# What % of Paywalls Have a Visible Close or Restore Option?

Source: Lazyweb Research
Published: 2026-07-07
Sample size: n=1204
Tags: paywall, cancellation, ux-patterns, monetization, mobile
HTML: https://www.lazyweb.com/research/what-percent-of-paywalls-have-a-visible-close-button
Markdown: https://www.lazyweb.com/research/what-percent-of-paywalls-have-a-visible-close-button.md

**Answer.** At least 314 of 1,204 described mobile paywalls show a close or dismiss button, and at least 212 show a restore-purchases link, in Lazyweb's corpus.[1] Both are minority-described affordances, meaning many paywalls either omit them or place them where the vision description did not capture them. If you are deciding whether to show an escape hatch or a restore link, both are common enough to be standard but far from universal in the data.

> At least 314 of 1,204 described mobile paywalls show a close/dismiss button and at least 212 show a restore-purchases link (Lazyweb Research, July 2026).

## The finding

Detected in vision descriptions across 1,204 paywalls: a close/dismiss/X button on at least 314 screens, and a restore-purchases link on at least 212.[1] Both are lower bounds — small top-corner X marks and footer restore links are exactly the elements a description can miss — so read these as floors, not ceilings.[2] Restore links and close buttons are both established paywall furniture, but neither shows up on a majority of described paywalls in the raw counts.[1]

## Detected anatomy elements

| Element | Paywalls (of 1,204) |
|---|---|
| Close / dismiss / X button | 314 |
| Restore-purchases link | 212 |

Both are lower bounds from vision-description regex.[1]

## How to apply it

A visible close button is the difference between a soft paywall (dismissible) and a hard one, and at least 314 paywalls show one — but the rest either use a hard paywall or hide the exit.[1] A restore-purchases link (at least 212) is effectively required for App Store compliance and to avoid support tickets from returning subscribers, so its true incidence is almost certainly higher than the described floor.[2] Include a restore link by default; treat the close button as a hard-vs-soft-paywall product decision rather than a given.[1]

## Caveats

Both figures are lower bounds from regex over LLM vision descriptions and use the representative-screenshot method.[1] A restore link or a corner X can be present on screen without being described, so the true rates are higher — do not read 314/1,204 or 212/1,204 as precise incidence.[2]

## The numbers

| Stat | Computed from |
| --- | --- |
| close/dismiss button on 314 paywalls; restore link on 212 (of 1,204) | misc_anatomy_elements: close button 314, restore link 212, n=1,204 |
| lower bounds via description regex | misc_anatomy_elements description and universe note: text-pattern stats are lower bounds (regex over vision descriptions) |

## Methodology

Universe: 1,204 mobile paywalls with vision descriptions across ~800 tracked apps. Close/dismiss and restore elements detected by regex over LLM vision descriptions. July 2026 pull. Key caveat: lower bounds — small corner and footer elements are undercounted, so true incidence is higher.

## Sources & citations

- [1] Lazyweb Research analysis of 1,204 described paywalls (~800 tracked apps), July 2026. Close/dismiss buttons and restore-purchases links detected via regex over LLM vision descriptions (representative-screenshot method); both lower bounds.

## Related questions

- [How Many CTAs Does the Median Paywall Have?](https://www.lazyweb.com/research/how-many-ctas-does-the-median-paywall-have)
- [Dark or Light: What Background Do Mobile Paywalls Use?](https://www.lazyweb.com/research/dark-or-light-what-background-do-mobile-paywalls-use)
- [What Is the Most Common Primary Paywall CTA Button Text?](https://www.lazyweb.com/research/what-is-the-most-common-primary-paywall-cta-text)