← All posts

Calculating LTV for mobile games

Jun 9, 2026

LTV — lifetime value — is the average revenue a player generates over their lifetime in your game. It is the number that tells you how much you can afford to spend acquiring a player. Get it wrong and you either starve growth or buy your way to bankruptcy. This guide walks through how to actually compute it.

Cumulative LTV, not a magic constant

In games, LTV is almost always cumulative: revenue accumulated per player up to day N since install.

LTV(N) = total revenue from a cohort in their first N days / cohort size

So you do not have one LTV — you have a curve: LTV(0), LTV(7), LTV(30), LTV(90), LTV(180). The curve rises fast early and then flattens as the cohort stops spending. The shape matters more than any single point, because it tells you when you recover your acquisition cost. See the lifetime value glossary entry for the full definition.

A subtle point: LTV(N) is a cohort metric, denominator-anchored on day 0. Everyone in the cohort stays in the denominator forever, including the 95% who never pay. That is correct — they cost money to acquire too.

Payer distribution and whales

Average LTV hides an extreme distribution. In most free-to-play games, revenue is brutally concentrated:

  • The vast majority of players never pay at all.
  • A small slice of payers (“dolphins”) spend modestly.
  • A tiny fraction (“whales”) drive the bulk of revenue.

If you only look at the average, you will misread everything. Pull the payer distribution — bucket players by total spend (e.g. $0, $0.01–$10, $10–$100, $100+) and look at how much each bucket contributes. When the top bucket carries half your revenue, your retention strategy for those players is your revenue strategy. The game analytics views in Keentics surface this whale curve directly.

ARPU vs ARPPU

Two averages, often confused, both useful:

  • ARPU — Average Revenue Per User. Total revenue ÷ all active users. Includes non-payers.
  • ARPPU — Average Revenue Per Paying User. Total revenue ÷ paying users only.

The relationship:

ARPU = ARPPU × pay-rate

This decomposition is the actionable part. If ARPU is low, is it because few people pay (low pay-rate → fix the first-purchase funnel) or because payers spend little (low ARPPU → fix pricing and depth)? Two completely different roadmaps. Use a funnel analysis on the path to first purchase to find where would-be payers drop.

LTV against CAC and ROAS

LTV only means something next to what a player costs:

  • CAC — Customer Acquisition Cost — what you paid to acquire that player.
  • ROAS — Return On Ad Spend — revenue ÷ ad spend, usually quoted at a day window (D7 ROAS, D30 ROAS).

The core rule: LTV must exceed CAC, with margin. Because LTV is a curve, the real question is when cumulative LTV crosses CAC — your payback period. A channel that pays back by D30 is a money printer; one that pays back at D400 is a loan you may never collect, because the cohort decays first.

break-even: LTV(N) ≥ CAC
D30 ROAS = revenue through day 30 / ad spend

Always compute this per channel and per cohort, never blended. A blended ROAS of 1.2 can easily hide one channel at 2.5 and another at 0.4 that you should kill today.

Forecasting LTV

You rarely have 180 days to wait. The practical approach: use early signals to project the mature curve.

  1. Anchor on early LTV — LTV(3) or LTV(7) is observable within a week.
  2. Apply a maturity multiplier — from past cohorts, learn the ratio LTV(180) / LTV(7). If history says mature LTV is ~3.5× your D7 value, you can project new cohorts within days of a campaign launching.
  3. Refit constantly. The multiplier drifts when you change monetization, seasonality hits, or channel mix shifts. Treat it as a living estimate, not a constant.

This is rough, but rough-and-early beats precise-and-too-late when you are deciding whether to scale ad spend tomorrow. Pair it with retention analysis — LTV is just retention multiplied by spend, so a retention crack will always show up as an LTV cap.