What's the Median vs P90 Onboarding Length Across Tracked Apps?

Across 129 tracked onboarding flows, the median is 11 screens and the p90 is 27 [1]. That means the longest 10% of onboardings are more than twice the length of the typical one. The average (13.9) sits between them because a small number of 30-70 screen flows drag it upward [1].

Median onboarding is 11 screens but p90 is 27 — the top decile runs 2.5x the typical flow (n=129, July 2026).

Lazyweb Research · n=129 · Published 2026-07-07

onboardingux-patternsmobilesaasdesign

The finding: a 16-screen gap between median and p90

The median tracked onboarding is 11 screens; the p90 is 27 [1]. The 16-screen spread between them shows how uneven onboarding length is across apps. For context, p25 is 6 and p75 is 17 [2], so the middle 50% of apps live in a fairly tight 6-17 band, and the divergence happens entirely in the top decile.

Where the long tail lives

BucketFlows
2-5 screens27 [3]
6-1036 [3]
11-1529 [3]
16-2013 [3]
21-3013 [3]
31+11 [3]

Of 129 flows, 11 exceed 30 screens [3]. These are concentrated in Health & Fitness, whose category median is 20.5 [4] — roughly double the overall median.

How to apply it

Report both numbers when you benchmark. The median (11) tells you what typical looks like; the p90 (27) tells you what a long flow looks like before it's a true outlier. If a stakeholder cites a 30-screen competitor onboarding as the norm, the data says only about 9% of tracked flows reach 30 [3] — it's an outlier, not a standard.

The numbers

StatComputed from
median 11; p90 27.2; avg 13.9onboarding_headline_distribution
p25 = 6; p75 = 17onboarding_headline_distribution percentiles
buckets: 2-5:27, 6-10:36, 11-15:29, 16-20:13, 21-30:13, 31+:11 (n=129); 12 flows >=30onboarding_length_histogram + is_N_too_many_thresholds >=30
Health & Fitness median 20.5 (n=26)category_health_fitness
Methodology. Universe: 129 canonical onboarding flows (125 companies) within 2,468 tracked flows. Percentiles and buckets computed over screens-per-canonical-flow. Caveat: right-skewed distribution, so median is the better central estimate than average.

Sources & citations

  1. [1] Lazyweb Research analysis of 129 canonical onboarding flows (subset of 2,468 tracked flows), July 2026. Median and p90 via PERCENTILE_CONT over screens-per-flow.
  2. [2] Lazyweb Research analysis of 129 canonical onboarding flows (subset of 2,468 tracked flows), July 2026. p25/p75 from the same distribution.
  3. [3] Lazyweb Research analysis of 129 canonical onboarding flows (subset of 2,468 tracked flows), July 2026. Length buckets and >=30 threshold count.
  4. [4] Lazyweb Research analysis of 26 Health & Fitness onboarding flows, July 2026. Per-category median; n=26 across 24 companies.

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