How Do Apps Let Users Exit The Paywall — X Button, 'Maybe Later', Or Not At All?
Across 2,708 mobile paywall screens tracked by Lazyweb Research, the most common visible exit is an X/close button (517 screens, 120 companies), followed by a 'Maybe later'-style text link (193 screens, 62 companies) and a restore-only escape (151 screens, 56 companies); 1,847 screens (68.2%) offer no visible exit [1]. When apps do give a way out, a tappable X outnumbers a text link by about 2.7 to 1 [1].
Among 2,708 paywall screens, a visible X (517) outnumbers a 'Maybe later' text link (193) by roughly 2.7 to 1 — Lazyweb Research, July 2026.
The exit taxonomy
Each of 2,708 screens is placed in exactly one bucket, with a visible X taking precedence [1]:
| Exit type | Screens | Companies | Share |
|---|---|---|---|
| Visible X / close button | 517 | 120 | 19.1% |
| 'Maybe later' text link | 193 | 62 | 7.1% |
| Restore-only escape | 151 | 56 | 5.6% |
| No visible exit (hard) | 1,847 | 229 | 68.2% |
The three visible-exit buckets sum to 861 screens (31.8%) [1].
Where the X signal comes from
The visible-X count is driven by vision descriptions, not button text: only 7 CTA rows are literally 'X' or 'close', while 601 of 2,706 screens with a vision description (22.2%) have a close/X control explicitly described [2]. That is why every visible-X figure should be read as an extraction-based lower bound [2].
How to apply it and caveats
If you want the lightest-touch soft paywall, the corpus default is a corner X — it is the single most common exit and reads as standard iOS/Android behavior [1]. A 'Maybe later' link is the second option and appears on 62 companies [1]. Restore-only is not really an exit for new users; it is a purchase-recovery link that happens to be the only escape on 151 screens [1]. Caveat: these are what-shipped counts, never lift measurements [1].
The numbers
| Stat | Computed from |
|---|---|
| X 517 (120 cos) | text link 193 (62) | restore-only 151 (56) | no exit 1,847 (229) | exit_type_taxonomy |
| 601 of 2,706 vision-described screens (22.2%); only 7 literal X CTA rows | vision_close_mentions: 601/2,706; dismiss_copy_leaderboard literal X/close = 7 |
Sources & citations
- [1] Lazyweb Research analysis of 2,708 paywall screens (mobile app corpus, 252 companies), July 2026. Mutually exclusive exit taxonomy with visible-X precedence; extraction-based lower bounds. ↩
- [2] Lazyweb Research analysis of 2,706 vision-described paywall screens (mobile app corpus), July 2026. Close/X control detected via vision-description regex; CTA text almost never contains a literal X. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.