How Many Paywalls Show a Restore-Purchases Link or Close Button?
Of 1,204 described paywalls in Lazyweb's corpus, at least 314 show a close/dismiss button and at least 212 show a restore-purchases link[1]. Both are common escape and utility affordances, but neither appears on a majority of documented paywalls. Include a restore link for compliance and a considered dismiss control based on how hard you want the wall to be.
At least 314 of 1,204 described paywalls show a close button and 212 show a restore-purchases link (Lazyweb Research, July 2026).
The finding
Scanning 1,204 canonical paywall descriptions, a close/dismiss (X or 'close') control was detected on 314 paywalls and a restore-purchases link on 212[1]. Both are lower bounds, so actual prevalence is higher — but even at face value, close buttons appear on about a quarter and restore links on roughly a sixth of documented paywalls[1].
How to apply it
A restore-purchases link is effectively required on iOS subscription paywalls and appears on at least 212 screens[1]; include it as a standard footer link. The close button is a product decision: showing one (314 paywalls)[1] makes the wall softer and less coercive, while omitting it creates a harder wall. Match the choice to how much friction your funnel can absorb, and pair a visible close with the dismiss-role CTA data on the role-mix page.
Caveats
Both counts are lower bounds from regex over LLM vision descriptions — a restore link or X button can be present without being described, especially small tap targets[1]. Treat 314 and 212 as floors. These are structural counts, not statements about whether dismiss affordances change conversion.
The numbers
| Stat | Computed from |
|---|---|
| close/dismiss button 314; restore-purchases link 212 of 1,204 described paywalls (lower bounds) | misc_anatomy_elements: regex on vision descriptions |
Sources & citations
- [1] Lazyweb Research analysis of 1,204 described paywall screens (~800 tracked apps), July 2026. Close button and restore-purchases link detected via regex on LLM vision descriptions; both are lower bounds. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.