App ROAS is the revenue your users generate, through in-app purchases, subscriptions, and in-app ad revenue, divided by what you spent to acquire them, the number that actually says whether the app makes money.
You already know CPI only measures the download, so App ROAS is where you go to find out if any of those downloads were worth it. You take the money the acquired users bring back to you and you set it against the spend that brought them in. If a campaign returns more rupees than it costs, it earns its budget. If it does not, no amount of cheap installs will save it, and this is the ratio that tells you which is which.
On Android and on the web, you can watch this fairly directly, because the events flow back to you with reasonable speed and detail. You see who paid, when, and how much, and you can trace revenue to the campaign that earned it. That clean line of sight is what lets you scale winners with confidence and cut losers before they drain the account.
On iOS, it gets murky. Apple's SKAdNetwork limits and delays the data, so campaign-level ROAS arrives fuzzy, aggregated, and late, and you cannot read it as cleanly as web ROAS. Do not treat the platform number as gospel. Blend the SKAN view with your measurement partner's reporting and lean on cohort-based lifetime value, so you are judging campaigns on what a group of users is really worth over time rather than on one incomplete snapshot.
CPI tells you what you paid to get them in the door; App ROAS tells you whether they were ever worth letting in.
Sources
- Apple SKAdNetwork returns aggregated, time-limited postbacks that constrain how precisely iOS campaign revenue and ROAS can be measured. appsflyer.com · verified 9th August 2026
Last checked 9th August 2026. Next check 15th August 2026.
