Churn is retention's complement. If 40% of users return in month 2, then 60% churned. For engagement-tracked products (free apps, games), churn usually means "stopped opening the app for some window" — say, a user who hasn't launched in 14 days is considered churned. For subscription businesses, churn is more literal: the percentage of paid subscribers who cancel in a given period.
Those engagement-churn numbers are uncomfortable but accurate: roughly 3 in 4 installs never return on day 1, and the median app has lost 96% of installs by D30. The "good D30 retention" benchmarks blogs cite (15-30%) describe the top-10%, not the median. This data is engagement churn — for subscription-specific monthly churn, see the subscription-tier benchmarks below.
Subscription tenure has a closed-form relationship with monthly churn rate: expected tenure ≈ 1 ÷ monthly churn rate. A 5% monthly churn rate implies an average tenure of 20 months. A 10% monthly churn implies 10 months. A 2% monthly churn implies 50 months. This is also why churn directly drives LTV: for a $10 / month subscription, LTV ≈ $10 ÷ monthly churn rate, so 5% churn = $200 LTV, 10% churn = $100 LTV, 2% churn = $500 LTV.
Two distinct churn types matter, with very different fixes. Voluntary churn — the user actively cancels — is what product, marketing, and customer service can influence: paywall pricing, in-app value delivery, retention offers, content freshness, support quality. The interventions are creative and product-driven, and the lift compounds over time.
Involuntary churn — billing failures from expired cards, insufficient funds, foreign-transaction declines — typically accounts for 20-30% of total subscription losses and is purely operational. The fix is "smart dunning": automatic retry logic at different cadences, in-app prompts to update payment method, email recovery flows. Best-practice dunning systems recover 30-50% of involuntary cancellations within 14 days. RevenueCat, Stripe, Adapty, and Apple's own Billing Retry all offer increasingly sophisticated dunning.
A 1-percentage-point reduction in monthly churn compounds into a roughly 25% LTV uplift. The math: at 5% churn, LTV ≈ ARPPU × 20 months. At 4% churn, LTV ≈ ARPPU × 25 months. That's a 25% bump in lifetime revenue per paying user, applied to every future cohort — without buying a single extra install. This is why mature subscription apps invest heavily in retention tooling and treat churn as a first-class metric alongside acquisition.
Churn segmentation matters too. Blended churn averages over wildly different user types: new subscribers in their first 30 days (highest churn — the "first-month effect"), users on intro-priced offers about to step up to full price (high churn at the step-up), long-tenure loyalists (low churn). Reporting blended churn hides where the leaks are; reporting by cohort + tenure bucket exposes them.
Common categorical benchmarks (subscription, 2026): productivity tools 4-7% monthly voluntary churn; streaming / entertainment 5-10%; dating 10-15% (high natural cycle); fitness 8-12%; news / content 5-9%. Free-to-play games measure "engagement churn" instead — a 30-day "did this user come back" rate, which typically lands at 50-80% by month 2 across most genres.
The distribution above shows the truth of engagement churn at scale. Almost no apps fall under 80% D30 churn — only a thin sliver of the top decile clears that threshold. The bulk of the catalog sits in the 90-97% D30 churn band; "average" is much closer to 96% than to anything the conventional benchmark wisdom suggests.
Median D1 / D7 / D30 churn rates by category (US, Q3 2025, MWM)
| Category | D1 churn | D7 churn | D30 churn |
|---|---|---|---|
| Social & Communication | 68.1% | 87.7% | 94.1% |
| Lifestyle & Well-being | 76.3% | 90.4% | 95.2% |
| Productivity & Tools | 77.0% | 91.1% | 95.5% |
| Education & Knowledge | 75.1% | 91.4% | 96.4% |
| Media & Entertainment | 75.1% | 92.0% | 96.6% |
| Game | 63.4% | 90.7% | 97.1% |
Category breakdown patterns: Social and Communication apps churn least at every horizon (the structural daily-need effect), while Games churn most by D30 despite competitive D1 numbers — the classic "gamers play hard, then stop" pattern. The D1-to-D30 gap is where category effects really show: Games lose 30+ percentage points more from D1 to D30 than Social, reflecting how steeply game engagement decays once the novelty fades.
Does churn vary by country?
The common assumption is that emerging markets churn faster — lower-intent users, storage pressure, more aggressive app deletion. The catalog data flatly contradicts it.
Median D1 / D7 / D30 churn by country — Tier-1 vs emerging markets (Q3 2025, MWM)
| Country | D1 churn | D7 churn | D30 churn |
|---|---|---|---|
| United States | 72.7% | 90.8% | 96.1% |
| United Kingdom | 72.8% | 91.0% | 96.2% |
| Germany | 72.6% | 91.0% | 96.2% |
| France | 72.7% | 91.0% | 96.2% |
| Japan | 72.6% | 90.9% | 96.2% |
| South Korea | 72.6% | 91.0% | 96.2% |
| Brazil | 72.9% | 91.1% | 96.2% |
| India | 72.5% | 90.9% | 96.3% |
Median D1 / D7 / D30 churn is essentially identical across every major market: D1 sits around 72.6%, D7 around 91%, D30 around 96% from the US to India, all within half a point of each other. Engagement churn is structural to the medium, not the geography — the leaky-bucket shape of a mobile cohort looks the same in São Paulo, Seoul, and San Francisco. The variation between apps inside any one market dwarfs the variation between markets. This is the loss-side mirror of the flat curve documented on the [[retention]] page.