How to calculate ROI on an app
The formula is simple: (total return − total investment) ÷ total investment. What people get wrong is the investment side. It is not just the build — it is the build plus three years of hosting, third-party APIs, maintenance, store fees and the support time somebody has to spend.
Over three years, running costs frequently match or exceed the original build cost. An app that cost AED 80,000 to build and AED 3,000 a month to run is a AED 188,000 investment by month 36, not an AED 80,000 one. Any ROI number calculated against the build cost alone is flattering fiction.
Payback period is the number that matters
ROI over three years is a nice headline. Payback period — the month your cumulative profit crosses zero — is what determines whether you survive to see it.
- Under 12 months is excellent and rare. Usually means an internal tool replacing expensive manual work.
- 12–24 months is a healthy consumer or B2B product. Most successful apps live here.
- 24–36 months is viable if you are funded and the growth curve is real, risky if you are self-funding.
- Beyond 36 months means the assumptions need to change — smaller build, higher price, or a different problem.
Churn quietly destroys subscription ROI
At 5% monthly churn you lose roughly half your subscriber base every year. Growth of 6% a month against 5% churn is net 1% — near-flat revenue that looks like growth on a chart of new signups.
Move the churn slider and watch the payback month jump. For most subscription products, a single point of churn reduction is worth more than several points of acquisition growth, and it is usually cheaper to achieve. Onboarding, activation and the first-week experience are ROI features, even though they never appear on a feature list.
The cheapest way to improve ROI is to spend less building
Every dirham you cut from the build is a dirham you do not have to earn back. This is the real argument for shipping an MVP: not that it is cheap, but that it moves the break-even point months closer while you are still learning whether the assumptions hold.
Use the MVP cost calculator to price the smallest credible version, the feature estimator to see which features to defer, and the app cost calculator for the full build. Then come back and re-run the return with the lower number.