the middle of the distribution. Adapty’s 2026 study found soft paywalls at 4.85% of paywall views converting to payment and hard paywalls at 3.34%, which makes soft gating the winner on this specific metric by close to 50%.

That result reverses when you change the denominator, and the reversal is instructive rather than contradictory. A hard paywall shows itself to every user who opens the app, so its denominator includes everyone. A soft paywall is often reached by people who went looking for a premium feature, so its denominator is pre-filtered toward intent.

Measured from the install instead, RevenueCat’s study has hard paywall apps at a median 10.7% download-to-paid against 2.1% for freemium. Same underlying behaviour, opposite ordering.

Neither figure is the manipulated one. They are answers to “how good is this screen at closing the people who see it” and “how good is this model at producing payers from installs”, and a monetization decision needs both. The model choice behind them is covered in freemium vs free trial vs hard paywall.

For a practical target: if your paywall converts more than the 5% at the top of that study’s range and your install-to-paid is still weak, stop working on the paywall. The screen is doing its job and the users are not reaching it.

What share of installs should end up paying?

The share of installs that ends up paying runs to a median 2.6% on iOS and 0.9% on Android within 35 days, per RevenueCat’s 2026 study.

That is the number to hold against a cost per install, and with global iOS cost per install at $5.84 in the first quarter of 2026 on AppsFlyer data compiled by Digital Applied, it is also the number that explains why subscription apps are hard to make work on paid traffic alone.

Three things move it more than paywall quality does. Model is the largest, at 10.7% for hard gating against 2.1% for freemium in the same study.

Price is next: according to RevenueCat, high-priced apps convert a median 2.66% of downloads against 1.49% for low-priced ones, which runs against the intuition that a lower price buys volume and is worked through in how to price a subscription app.

Category does the rest, and the spread between categories is wider than the spread between competent and incompetent execution inside one.

Traffic source belongs in the same list even though it rarely appears in published benchmarks. Installs from a creator video, a search ad and an incentivised placement convert at rates that are not comparable, and a single mis-targeted campaign can pull a whole account below its category median while every other source performs normally. Segment install-to-paid by source before you conclude anything about the product.

Use the medians above as an order-of-magnitude check, 2.6% on iOS and 0.9% on Android per RevenueCat’s 2026 study. At 0.2% on iOS, something is structurally wrong. At 2% you are normal, and the way to find out whether you should be at 4% is your own history rather than someone else’s average.

Why is the Android number so much lower than iOS?

The Android gap is real, industry-wide and mostly not about the paywall. RevenueCat’s 2026 study has day-35 download-to-paid at 0.9% on Android against 2.6% on iOS, a gap of nearly three to one that shows up across categories.

Geography explains a large share of it. Google Play’s install base spans far more markets with far lower purchasing power, so an Android median is an average over a much wider price-sensitivity distribution than an iOS median. Comparing a US-only iOS cohort to a worldwide Android cohort is not a platform comparison at all.

Funnel entry explains more of it than most teams check. RevenueCat’s own engineering write-up on the Android gap argues the problem is usually where users enter the funnel rather than the trial or the paywall itself, which means the useful diagnostic is what share of Android installs ever see a paywall, not what share of paywall viewers pay.

The practical rule is to benchmark Android against Android, split by market, and look at paywall impressions per install before touching the paywall. A 0.9% Android median is not permission to ignore the platform, since Play volume can make a lower rate perfectly viable, but it is permission to stop treating the gap as a bug in your own funnel.

How much does paywall placement move the number?

Placement moves it more than design does. Adapty’s 2026 study found paywalls triggered after a measurable value moment producing 2.1 times the trial start rate of paywalls shown immediately on open.

Timing has a much shorter tail than the standard advice assumes. According to the same Adapty study, 89.4% of all trial starts happen on install day, with days 1 to 3 contributing 2.0% and the whole stretch from day 4 to day 30 contributing under 2.4% between them.

The common instruction to wait a month before judging a paywall is mostly wrong, then: by day three the result is in, and Education is the only category in the study where users routinely install and commit weeks later.

If the trials are not arriving at all, the cause is usually upstream of this screen: why app installs are not converting into trials works through the ad to paywall seam and the 30-day window the published benchmarks use.

Design effects are real and smaller. Superwall, analysing more than 40 million paywall opens, measured multi-page onboarding paywalls converting at 12.41% against 9.07% for single-page. A 37% relative lift is worth having. It is also the kind of gain that gets erased by leaving the paywall in session three where most users never reach it.

Recovery is the placement question teams skip. Superwall’s study of exit offers across 18 apps and roughly 500,000 users attributed 17% of total revenue to offers shown at the moment of abandonment, which means the second paywall is doing a meaningful share of the work of the first.

How should you benchmark your own paywall?

Benchmark against your own previous cohorts, and use published medians only to check you are in the right order of magnitude. Category, price, platform and traffic source each move these numbers by more than the gap between a good paywall and a mediocre one, so an external average cannot tell you whether you are underperforming.

Set up the reading in two cuts. Cohort by install date rather than a rolling window, because trailing traffic in a growing app makes a working paywall look broken. Then segment by platform and by acquisition source, since a blended figure across iOS and Android is an average of two populations that behave almost three to one apart.

Watch for the two failure modes that a single blended number hides. A high paywall-view-to-payment rate with very few paywall views usually means the paywall is too hard to reach, and the fix raises revenue while lowering the rate you were proud of. A high trial-start rate with weak trial-to-paid usually means the offer is being reached by people the ads recruited for a different product.

That last one is an acquisition problem wearing a monetization problem’s clothes, and it is the seam where most app teams lose money: the ad promises one thing, the paywall sells another, and each team reads its own metric as healthy. Running both ends as one experiment is the part of an app launch we take on directly.

The short version

A good paywall conversion rate is 4.85% of paywall views for a soft paywall and 3.34% for a hard one, per Adapty’s State of In-App Subscriptions 2026. Measured from the install, the median is 2.6% on iOS and 0.9% on Android by day 35, per RevenueCat’s State of Subscription Apps 2026. Measured from trial starts, the global average is 25.6%, from 35.0% in Health and Fitness to 19.1% in Entertainment.

Name your denominator before you compare anything. Report paywall view to payment for paywall work, install to paid for acquisition, and trial to paid for pricing, because each one answers a different question and mixing them produces confident wrong conclusions.

Then fix placement before design. Per Adapty, a paywall shown after a real value moment starts 2.1 times as many trials as one shown on open, and 89.4% of trial starts happen on install day, so the first session is very close to the whole opportunity.

An exit offer at the point of abandonment carried 17% of revenue in Superwall’s study of 18 apps. Benchmark yourself against your own last cohort, split by platform and by acquisition source.