Which Apps Have Run The Most UI Experiments In 2026?
- By Ali Abouelatta
- n = 4,814
- Published
- Updated
- 2 min read
- experiments · mobile · saas · ux-patterns · monetization
Short answer
Of 4,814 detected experiments (July 2026), 635 were detected in calendar 2026 — just 13% of the corpus. source[1] Finch leads with 30 experiments (all in 2026), followed by ChatGPT (12) and Babbel (8); most companies had single-digit 2026 activity. source[2] These are observed variations, not confirmed A/B tests.
Only 635 of 4,814 detected experiments (13%) landed in 2026; Finch alone accounts for 30 of them (July 2026).
Bar chart · Experiments detected in 2026
Experiments detected in 2026 — The 2026 leaderboard
One block is about 1.1 Experiments detected in 2026
- Finch
- 30 (all detected in 2026) [2]
- ChatGPT
- 12 [2]
- Babbel
- 8 [2]
- AllTrails
- 4 [2]
- Bluesky
- 4 [2]
- Gauthmath
- 4 [2]
- Breathwrk
- 4 [2]
- FaceApp
- 4 [2]
The finding
Of the 4,814 detected experiments, only 635 were detected in calendar 2026 — about 13% of the corpus. source[1] That means the vast majority of detected activity (4,179 experiments) predates 2026, and "what did company X change this year" is only honestly answerable for a handful of apps. source[1] Finch dominates the current year with 30 experiments, all in 2026. source[2]
The 2026 leaderboard
The current-year leaders are few; below them, 2026 counts drop into low single digits.
How to apply it
When a stakeholder asks "who is actively iterating right now," this is the honest shortlist: Finch and ChatGPT are the only two companies with a defensible double-digit 2026 volume. source[2] For everyone else, frame findings as "most recent detected changes" with explicit dates rather than a current-year claim, because 87% of the corpus predates 2026. source[1] This is the single most important guardrail for any "this year" page in the family. All figures are detected variations. source[1]
Caveats
All figures are observed variations with LLM-inferred rationale, not company-confirmed A/B tests — no lift is measured. source[1] Detection date reflects when Lazyweb captured the diff, not necessarily when the company shipped it. source[1] Companies with fewer than 8 experiments in 2026 cannot support a standalone current-year page. source[2]
The numbers
| # | Stat | Computed from |
|---|---|---|
| [1] | 635 of 4,814 experiments in 2026 (13%); 4,179 predate 2026 | experiments_2026 (635/4814); total_detected_experiments (4814) |
| [2] | 2026 leaders: Finch 30, ChatGPT 12, Babbel 8, AllTrails 4, Bluesky 4, Gauthmath 4, Breathwrk 4, FaceApp 4 | company_total in-2026 splits: finch 30, chatgpt 12, babbel 8, alltrails 4, bluesky 4, gauthmath 4, breathwrk 4, faceapp 4 |
Methodology
Universe: 4,814 detected experiments across 276 companies, July 2026; 635 in calendar 2026. Extraction: COUNT DISTINCT experiment_id by year. Caveat: detection date is capture date, not ship date; detected variations only.
Sources
- [1]
Lazyweb Research analysis of 4,814 detected experiments (276 companies, ~800-app mobile corpus), July 2026
635 experiments have experiment_created_at >= 2026-01-01.
- [2]
Lazyweb Research analysis of 4,814 detected experiments (276 companies), July 2026
Per-company in-2026 counts (COUNT DISTINCT experiment_id, experiment_created_at >= 2026-01-01).
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.