What is the difference between ROAS, LTV and payback period?
ROAS compares revenue with advertising spend over a stated scope and window. LTV measures or models value per customer or acquired user across a chosen horizon. Payback is when cumulative value on the same denominator reaches acquisition cost. A ratio, a forecast and a crossing time are different measurements, not interchangeable proof of profitability.
- What you need: Same-currency revenue and spend, an explicit acquired-user denominator, cohort dates and an observed or modeled horizon.
- What you get: A comparable ratio, per-user value and conditional crossing point under your declared definitions.
- Limits: Future renewal or retention assumptions are not observations. These metrics do not prove bank payout, complete profit or a guaranteed bid ceiling.
These metrics are useful together only when their populations, units and timing are explicit. Label revenue as gross or net, spend as advertising cost or a broader cost measure, and users as installs, trial starters or paying subscribers. A shared currency symbol does not make different denominators comparable.
Match the measurement to the question
| Metric | Definition used here | Question | Limit |
|---|---|---|---|
| ROAS | Revenue ÷ ad spend over the same scope | Revenue per unit of ad spend? | Gross revenue is not profit or net proceeds. |
| LTV | Observed or modeled value per acquired user over a stated horizon | Value under these assumptions? | Forecasts extend beyond observed evidence. |
| Payback | First point where cumulative per-user value reaches same-denominator CPA | When does this model cover acquisition cost? | A crossing is not a bank deposit. |
Illustrative example: one cohort, different answers
Consider a synthetic cohort of 100 acquired users, $200 advertising spend and $150 gross revenue in a declared reporting window. Gross ROAS is 0.75× ($150 ÷ $200). Gross revenue per acquired user is $1.50 ($150 ÷ 100), compared with a same-denominator CPA of $2.00 ($200 ÷ 100). These observed-window figures do not establish a future LTV or payback date.
Now consider a separate hypothetical net-revenue scenario using that acquisition denominator. Cumulative modeled net revenue of $1.00, $1.80 and $2.40 at periods 1, 2 and 3 first covers a $2.00 CPA at period 3. Its horizon value is $2.40 per acquired user. Do not substitute this scenario for the observed $1.50 gross figure or a known bank receipt.
Subscription billing and retention LTV are different models
The Subscription Payback Calculator models expected charges, conversion, renewal, refunds and commission. Compare charge-time net revenue with consistently defined CPA. See subscription app payback for a charge-by-charge example.
Retention & LTV fits supplied cohort retention observations and applies an ARPDAU assumption across retained-user days. Mixed cohorts use their supplied population weights. A projected horizon is a scenario; fit spread is not a statistical confidence interval, and averaged descriptive fit parameters are not the mixed curve used for revenue.
Neither model reads your store terms, operating costs, cash balance or future customers. Its output is not a platform payout schedule or guaranteed commercial performance.
Compare on the same footing
- Use the same currency and document gross/net treatment.
- Choose an acquisition denominator before entering CPA.
- Label observed dates separately from forecast horizons.
- Use the same horizon and cohort population for scenario comparison.
- Show assumptions and horizons with no payback.
For a simple subscription illustration, a $10 charge with a deliberately assumed 15% commission and no refunds contributes $8.50 modeled net revenue. Three certain charges yield cumulative values of $8.50, $17.00 and $25.50; a $20 CPA crosses at charge 3. These fees and renewals are synthetic assumptions, not current platform-policy claims.
Inspect data before extending forecasts
Check actual cohort and revenue definitions before choosing model inputs. Use Parity Reconciler when supplied exports need comparison. Read the discrepancy guide before treating mismatched metrics as a change in business performance.