Hard paywall or freemium is the wrong first question.

RevenueCat’s 2026 State of Subscription Apps report, built from more than 115,000 apps and over $16 billion in revenue, puts a number on the oldest argument in subscription apps. Apps behind a hard paywall have a median download-to-paid conversion of 10.7%. Freemium apps sit at 2.1%. Five times better is the kind of figure that ends a debate, and it shouldn’t, because the same dataset says the model you pick matters less than one thing most teams never write down: the moment a new user can see what they’d be paying for.

What the five-times figure actually measures

The report defines download-to-paid conversion as the share of installs that start at least one paid subscription within 35 days of installing. It counts installs, not trial starts, so it captures everyone who arrived, including the people who never saw the paywall twice.

On that measure the gap is real and it carries into money. Median revenue per install at day 60 is $3.09 for hard paywall apps and $0.38 for freemium apps, roughly eight times. If you buy installs, that difference is the whole game: the same ad spend returns eight times the revenue inside two months.

Then look at what happens to the people who do pay. RevenueCat’s benchmarks put one-year retention of yearly subscribers at 27% for hard paywall apps and 28% for freemium, a gap it calls statistically negligible. A hard paywall doesn’t buy you more loyal subscribers. It buys you more of them, sooner.

One caution the numbers can’t settle on their own: this is observational data. Apps choose their model, and the kinds of apps that choose a hard paywall (often well-funded, often buying traffic, often in categories where the promise is easy to state) may convert better for reasons that have nothing to do with the paywall.

The spread inside each model is bigger than the gap

The more useful number is how far hard paywall apps differ from each other. The low end of the range RevenueCat charts for them is 4.2%, and the top 10% reach 38.7%. A top-quartile freemium app clears 4.5%, which already beats the bottom of the hard paywall range. The report’s own reading is that the width of that spread suggests execution matters more than the model alone.

Put plainly, a well-run app on either model can beat a badly run app on the other.

Freemium pays later, and late is easy to miss

The two models differ most in timing. For both, about a third of all conversions happen on the day of install, and more than 60% have happened by day 7. After that they split. In freemium apps, 23% of conversions don’t arrive until six weeks or more after the download. Hard paywall apps see a spike between days 4 and 7 instead, which the report attributes to 7-day trials expiring.

That has two practical consequences. First, if you run freemium and judge your funnel at 30 days, you’re missing close to a quarter of what that cohort will eventually pay. Second, freemium is slower cash. A bootstrapped developer paying for installs has to wait out those six weeks before the money comes back, and the hard paywall’s day-one revenue is worth more to someone who needs it this month.

The same switch went both ways

RevenueCat’s Sub Club podcast carried a useful pair of results from growth advisor Phil Carter, who has moved client apps between models. One established app went from a hard paywall to freemium, telling new users in onboarding that the product was free and always would be, then offering the best version for seven days. Together with other pricing and packaging changes, he reported a 75% increase in lifetime value per user. A different client tried the same move and saw subscriber conversion fall by more than 50%. They pulled it after a couple of weeks.

His summary is the line worth keeping: “Freemium isn’t a pricing strategy; it’s a product strategy.” Whether giving the product away grows the business depends on what people get from the free part, and that’s a property of the product, not the paywall.

Ask when the value shows up

So before choosing a model, answer one question honestly: at what point does a new user feel the thing they’d pay for? Most apps fall into one of three answers.

Before they use it. The value is legible from the promise: a training plan, a meditation course, a filter pack, a language course. People can judge it from your screenshots and a 30-second onboarding, so asking for a trial up front is a fair trade. These are the apps that fill the top of the hard paywall range.

After they’ve put something in. A habit tracker, a budget, a notes app, a design tool: the value appears once the user’s own data, time or work is inside. A hard paywall here asks people to pay for a future they can’t see yet, and many won’t. Freemium, or a trial long enough to reach that point, lets them get there first.

When other people are there. Shared lists, communities, anything with user-generated content or a network effect. Every free user makes the product better for the paying ones, which is why Carter lists network effects and UGC among the reasons to test freemium at all.

Then put your constraints next to the answer. Carter’s advice is to stay with a hard paywall if you’re bootstrapped or short of capital, need cash to fund paid acquisition, solve an acute problem, or lack the product analytics to test what to gate. Freemium is for teams with strong retention, the runway to absorb a conversion dip, and the instrumentation to see what’s working.

If you go freemium, gate the output

The common mistake is gating the first experience, so free users never reach the moment that sells the upgrade. Gate the thing that delivers value repeatedly instead: exports, saved items, projects, history. Let people do the full job once and meet the limit on the second or third time round, when they already know it’s worth paying for.

That’s how we set up ShotCanvas. The free tier caps monthly exports, saved drafts and the number of templates, while the editor itself is open, so someone can design a full screenshot set before deciding whether they need more. Show the limit before it bites, too. A counter of exports left this month is a reminder; a wall mid-task is an ambush.

If you go hard, earn the ask

A hard paywall concentrates everything into the minutes before the ask. Hannah Parvaz, quoted in the report, puts the failure mode simply: “If a paywall appears before context is established, it feels jarring.” Onboarding has to show the product working, ideally on the user’s own input, before the price appears.

Trial length is the other lever. Across the dataset, trials of 17 to 32 days have a median trial-to-paid rate of 42.5%, against 25.5% for trials of four days or fewer, yet 46.5% of trials are now four days or shorter. Short trials also lose people instantly: 55.4% of 3-day trials are cancelled on day zero, compared with 39.8% of 7-day trials. Part of that gap is the apps that choose long trials, so treat it as a reason to test a longer trial, not proof that one will work. Our post on counting trial conversions covers how to read that test without fooling yourself.

And remember where the sale really starts. For a hard paywall app, the store listing is the only part of the product a person sees before they’re asked to commit, so the screenshots have to carry the promise on their own.

Figures are from RevenueCat’s State of Subscription Apps 2026, its benchmark summary and the hard paywall vs. freemium write-up of Phil Carter’s Sub Club interview.