Tools
LTV Calculator
Estimate revenue per acquired user from ARPDAU and D1/D7/D30 retention. Compare acquisition cost against a target return over 90, 180 or 365 days.
Cohort Economics
Turn Retention Into Value.
Estimated lifetime value and an acquisition-cost ceiling at your target return.
The Retention-Curve Fit
For each pair of anchors, the curve is R(t) = R(a) × (t ÷ a)^b, where b = ln(R(b-day) ÷ R(a)) ÷ ln(b-day ÷ a). One segment joins D1 to D7; another joins D7 to D30. After D30, we extrapolate that second segment’s power-law slope. Installation day has retention of 100%. A zero endpoint uses a linear bridge to zero because logarithms of zero are undefined.
Estimated LTV = ARPDAU × [1 + sum of R(t) from day 1 through the selected horizon]. Max affordable CPI = LTV ÷ (target ROAS ÷ 100). This assumes stable revenue per active day, a declining retention curve and no discounting. It is a finite-horizon scenario, not a statistical confidence interval or true infinite lifetime.
Observed revenue cohorts are better evidence than extrapolation. Test horizons against mature cohorts, model iOS and Android separately and stress-test the long tail. Reactivation, subscriptions, ad-load changes and payer concentration can break these assumptions. Use net ARPDAU when your target is contribution rather than gross revenue.
Common Questions
How is retention-based LTV estimated?
The model fits a power-law curve through D1/D7 and D7/D30 retention anchors in log-log space, then extends the last slope to the selected horizon. It multiplies modeled active days by constant ARPDAU, including installation day at 100% retention.
Should I use gross or net ARPDAU?
Use net revenue per daily active user for contribution economics: remove store fees, refunds and any costs you want the model to account for. Keep cohort, platform and geography consistent with retention.
What is the maximum affordable CPI?
At a target ROAS of 100%, maximum CPI equals modeled LTV. At 150%, it equals LTV divided by 1.5. This is a modeled ceiling, not a recommended bid or guaranteed return.
Why does the model require retention to decline?
This simplified decay model assumes D1 ≥ D7 ≥ D30. Reactivation, small samples and calendar effects can produce nonmonotonic observations; use a cohort model appropriate to those effects rather than forcing the data here.
Can this estimate subscription lifetime value?
Not reliably when renewals, trials and subscriber survival dominate revenue. Constant ARPDAU and active-user retention do not model billing cycles or changing payer mix. Compare against observed revenue cohorts before setting budgets.
Related Content
Next Step
Bring your cohort economics, creative and markets. Let’s scope a test for Adora X.