Ads, Subscription or One-Time Purchase? Revenue Models for Apps

I started building Android apps with Java in 2020 and shipped my first apps to the store in 2022 — that was the first time something I built in production actually earned money. Since then I have tried both ads and in-app purchases. The most important thing I learned is this: the wrong revenue model loses users faster than a bad interface does.

Three models, three different questions

To decide which model to pick, you only need to ask your app one question: how often, and for how long, will someone use this app?

Ads: three formats, and each one belongs somewhere different

Treating ads as "dropped one in, done" is the most expensive mistake there is. There are three formats, and their effect on the user is wildly different:

• Banner: Sits fixed at the bottom of the screen. Low revenue, but it does not break the flow. Reasonable on screens with long lists.
• Interstitial: Takes over the full screen. The revenue is high, and so is the annoyance. Put it in the wrong place and people uninstall the app.
• Rewarded: The user chooses to watch it and gets something in return. To me this is the most honest model — because the user knows they are making a trade.

When should you show an ad?

The rule I follow is simple: an ad should appear after the user finishes something, not in the middle of it. When a level is completed, after a record is added, after a result is shown.

You also have to cap the frequency. An app that shows an ad on every screen transition earns more in the short term, but earns less overall because it loses the user.

Subscription: is the value continuous?

A subscription has exactly one real requirement: value has to be produced continuously enough to justify the user paying again every month. If the app solves one job once, a subscription is the wrong model; the user cancels a month later, and they are right to.

• Offer a free trial, but say clearly when it ends.
• Do not make cancelling hard — it is both a store rule and the main source of bad reviews.
• Offer an annual option; revenue per user goes up noticeably.

The line between free and paid

This is the most critical product decision. The free version has to be limited enough that the user considers paying, but not so limited that they never get to see the app work at all.

The measure I use is this: the free version should be able to deliver the app's promise once. The user should hit the limit after they have said "this really does work", not before.

What should you measure?

Download counts tell you nothing about revenue. What you should be looking at:

• Retention: day 1, day 7 and day 30. If this is low, it is too early to talk about the revenue model — it means you have a product problem first.
• Daily revenue per user: look at ads and purchases together.
• Conversion rate: the percentage of users who move from free to paid.
• Uninstall rate: always watch this after you increase ad density.

A revenue model is not a layer you bolt on afterwards, it is part of product design. Deciding where the ad goes also decides how you design that screen.

• Revenue model
• Ads
• Subscription
• AdMob
• Product decision