Session length is the time a user spends in the app per session, measured from app open to app close (or background after an inactivity timeout). Industry-standard inactivity timeout is 30 seconds — if the user backgrounds the app and returns within 30 seconds, it's still the same session; longer than that, a new session starts on next foreground.
The median across the measurable catalog is much lower than most "good session length" blog posts suggest. Most app sessions sit between 30 seconds and 5 minutes; the apps people imagine when they think "session length" — games, streaming, deep social — are the top decile, not the middle. Use the distribution below to find where your app actually sits.
Category benchmarks (median session length, 2026 anchors):
- Hyper-casual games: 60-180 seconds. Short, repeatable sessions are the model.
- Casual games: 5-15 minutes. Longer engagement loops, level-based progression.
- Mid-core games: 15-45 minutes. Deep gameplay loops.
- Social / messaging: 5-25 minutes. Variable — quick check-in vs deep scroll.
- Streaming / media: 15-60 minutes. Content-length-anchored.
- Productivity: 10-60 minutes. Work-task-driven, often longer.
- Utilities: 30-120 seconds. Task-completion-driven.
- News / content reading: 3-10 minutes. Article-completion-driven.
Cross-category comparisons mislead — utility users opening an app for 30s to check the weather aren't "less engaged" than a casual gamer's 10-minute session.
Why median beats mean for session length: distributions are heavy-tailed. A small fraction of users have marathon sessions (an hour-plus binge in a streaming app, a 3-hour gaming session) that skew the mean upward. The median user has a much shorter session. Reporting "average session length 35 minutes" when the median is 5 minutes paints a misleading picture. Always report median + percentile spread (P25 / P50 / P75 / P95), not mean alone.
The distribution shape tells you how much room for growth your category has. Productivity apps clustering at 30-60s reflects the in-and-out task-completion model; games clustering at 2-5min reflects level-based loops. The handful of apps in the 10-minute-plus tier are mostly content streaming and immersive games — that's a structurally different product, not a session-length goal for a productivity app.
Session length × session frequency is what matters. A 5-minute session 4× per day = 20 minutes of daily engagement. A 20-minute session 1× per day = same total. Different products optimize different shapes. Hyper-casual games target frequent short sessions; streaming apps target fewer longer sessions. Watch both metrics together — lifting session length while frequency drops can be neutral or even bad (the same total engagement, with more friction).
Median session duration & time-per-user by category (US, MWM)
| Category | Median session | Median time / user / day |
|---|---|---|
| Game | 4m 2s | 12m 18s |
| Media & Entertainment | 1m 18s | 3m 44s |
| Social & Communication | 1m 5s | 3m 46s |
| Education & Knowledge | 1m 4s | 3m 8s |
| Lifestyle & Well-being | 57s | 2m 34s |
| Productivity & Tools | 42s | 2m 3s |
Two patterns stand out in the category breakdown. Games dominate — both per-session and per-day, by a wide margin. The non-game spread is tight: 40-90 seconds per session, 2-4 minutes per day across productivity, lifestyle, social, education, and media. The "time per user / day" column is the more useful one for non-game apps because it folds in frequency — a chat app's bursty 1-minute sessions add up the same way a streaming app's longer sessions do.
Common pitfall: optimizing session length by adding friction (forcing users through more screens, longer animations, more onboarding steps). This artificially inflates the metric without lifting genuine engagement, and often hurts retention. Real session-length lift comes from giving users a reason to stay — better content, deeper features, more reasons to come back the next session.