How to calculate mobile app LTV

Start with a cohort, not a dashboard total.

A cohort is a group of users with a shared starting point. For example, everyone who first installed your iOS app through a specific campaign in January. Keep that group fixed while you follow its revenue.

Pick an observation window: day 30, day 60, or day 90 after each install. Only compare cohorts that have had time to reach it. In these notes, D90 means days 0 through 90, inclusive.

RevenueCat’s realized LTV documentation explains this cohort approach and the difference between value per customer and value per paying customer. Here, we use net proceeds per install so the result is useful for acquisition decisions.

Add the proceeds. Divide by installs.

Purchase proceeds should already account for store commissions, applicable transaction taxes, fees, and refunds. Use the ad revenue attributable to the same cohort and window, after applicable deductions. Avoid subtracting any fee twice.

This basis can differ from the “revenue” number in your analytics tool. Check the definition before copying it across. Do not use all of this month’s app revenue with only this month’s new installs.

The calculation is ($2,400 + $600) ÷ 1,000. For a subscription-only app, the ad revenue input is zero. For an ad-supported app with no purchases, purchase proceeds are zero. For a hybrid app, include both without double counting.

Value is not the same as contribution.

Serving users costs money. Subtract the cohort’s variable costs, such as AI inference, usage-based infrastructure, or support attributable to those users.

If the example cohort costs $400 to serve, contribution before acquisition is ($3,000 − $400) ÷ 1,000 = $2.60 per install. That is the amount available to cover acquisition, fixed costs, and profit. Your team’s salaries and other fixed costs are not included in this example.

Label what you know, and what you expect.

D90 realized value measures what happened through day 90. It is not a forecast of the user’s entire lifetime. Later renewals, ad impressions, and refunds can change the outcome.

A simple revenue-divided-by-churn forecast assumes stable revenue and churn in matching periods. A small change in churn can move the estimate dramatically. New apps and mixed subscription plans rarely give you enough evidence to treat that shortcut as a spending guarantee.

Start with measured cohorts. Keep country, platform, channel, and plan mix in view. Add forecasts only when you can compare earlier predictions with actual outcomes.

Calculate your cohort’s value, then read how to compare it with acquisition cost.