Web Vs Mobile: Which Page Areas Do Companies A/B Test On Each?
Across 2,160 area-annotated experiments tracked by Lazyweb Research (1,280 mobile / 880 web), the tested surfaces split cleanly by platform: NAV, FORM, and ANNOUNCEMENT BAR are essentially 100% web, while OFFER (85% mobile), PRICING (84% mobile), and HEADER (100% mobile) are paywall-driven mobile areas [1][2]. The hero is the one area both platforms test heavily, splitting 272 web / 216 mobile [3]. Detected experiments overall are 92% mobile (4,449 of 4,814), so web reads are absolute counts only [4].
NAV, FORM and ANNOUNCEMENT BAR are 100% web while OFFER and PRICING run ~85% mobile — Lazyweb Research analysis of 2,160 annotated experiments (1,280 mobile / 880 web), July 2026.
The finding: platforms test different real estate
Of the 2,160 annotations, 1,280 are mobile (from 147 companies) and 880 are web (from 106 companies) [1]. The area mix differs sharply by platform: web owns NAV (all 144), FORM (all 37), and ANNOUNCEMENT BAR (all 63), while mobile owns the monetization furniture — OFFER runs 255 mobile / 46 web, PRICING 254 mobile / 47 web, and HEADER is 100% mobile (29) [2][3]. That reflects what each surface is for: mobile experiments concentrate on paywalls, web experiments on marketing pages [2].
Where each platform concentrates
Selected areas by platform split [2][3]:
| Area | Mobile | Web | Skew |
|---|---|---|---|
| NAV | 0 | 144 | web-only |
| FORM | 0 | 37 | web-only |
| ANNOUNCEMENT BAR | 0 | 63 | web-only |
| OFFER | 255 | 46 | ~85% mobile |
| PRICING | 254 | 47 | ~84% mobile |
| HERO | 216 | 272 | mixed |
| CTA | 141 | 91 | mostly mobile |
| SOCIAL PROOF | 14 | 34 | web-leaning |
HERO is the shared battleground, tested heavily on both [3].
How to apply it
Match your test backlog to your platform's tested surfaces: on mobile, the corpus concentrates on offer framing and pricing inside the paywall; on web, on nav de-chroming, hero, and signup forms [2][3]. If you run both, the hero is the one area worth testing on each — it is the only high-volume area split roughly evenly [3]. Don't force web-only patterns (announcement bars, nav cuts) onto a mobile paywall, or vice versa [2].
Caveats
Detected experiments are 92.4% mobile (4,449) vs 7.6% web (365), so the web corpus is thin and its area cuts are absolute counts only, not percentage league tables [4]. Platform splits are annotation counts, not lift [4]. Company counts overlap across platforms [1].
The numbers
| Stat | Computed from |
|---|---|
| 2,160 annotations = 1,280 mobile (147 companies) + 880 web (106 companies) | platform_split_annotations; universe |
| Web-only areas: NAV (144), FORM (37), ANNOUNCEMENT BAR (63). Mobile-heavy: OFFER 255/46, PRICING 254/47, HEADER 29 all mobile | platform_area_split; area_NAV; area_FORM; area_ANNOUNCEMENT_BAR; area_OFFER; area_PRICING; area_HEADER |
| HERO 272 web / 216 mobile; CTA 141 mobile / 91 web; SOCIAL PROOF 34 web / 14 mobile | area_HERO; area_CTA; area_SOCIAL_PROOF |
| Detected experiments: mobile 4,449 (92.4%) / web 365 (7.6%) of 4,814 | platform_split_experiments |
Sources & citations
- [1] Lazyweb Research analysis of 2,160 annotated experiments (~800 tracked apps), July 2026. Platform split of area annotations; 1,280 mobile / 880 web. ↩
- [2] Lazyweb Research analysis of platform-by-area splits (~800 tracked apps), July 2026. Per-area mobile/web counts within 2,160 annotations. ↩
- [3] Lazyweb Research analysis of per-area platform counts (~800 tracked apps), July 2026. HERO/CTA/SOCIAL PROOF web-mobile breakdowns. ↩
- [4] Lazyweb Research analysis of 4,814 detected experiments (~800 tracked apps), July 2026. Web is thin (365 experiments); web cuts are absolute counts only. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.