Life · Infinity Games
200+ paywall tests, and the one that won
The old paywall showed up before the value did. I treated every paywall like an experiment, and ran over 200 of them.
DraftPublished Updated 2 min read
This is a working draft. I’m still filling it in.
+23%
Revenue from one paywall
+40%
Paid growth on iOS
20×
In-app purchases
37%
Of monthly users on paid plans
On this page
How I tested: watch, test, tweak
I treated every important flow like something I could test, watch and tweak, one experiment at a time.
Every flow ran on the same loop: guess, test, learn, change. We ran it on acquisition, onboarding, engagement and the paywall.
Watch what people do
Mixpanel funnels showed where women dropped off, flow by flow.
Test one change
Every drop-off became one focused A/B test, from onboarding steps to paywall copy.
Nail the first impression
App Store tests tried out screenshots and messages before anyone even installed.
Listen
200+ interviews and threads about stress, mood and cycles showed us where our words didn't land.
This loop kept us honest. A change only shipped if it moved a number, and the interviews told me why it moved. I built the winning ideas into patterns we could reuse and A/B test again.
200+ versions, each one a small bet
The paywalls that won were rarely the prettiest. They said the value out loud, right when a woman was already feeling it.
I tested headlines, prices, plan options, buttons, illustrations, even punctuation. Each version was a small bet I could measure.
I planned a paywall for each step of the journey and priced each one to the moment. Onboarding offered a free trial. If she closed it, she saw a lower price. The home screen ran a banner, generic or seasonal. Locked premium features had their own paywall. When she came back to the app, the price matched that visit.

Onboarding

Upgrade
The winner: Easter Surprise
One paywall beat every other version on revenue: Easter Surprise bumped it 23%.
Four things made it work:
- A simple, direct message.
- It went live over Easter weekend.
- An image that felt real.
- A short offer with a deadline.
It offered half price for good, but only for a few days. If you tried to close it, a sad bunny asked: “Are you sure you don’t want a discount?”

Together with the onboarding and pricing tests, the paywall work brought +40% paid growth on iOS and 20× in-app purchases. 37% of monthly users were on paid plans.
What I learned
- Better-looking designs didn’t always win. In several tests, the less polished one beat the clean one. Paywalls win on clarity and timing.
- Small bets you can measure add up. 200 tests taught me more than any big redesign.
Topics: experimentation, growth, data. Also as plain markdown.