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.

Lazyweb Research · n=2160 · Published 2026-07-07

webmobileexperimentspaywalllanding-pageux-patterns

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]:

AreaMobileWebSkew
NAV0144web-only
FORM037web-only
ANNOUNCEMENT BAR063web-only
OFFER25546~85% mobile
PRICING25447~84% mobile
HERO216272mixed
CTA14191mostly mobile
SOCIAL PROOF1434web-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

StatComputed 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 mobileplatform_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 mobilearea_HERO; area_CTA; area_SOCIAL_PROOF
Detected experiments: mobile 4,449 (92.4%) / web 365 (7.6%) of 4,814platform_split_experiments
Methodology. Universe is 2,160 area annotations (1,280 mobile / 880 web) over 1,126 detected experiments within 4,814 detected experiments in the ~800 tracked-apps corpus. Splits are annotation counts from LLM-annotated before/after screenshots, July 2026; web is thin, so web cuts are absolute counts only.

Sources & citations

  1. [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. [2] Lazyweb Research analysis of platform-by-area splits (~800 tracked apps), July 2026. Per-area mobile/web counts within 2,160 annotations.
  3. [3] Lazyweb Research analysis of per-area platform counts (~800 tracked apps), July 2026. HERO/CTA/SOCIAL PROOF web-mobile breakdowns.
  4. [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.

Related questions

Explore the underlying screens, flows, and A/B tests inside Lazyweb. More research