# Dark or Light: What Background Do Mobile Paywalls Actually Use?

Source: Lazyweb Research
Author: Ali Abouelatta, Lazyweb Research
Published: 2026-07-07
Updated: July 2026
Sample size: n=1107
Category: [Paywall anatomy and plan-structure benchmarks](https://design.lazyweb.com/blog/category/paywall-anatomy-and-plan-structure-benchmarks)
Tags: paywall, design, monetization, mobile, ux-patterns
HTML: https://design.lazyweb.com/blog/dark-vs-light-paywall-background-benchmark
Markdown: https://design.lazyweb.com/blog/dark-vs-light-paywall-background-benchmark.md

**Answer.** Across 1,107 mobile paywall screens with a classified background in Lazyweb's corpus, 70.9% use a dark background, 19.1% light, and 10.0% colored or other[1]. Dark is the clear default for paywalls, but the split varies sharply by category. Treat dark as the safe convention and light as a deliberate, brand-led exception rather than the norm.

> 70.9% of 1,107 classified mobile paywalls use a dark background vs 19.1% light (Lazyweb Research, July 2026).

## The finding

Of 1,204 canonical paywall screens with a vision description, 1,107 had a classifiable background field[1]. Dark backgrounds dominate at 785 screens (70.9%), light at 211 (19.1%), and colored/other at 111 (10.0%)[1]. The remaining 97 canonicals lacked a background field and are excluded from the percentages[2]. So among paywalls where we can read the background, dark outnumbers light nearly 4 to 1.

## How to apply it

If you are shipping a paywall with no strong brand reason to go light, dark is the convention your users have been trained on: roughly 7 in 10 paywalls they have seen are dark. Light backgrounds cluster in specific verticals (see the Health & Fitness cut, where the split narrows to 56% dark). Use light when your brand system is light-first and the paywall must feel continuous with the rest of the app, not because it converts better in the abstract.

## Caveats

Background is read from a single representative screenshot per canonical (latest with vision JSON), classified from the vision 'background' field[2]. The field is missing on 97 of 1,204 described paywalls, so the denominator is 1,107, not the full 1,297 corpus[1]. This is a structural/aesthetic census, not an outcome measurement: we are not claiming dark converts better, only that it is far more common.

## The numbers

| Stat | Computed from |
| --- | --- |
| dark 785 (70.9%), light 211 (19.1%), other 111 (10.0%) of 1,107 classified | dark_vs_light_overall: representative screenshot per canonical, background field from vision JSON |
| 1,204 described paywalls; 97 lack a background field; 1,107 classified | dark_vs_light_overall denominator note |

## Methodology

Universe: 1,204 canonical paywall screens with a vision description across ~800 tracked apps; background classified from the vision JSON background field on one representative screenshot per canonical, July 2026. Caveat: 97 canonicals lack the field and are excluded from percentages.

## Visual references

Real, in-market app screens tracked by Lazyweb Research:

- Disney: https://zlfyzdmohcskkucuunmk.supabase.co/storage/v1/object/public/screenshots/ent6_disney/lazyscreenshots/2023_12_16_11-33-58_96F3C97D.png
- Nyt Games: https://zlfyzdmohcskkucuunmk.supabase.co/storage/v1/object/public/screenshots/uploaded_nyt_games/paywall/compare/2026-05-29/1780027702930_00_NYTGames-20260529-040822-01.png
- Lockwidget: https://zlfyzdmohcskkucuunmk.supabase.co/storage/v1/object/public/screenshots/graphics8_lockwidget/backlog/2025_08_30_22-34-31_6F6833F3.png
- Headway: https://zlfyzdmohcskkucuunmk.supabase.co/storage/v1/object/public/screenshots/edu17_headway/lazyscreenshots/2024_07_06_13-41-11_0F76CB5C.png

## Sources & citations

- [1] Lazyweb Research analysis of 1,107 classified paywall screens (~800 tracked apps), July 2026. One representative screenshot per sc_canonical_id, background classified from the vision_description_json background field.
- [2] Lazyweb Research analysis of 1,204 described paywall screens (~800 tracked apps), July 2026. Background field present on 1,107 of 1,204 canonicals with a vision description.

## Related questions

- [Which App Categories Use Light Paywalls Instead of Dark?](https://design.lazyweb.com/blog/dark-vs-light-paywall-background-by-category)
- [Should a Health and Fitness Paywall Use a Dark or Light Background?](https://design.lazyweb.com/blog/health-fitness-paywall-dark-or-light-background)
- [How Many CTAs Does the Median Mobile Paywall Have?](https://design.lazyweb.com/blog/how-many-ctas-median-paywall-has)
- [How Many Plan Options Do Mobile Paywalls Show — 1, 2, or 3?](https://design.lazyweb.com/blog/how-many-plan-options-do-mobile-paywalls-show)
- [What % of Paywalls Pre-Select the Annual Plan?](https://design.lazyweb.com/blog/what-percent-of-paywalls-preselect-the-annual-plan)
- [What % of Paywalls Show Social Proof (Ratings, Member Counts)?](https://design.lazyweb.com/blog/what-percent-of-paywalls-show-social-proof)
- [Do Paywalls Use Feature Checklists or Benefit Carousels?](https://design.lazyweb.com/blog/do-paywalls-use-feature-checklists-or-benefit-carousels)
- [What % of Paywalls Mention a Free Trial?](https://design.lazyweb.com/blog/what-percent-of-paywalls-mention-a-free-trial)

## Explore the evidence

- [Cal AI app tree](https://design.lazyweb.com/company/cal-ai/app-tree) — Full 45-frame August 2026 capture
- [Cal AI trial introduction](https://design.lazyweb.com/flow/cal-ai/onboarding-paywall) — Three captured screens · August 2026
- [Cal AI: sign-in, trial, then reminder](https://design.lazyweb.com/blog/cal-ai-sign-in-trial-reminder-sequence)
- [Paywall before or after account creation?](https://design.lazyweb.com/blog/do-apps-paywall-before-or-after-account-creation)

## Continue your research

[Connect Lazyweb to your coding agent](https://www.lazyweb.com/mcp-install)