What Is Flo A/B Testing In Its Health App?
- By Ali Abouelatta
- n = 52
- Published
- Updated
- 2 min read
- retention · experiments · mobile · ux-patterns · design
Short answer
Lazyweb Research detected 52 distinct experiments at Flo (July 2026), one of the highest totals among women's-health apps in the corpus, with a detected paywall footprint of at least 6. source[1] The bulk of Flo's detected activity sits on unlabeled non-paywall surfaces, signaling engagement-led iteration. These are observed before/after variations with inferred rationale, not company-confirmed A/B tests.
Lazyweb Research detected 52 Flo experiments (July 2026), among the most of any health-tracking app tracked.
The finding
Lazyweb Research detected 52 distinct experiments at Flo. source[1] Only 6 are on the paywall surface, so the visible iteration is concentrated on the core tracking and content experience rather than monetization. Flo is a useful benchmark for a health-tracking app that experiments heavily on engagement.
How to apply it
If you run a health-tracking app, Flo's profile — high total volume, small paywall share — is evidence that most detectable iteration in this category happens on the core loop, not the paywall. Benchmark your own paywall-to-total ratio against it before assuming monetization is where the testing energy goes. Only 1 Flo experiment was detected in 2026, so this is a cumulative signal. source[1]
Caveats
All figures are observed variations with LLM-inferred rationale, not company-confirmed A/B tests — no lift is measured. source[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 |
|---|---|---|
| [1] | 52 distinct experiments; at least 6 paywall | company_total:flo (value 52; paywall 6, in-2026 1) |
| 02 | 1,425 of 4,814 experiments have no screen categorycat_null | screen_category_null_on_experiments (1425/4814) |
Methodology
Universe: 52 distinct Flo experiments (COUNT(DISTINCT experiment_id)) within 4,814 detected before/after UI diffs across 276 companies, July 2026. Extraction: LLM-inferred rationale on observed variations. Caveat: detected variations only, never company-confirmed A/B tests.
Sources
- [1]
Lazyweb Research analysis of 52 detected experiments (Flo, ~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.