Streaming Subscription Statistics

A monthly churn rate around the mid single digits sounds small until it compounds: at that level a service replaces a large share of its base every year simply to stand still. That is why release calendars, bundles and annual plans have become retention instruments rather than marketing choices.

Churn figures are only comparable when the denominator matches. Some analyses exclude subscribers acquired through a telco or device bundle, some count a paused subscription as cancelled, and promotional cohorts churn far above the average when the discount ends. Compare like-for-like cohorts or not at all.

When a service offers a cheaper ad-supported plan, a large share of new subscribers take it, and some existing subscribers downgrade. Whether that helps depends on advertising yield: an ad tier earns its place only when subscription revenue plus advertising revenue per user beats the price gap.

This is also why subscriber counts became a weaker signal. Two services reporting the same growth can have very different revenue trajectories if one grew mainly on its ad tier and the other on full-price plans. Revenue per subscriber, published inconsistently, is the number that actually settles it.

Households in mature markets hold several services at once but keep a fairly stable total spend. A new service therefore usually arrives by displacing another, not by expanding the budget — which makes bundles attractive to operators, because a bundled service is harder to isolate and cancel.

Survey figures on the number of services per household vary with the question asked. Counting any service used in the past month gives a higher number than counting services the respondent pays for directly, and free ad-supported apps inflate both if they are not excluded explicitly.

Services disclose what flatters them. Some report paid subscribers, some accounts, some monthly active users including ad-tier viewers on shared plans. When a company changes the metric it reports, treat the series as broken rather than continuous, and rebuild your comparison from the new definition forward.

For planning, pair any subscriber figure with a viewing figure from an independent panel. Subscriptions measure willingness to pay; panels measure attention. The two diverge most for services carried inside a bundle.