Building a Subscription Paywall That Doesn't Feel Hostile
A client's original paywall design blocked every feature behind a 7-day trial signup screen shown immediately on first launch, before the user had done anything in the app. Conversion to trial start was decent at 18%, but trial-to-paid conversion was poor at under 4%, and App Store reviews were full of complaints about being asked to pay before understanding what the app even did.
We restructured around a value-first sequence: let users complete one full core action free — in this case, generating one real workout plan — before showing any paywall at all. Trial starts dropped slightly to around 14%, but trial-to-paid conversion nearly tripled to 11%, because the people starting trials had already experienced enough value to have a real reason to convert.
Pricing presentation mattered more than we expected going in. Showing three tiers — monthly, annual, and a lifetime option — with the annual plan visually marked as 'most popular' and displaying the effective monthly cost alongside the annual price, rather than just the lump sum, increased annual plan selection from about 20% to 45% of new subscribers, meaningfully improving revenue predictability for the client.
We avoided dark patterns specifically flagged in Apple's and Google's review guidelines and in general UX ethics — no pre-checked auto-renew upsells, no countdown timers implying false urgency, and a cancellation flow that's exactly as many taps as the subscription flow, not buried three menus deep. Apple has actually rejected submissions from us in the past for cancellation flows that were harder to find than the subscribe flow, so this is also a practical App Store review requirement now.
Post-cancellation, we added a short exit-survey step, optional, one tap to skip, asking why someone cancelled, which the client used to fix two real product gaps — missing platform support and a confusing pricing tier — rather than treating cancellation purely as lost revenue to win back with a discount offer. That data ended up being more valuable to the roadmap than most of our formal user research sessions that quarter.
