A hard paywall makes users pay or start a trial before they can use the app; a soft paywall can be dismissed, so people can keep using a free version.
Hard paywalls usually convert far more downloads into payers, and RevenueCat's 2026 data shows one-year retention is about the same for both.
Soft paywalls trade short-term revenue for reach, which pays off only when free users bring real value, such as word of mouth or network effects.
Below are the definitions, the trade-off in numbers, what real app tests from the LiftArchive database show, and how to test a switch without guessing.
Hard paywall vs soft paywall: definitions
A hard paywall locks the core product. A new user sees the paywall, often at the end of onboarding, and can't get past it without starting a trial or a paid plan.
There's no close button that leads to a usable free app.
A soft paywall has a way out. The user can close it and keep using the app with limits: fewer features, capped usage, a smaller content library, or ads.
The paywall comes back at key moments, such as when someone hits a limit.
This is the model most people call freemium.
There are hybrids in between:
- Metered: a set number of free uses, then a hard gate.
- Dismissible with a second offer: closing the paywall triggers a discount or a different plan.
- Ad-supported exit: the paywall offers "continue with ads" as the free path.
- Hard paywall with a trial: technically hard, but the free trial gives users a taste first. RevenueCat counts these as hard paywalls in its benchmarks.
Hard vs soft is a different question from trial vs no trial.
For that one, see freemium vs free trial.
The revenue vs retention trade-off
The common belief is that hard paywalls win on revenue and soft paywalls win on retention.
The first half holds up in benchmarks. The second half mostly doesn't.
RevenueCat's State of Subscription Apps 2026, built on data from more than 115,000 apps, reports these medians:
| Metric | Hard paywall | Freemium (soft) |
|---|---|---|
| Day 35 conversion to paid | 10.7% | 2.1% |
| Revenue per install, Day 60 | $3.09 | $0.38 |
| Yearly subscribers retained after one year | 27% | 28% |
So hard paywall apps earned about 8x more per install in the first 60 days, and their paying users stuck around about as long.
RevenueCat also notes that hard paywall conversion fell from 12.1% in 2025, while freemium held at 2.1%.
Two caveats matter. First, these are medians across very different apps, not a controlled experiment, and apps that pick a hard paywall may simply have stronger value propositions.
Second, the retention figure covers paying subscribers only.
A soft paywall keeps far more people in the app for free, and that's where its value lives.
What a soft paywall can buy you:
- Reach: more active users who can recommend the app, share content, or invite others.
- Later conversion: people who need weeks to see the value get that time.
- Ad revenue: free users can still earn money.
- Fewer bad payers: users who aren't forced in are less likely to refund or complain.
What it costs: a much lower share of users paying, more support and server load per dollar earned, and a product that has to be good enough for free to keep people coming back.
What the tests show
The LiftArchive database holds several real cases of apps moving between hard and soft paywalls.
They point in both directions, which is the point: the right answer depends on the app.
Tests where the hard paywall won
Headspace ran an A/B test that locked all content behind the subscription instead of leaving some free.
The team reported a significant rise in paid subscribers and, perhaps surprisingly, higher content engagement from trial users who had committed early.
V1 Sports replaced its freemium model with paid plans that came with trials and saw +90% revenue within 12 months.
That's a before-after comparison, not a controlled test, so other changes over that year could have played a part.
In a case shared by consultant Phil Carter, a subscription app that switched to freemium saw more than 50% lower subscriber conversion and had to reverse the change.
Paywall timing tells a related story. When Greg showed a paywall in onboarding to everyone, trial sign-ups rose about 400%, with sign-up-to-trial going from 3% to 15%.
Rootd moved its paywall to the start of onboarding and reported 5x revenue (before-after).
Both suggest that more people seeing the paywall early does more than a free path does. More on that in paywall timing.
Tests where the soft paywall won
Opal replaced its hard paywall with a generous free tier.
Pay penetration dropped from 20% to 9%, but the free tier pulled in high school and college students, who became two-thirds of daily active users, according to RevenueCat's write-up.
Those students spread the app by word of mouth and led to a new schools product.
Opal's team calls the drop short-term; the long-term payoff is their claim, not a measured number.
The same Phil Carter write-up includes an established app that moved to freemium with a multistep paywall and saw +75% LTV per user.
That result came bundled with new messaging, so the soft paywall alone can't take all the credit.
Ryn VPN added a "Continue with Ads" option to its paywall and reported +158% revenue, with conversion going from 0.08% to 1.2% and ARPPU up 98% (vendor-reported by Superwall, before-after).
The test that warns against forcing it
BoldVoice A/B tested removing the close button entirely.
Trial starts went up, but refunds rose by more than the gain.
A second variant that kept the close button but showed a discount offer that couldn't be closed had the same result.
BoldVoice dropped both and now judges paywalls on net revenue after refunds per exposed user, on a matured cohort.
The pattern across these cases: hard paywalls tend to win when the app delivers clear value fast and users already expect to pay.
Soft paywalls win when free users create value of their own, such as Opal's student word of mouth or Ryn VPN's ad revenue.
Forcing people who aren't ready tends to come back as refunds.
Which app categories suit each
No category rule is absolute, but the tests and RevenueCat's hard paywall vs freemium guidance point to some fits.
Hard paywall tends to fit:
- Health, fitness, and meditation apps with a clear promise and a guided program. Headspace and V1 Sports both did better with less free access.
- Utilities that solve an acute problem, like scanners, VPNs, or photo editors, where the user arrives with a job to do now.
- Bootstrapped apps that need cash flow and don't have the analytics to manage a free funnel. RevenueCat lists both as reasons to stay hard.
- Apps with high per-user costs, such as AI features, where free users cost real money.
Soft paywall tends to fit:
- Social, communication, and collaboration apps with network effects, where every free user makes the app better for others.
- Apps with a viral or word-of-mouth loop, like Opal among students.
- Apps that can earn from ads, so a free user isn't a pure cost.
- Apps that have hit an organic growth ceiling with strong product-market fit and enough runway to wait for later conversion.
If you're unsure, the benchmarks favor starting hard.
RevenueCat's guidance describes freemium as a product strategy that needs "a lot more sophistication," not just a pricing switch.
How to test the switch
Moving between hard and soft paywalls is a big change, so test it like one. Here's a practical setup.
- Pick the right metric. Trial starts alone will mislead you, as BoldVoice found. Use net revenue after refunds per user who saw the paywall, measured on a cohort old enough to have renewed. Track retention and refunds alongside it.
- Randomize new users only. Existing users already have expectations. Split new installs between the hard and soft versions using your paywall tool or a feature flag.
- Wait for cohorts to mature. A soft paywall converts later, so a two-week readout favors the hard version. Plan to read results after at least the first renewal of your main plan.
- Count the free users' value. If free users drive referrals, ads, or invites, measure those too. Opal's case shows a lower pay rate can hide a bigger audience.
- Try hybrids before going all the way. A metered free tier, a "continue with ads" exit, or a hard paywall with a trial can capture much of the upside with less risk.
- Keep it honest. A close button that doesn't work, or an exit offer that can't be dismissed, can backfire in refunds and reviews. Make the subscription terms and the way out clear.
For a before-after switch instead of an A/B test, compare at least a few matching cohorts on each side, and watch for seasonality and marketing changes.
Several cases above are before-after, which is why their numbers are a signal rather than proof.
For more on running app experiments, browse all paywall tests, or start with the paywall guide.