What Is Crumbl A/B Testing On Its Home Screen?
Lazyweb Research detected 43 distinct experiments at Crumbl (July 2026), of which at least 39 are on the home surface — one of the most home-concentrated test programs in the corpus. [1] With no detected paywall experiments, Crumbl's iteration is almost entirely on merchandising the weekly menu, not monetization. These are observed before/after variations with inferred rationale, not company-confirmed A/B tests.
Lazyweb Research detected 43 Crumbl experiments (July 2026), at least 39 of them on the home screen.
The finding
Lazyweb Research detected 43 distinct experiments at Crumbl, with at least 39 on the home surface and 0 on any paywall. [1] Crumbl is the corpus's clearest example of a food-ordering app treating its home screen — where the rotating weekly menu lives — as the single most-tested surface. Nearly every detected experiment is about merchandising, not checkout or subscription.
How to apply it
Crumbl is the benchmark for a menu-driven ordering app: when the product rotates weekly, the home/merchandising surface is where iteration concentrates (at least 39 of 43 detected experiments). If your ordering app tests checkout heavily but leaves the home menu static, Crumbl is evidence the merchandising surface is the higher-leverage test bed. No Crumbl experiments were detected in 2026. [1]
Caveats
All figures are observed variations with LLM-inferred rationale, not company-confirmed A/B tests — no lift is measured. [1] Surface splits are lower bounds because screen category is unlabeled on 1,425 of 4,814 corpus experiments. [cat_null]
The numbers
| Stat | Computed from |
|---|---|
| 43 distinct experiments; at least 39 home | company_total:crumbl (value 43; home 39, signup 3, in-2026 0) |
| 1,425 of 4,814 experiments have no screen category | screen_category_null_on_experiments (1425/4814) |
Sources & citations
- [1] Lazyweb Research analysis of 43 detected experiments (Crumbl, ~800-app mobile corpus), July 2026. COUNT(DISTINCT experiment_id) on before/after diffs; surface splits from is_paywall + screen_category. ↩
- [cat_null] Lazyweb Research analysis of 4,814 detected experiments (276 companies, ~800-app mobile corpus), July 2026. screen_category is NULL on 1,425 experiments, so all surface splits are lower bounds. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.