Where do web products concentrate their experiments, and what does that mean for my roadmap?

Across 880 web experiment annotations, testing concentrates on the hero (272), navigation (144), CTA (91), value props (82), and announcement bar (63), with pricing tables (47) and forms (37) close behind ^[1]. Notably, web is far more heavily annotated per experiment than mobile — 2.4 annotations per web experiment vs 0.29 per mobile ^[2]. Web testing is a messaging-and-structure game, not a single paywall button.

Web experiments carry 2.4 area annotations each vs 0.29 on mobile — Lazyweb Research, July 2026.

Lazyweb Research · n=365 · Published 2026-07-07

experimentsweblanding-pagepricingdesign

Finding: hero, nav, and pricing structure dominate web tests

Web experiment annotations concentrate on the top of the page and its structure:

Web area	Annotations
HERO	272
NAV	144
CTA	91
VALUE PROPS / FEATURES	82
ANNOUNCEMENT BAR	63
PRICING TABLE	47
FORM	37
SOCIAL PROOF	34
SECONDARY ACTIONS	33
LEGAL/TRUST	19

Web also gets richer annotation coverage: 880 annotations from 365 experiments (2.4 each) versus 1,280 from 4,449 mobile experiments (0.29 each) ^[1]^[2].

How to apply this

For a web product, weight your test backlog toward hero messaging, navigation, and value-prop clarity first — that is where the corpus concentrates and where a category leader like Calendly reframed its own positioning ('More than a scheduling link') ^[3]. Pricing-table restructuring is a real but lower-volume lever. Announcement bars and forms are web-native test areas with no mobile-paywall equivalent, so borrow those ideas from web peers, not from app paywalls.

Caveats

Web experiments (n=365) are ~12x fewer than mobile; these are absolute annotation counts, not rates implying parity ^[2]. All web experiments share one placeholder canonical screen, so you cannot compute where on the page each test sat ^[4]. The Calendly hero example is a detected before/after change with inferred rationale, not a measured lift ^[3].

The numbers

Stat	Computed from
Web area annotations: HERO 272, NAV 144, CTA 91, VALUE PROPS 82, ANNOUNCEMENT BAR 63, PRICING TABLE 47, FORM 37	experiment_area_mix_by_platform (web)
880 web annotations from 365 experiments (2.4/exp) vs 1,280 mobile from 4,449 (0.29/exp)	annotation_platform_split + experiment_platform_split
Calendly hero reframed from 'Book meetings that work for you' to 'More than a scheduling link'	qualitative: web HERO Calendly reframe
365 web experiments all share placeholder canonical screen id 23588	web_experiments_not_screen_anchored

Methodology. Universe: 365 web experiments with 880 area annotations, July 2026. Method: count annotations by area; compare annotation density across platforms. Caveat: absolute counts with thin web n; detected changes with inferred rationale, not measured lift.

Sources & citations

[1] Lazyweb Research analysis of 880 web experiment annotations (web corpus, 365 experiments), July 2026. Area mix for web experiments. ↩
[2] Lazyweb Research analysis of 4,814 detected experiments and 2,160 annotations (mobile + web corpora), July 2026. Web is annotated 2.4x per experiment vs 0.29 on mobile. ↩
[3] Lazyweb Research analysis of web hero experiments (web corpus), July 2026. Calendly hero reframe; detected change with inferred rationale. ↩
[4] Lazyweb Research analysis of 365 web detected experiments (web corpus), July 2026. All web experiments share one placeholder canonical screen; not screen-anchored. ↩

Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.