Are onboarding flows getting shorter in 2026?
Yes, modestly. Mean onboarding-flow length fell from 15.5 steps for flows first captured in 2023 to 12.4 in 2025 - about a 20% drop - with the median sliding from 12 to 10 [1]. The reliable comparison is 2023 (51 flows) versus 2025 (49 flows); the intervening 2024 and 2026 cohorts are too thin to lean on [2]. If you are cutting onboarding length, the trend supports you, but the shift is a few steps, not a wholesale collapse.
Mean onboarding length dropped from 15.5 steps (2023 flows) to 12.4 (2025) - about 20% shorter (July 2026).
The trend: about 3 fewer steps
Each onboarding flow is dated by the first-capture date of its member screens:
| First-captured year | Flows | Mean steps | Median steps |
|---|---|---|---|
| 2023 | 51 | 15.5 | 12 [1] |
| 2024 | 15 | 15.4 | 12 [2] |
| 2025 | 49 | 12.4 | 10 [1] |
| 2026 (partial) | 14 | 11.4 | 11 [2] |
Mean length is roughly flat 2023-2024 then drops about 20% in 2025 - a modest, real shortening [1].
Read the sample sizes carefully
Only 2023 (51 flows) and 2025 (49 flows) clear a healthy N; 2024 (15 flows) and partial-2026 (14 flows) are below N=20 and partly capture-driven, so treat the 15.4 and 11.4 means as directional at best [2]. The honest headline is the 2023-vs-2025 comparison: 15.5 to 12.4 mean, 12 to 10 median.
How to apply it
The market drifted toward leaner onboarding, so a 10-12 step flow is now nearer the norm than the 15-step flows of 2023 - useful cover if you are proposing cuts. But 'shorter' here means a couple of steps, not single-screen onboarding; the median is still 10 steps in 2025 [1]. Use it as a directional benchmark, not a target - pair any length change with an activation-rate test rather than optimizing step count in isolation.
Caveats
Flows are identified by %onboarding% in the flow name and dated by the minimum member-screen capture date [1]. 2024 and 2026 fall below N=20 and are labeled thin; 2026 is a partial year through July 4 [2]. Do not slice this trend by category - splitting the cohort drops most cells below usable N [3].
The numbers
| Stat | Computed from |
|---|---|
| Mean 15.5 (2023, 51 flows) -> 12.4 (2025, 49 flows); median 12 -> 10 | onboarding_length_over_time |
| 2024 = 15.4 mean (15 flows, N<20); 2026 partial = 11.4 mean (14 flows, N<20) | onboarding_length_over_time / smallSampleWarnings |
| Per-category time cuts unsupported - splitting cohorts drops cells below N=8 | smallSampleWarnings |
Sources & citations
- [1] Lazyweb Research analysis of 129 onboarding flows (mobile-app corpus), July 2026. Each onboarding flow (`%onboarding%` name) dated by MIN member-screen capture date; mean and median step counts by year. ↩
- [2] Lazyweb Research analysis of 129 onboarding flows (mobile-app corpus), July 2026. 2024 (15 flows) and partial-2026 (14 flows) are below N=20 and treated as thin/directional. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.