What pricing-page experiments have SaaS companies actually run?

Lazyweb detected 301 pricing-area experiments (254 mobile plan-picker diffs + 47 web pricing-table diffs), our model scoring web PRICING TABLE changes highest for impact (3.60 of 5 vs 3.52 for mobile PRICING) [1]. The recurring plays are tier restructuring — collapsing to a single hero price, adding a $0 floor, deleting a mid tier, splitting into audience tracks — plus billing-cadence and badge moves [1]. Every one is a detected before/after diff with inferred rationale, not a confirmed A/B win.

Lazyweb detected 301 pricing-area experiments (254 mobile + 47 web) in July 2026; web pricing-table changes scored highest impact at 3.60/5.

Lazyweb Research · n=301 · Published 2026-07-07

pricingmonetizationexperimentssaaswebmobile

The tested-area landscape

The 301 pricing-area experiments split by platform and carry model-assigned impact/confidence scores (1–5, for relative ranking only) [1]:

AreaPlatformCountAvg impactAvg confidence
PRICINGmobile2543.523.76
PRICING TABLEweb473.603.51

Web pricing-table changes edge out mobile on impact but with lower confidence — a thin, higher-variance set [1]. These scores rank experiments against each other; they are not measured lift.

The recurring plays (named before/after diffs)

The highest-scored detected pricing-table diffs, with the change and inferred rationale [1]:

PlayObserved changeInferred rationale
Collapse to hero price3-tier grid → single '$8/mo' headline + one CTARemove the downgrade escape hatch and plan-evaluation step
Widen entry with $03 paid tiers → Free $0 · $5 · $15 · $30 · Custom staircaseCapture hobbyists instead of bouncing at a $10 floor
Split into tracks3 per-person tiers → individual + team tracks, each with a free tierStop solo users parsing '/person' pricing; drop top anchor $21→$10
Remove the mid tier4 tiers → 3 (delete $29.99 'Business')Spare a three-way comparison; route agencies to sales
Trim flanks, hold anchorHobbyist $16→$12, Business $50→$40, keep 'Most Popular' Creator $24Lower entry barrier without discounting the anchor tier
Repriced per-seatGrowth $297.50/mo flat (5 seats) → $42.50/mo per userLet a small team start near $85 instead of a ~$300 anchor

Company attribution for these diffs is inferred from the pricing details in each summary, not a verified join [1].

How to apply this

Use this as a tested-play menu, not a leaderboard of proven wins. If your funnel problem is decision paralysis, the tested moves are collapse-to-hero-price or delete-the-mid-tier. If it's entry too expensive, it's add-a-$0-floor or trim-the-flanking-tiers. If it's wrong buyer confusion, it's split-into-individual-and-team-tracks. If it's a team pricing wall, it's switch-flat-bundle-to-per-seat [1]. Each was scored high-impact by our model but remains a detected diff with inferred rationale — reproduce it as your own A/B before trusting the direction.

The numbers

StatComputed from
301_experiment_annotations WHERE area IN ('PRICING','PRICING TABLE') — 254 mobile + 47 web
254, impact 3.52, conf 3.76PRICING area mobile experiments with model-assigned impact/confidence (1-5)
47, impact 3.60, conf 3.51PRICING TABLE area web experiments with model-assigned impact/confidence (1-5)
Methodology. Universe: 301 detected pricing-area experiments (254 mobile, 47 web), Lazyweb Research, July 2026. Experiments are detected before/after UI diffs with LLM-inferred rationale and model-assigned impact/confidence scores — relative ranking only, not measured A/B lift. Company names are inferred from summary text.

Sources & citations

  1. [1] Lazyweb Research analysis of 301 pricing-area experiments (detected before/after UI diffs), July 2026. 254 mobile PRICING + 47 web PRICING TABLE annotations. Impact/confidence are model-assigned 1-5 scores for relative ranking only, never measured lift. Company attribution is inferred from summary text, not a verified join.

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