N-day retention is the strict version of retention measurement: the user must return on EXACTLY day N after install. D1 N-day retention = the user opened the app on day 1 after install. D7 = opened on day 7. D30 = opened on day 30. If they came back on days 1, 2, 3, 5, 6, 8 but not day 7, they don't count toward D7 N-day retention.
The numbers above are catalog-wide medians measured at the N-day strict definition. Notice how steeply the metric collapses: D1 holds at 27%, but by D30 the median app retains under 4% of installs strictly on day 30. This is why "blog-post benchmarks" of D30 retention in the 15-30% range are usually rolling retention or top-decile N-day, not the median signal you'd compute by following the strict definition consistently.
Why N-day is the industry default: it's unambiguous and easy to communicate. "D30 retention = 12%" has exactly one meaning. The strictness produces a lower number than rolling retention (which counts day-N-or-after), but the strictness is the point — it's a sharper signal of whether users are habitually returning at the specific cadence the metric measures.
Day-of-week effects are visible in N-day retention curves. For products with strong weekday usage patterns (productivity apps, business tools), D7 retention is often higher than D6 or D8 because users align their week. The retention curve looks like a damped oscillation rather than a smooth decay. This is a feature, not a bug — it tells you the product has a weekly usage rhythm. If your D7 retention is much higher than D6 / D8, you likely have a strong weekday-anchored usage pattern.
When to use N-day vs rolling retention
- High-frequency products (messaging, social, daily-streak games): N-day works well. Users genuinely come back daily, so EXACT-day measurement isn't too noisy.
- Low-frequency products (utilities, weekly-use content apps, monthly-bill-pay apps): N-day is too noisy. A user who comes back on day 5, 12, and 28 has D30 = 0 on N-day but is clearly engaged. Use rolling retention instead.
- Comparison across industries: N-day is the lingua franca. When you read "D30 retention = X%" in a public report, assume N-day unless otherwise stated.
Common pitfall: comparing your rolling-retention numbers to industry N-day benchmarks. Your "D30 retention 35%" might be rolling retention (counting anyone who returned on day 30 or later), and the industry "D30 retention 12%" is N-day (only day-30 visitors). Apples and oranges. Always know which method you're using and what the comparison target uses.
The chart shows the N-day D30 distribution at catalog scale. The dominant buckets are 1-2.5% and 2.5-5%, with the bulk of catalog apps stacked between 1% and 10%. The "20%+ D30 N-day" bucket holds only a few thousand apps — that's the real population for which D30 rolling-retention case studies are written.
Median D1 / D7 / D30 retention by category (US, Q3 2025, MWM)
| Category | D1 | D7 | D30 |
|---|---|---|---|
| Social & Communication | 31.9% | 12.3% | 5.9% |
| Lifestyle & Well-being | 23.6% | 9.6% | 4.8% |
| Productivity & Tools | 23.0% | 8.9% | 4.5% |
| Education & Knowledge | 24.9% | 8.6% | 3.6% |
| Media & Entertainment | 24.9% | 8.0% | 3.4% |
| Game | 36.6% | 9.3% | 2.9% |
Category breakdown shows the textbook pattern: Games have the highest D1 (most install-day stickiness from onboarding flow) but the lowest D30 (the steepest mid-funnel collapse). Social & Communication has the most stable curve. The D1-to-D30 spread is what N-day retention exposes that single-number summaries hide.