Free is a price, not a business model
As of 2026, roughly 97 percent of apps on both Google Play and the App Store are free to download. Paid-up-front listings are about 3 percent and shrinking.
So "is it free" tells you almost nothing. The useful question is how it is funded, because that single fact predicts more about your experience over the next two years than any feature list: what data it will ask for, whether it will interrupt you, whether it will grow features you did not want, and whether it treats you finishing quickly as a success or a problem.
Here are the six models, what each does to the product, and how to identify which one you are inside.
1. In-app advertising
How it works. The app sells your attention. Banners, interstitials between screens, rewarded video, and native units styled to resemble content.
The scale. In-app advertising spend was around 390 billion dollars in 2025, making it the largest single revenue stream in mobile by total volume.
What it does to the product. The metric is time on screen and impressions served, so every design decision drifts toward more of both:
- Feeds, because a feed has no end and each screen is an impression
- Streaks and daily rewards, because they manufacture a reason to open
- Notifications aimed at re-engagement rather than information
- Extra taps in flows, because a flow with more screens has more ad slots
- Broad data collection, because targeted impressions are worth several times untargeted ones
Where it is genuinely fine. Entertainment, casual games, news, anything where time spent is the point. If you want to be there, an ad-funded model and your interests are not in conflict.
Where it fails. Utilities. A tool exists to be finished with, and advertising pays for the opposite. An ad-funded utility must work against its own purpose, and it will resolve that tension in favour of revenue because it has to make payroll.
How to spot it: the store listing says "Contains ads". The permissions are broader than the features explain.
2. In-app purchases
How it works. The app is free, and specific items cost money: currency, unlocks, cosmetics, extra lives.
The scale. Around 167 billion dollars in 2025, growing roughly 10 percent year on year.
What it does to the product. IAP revenue is famously concentrated. A very small percentage of users generate the large majority of it, which means the product gets optimised for that minority rather than for you.
In games this produces engineered friction: a difficulty curve that becomes unpleasant exactly where a purchase relieves it, timers that make waiting tedious, and currencies deliberately priced so you always have slightly the wrong amount.
Where it is fine. Genuine feature unlocks. A photo editor selling a professional toolset is straightforward and honest.
Where it fails. When the free experience is degraded on purpose to make the paid one feel necessary. The tell is whether the paid version adds something or removes an obstacle that was placed there deliberately.
How to spot it: "In-app purchases" on the listing, usually with a price range. A range topping out in the hundreds tells you a lot about who the app is built for.
3. Subscriptions
How it works. Recurring payment, monthly or annual, usually after a free trial.
The scale. Around 120 billion dollars in 2025, with average monthly subscription prices rising to roughly 10 dollars. Subscription apps also retain far better than average: Day 30 retention near 14 percent against a cross-category average closer to 5.
What it does to the product. Better incentive alignment than advertising, because the app makes money when you find it worth keeping. But it introduces two distinct pressures:
Feature accretion. A subscription has to feel worth paying for every month, which produces continuous feature addition. This is a major reason focused tools become sprawling suites: the renewal has to be justified, and "it still does the one thing well" is a hard sell twelve times a year.
Cancellation friction. The strongest revenue lever is making leaving difficult. This is where the roach motel pattern lives, and it is the pattern regulators have moved on most directly. The FTC's rule on negative option marketing requires cancelling to be at least as easy as signing up, and specifically targets flows that force a phone call or a chat session to escape something you started online. We cover the pattern alongside eleven others in dark patterns in apps.
How to spot it: trial language on the listing, and an in-app purchase entry with a recurring price.
4. Freemium
How it works. A functional free tier and a paid upgrade. Distinct from IAP because the upgrade is the whole product rather than an item inside it.
What it does to the product. Everything depends on where the line sits, and the line moves.
The recurring pattern is a generous free tier at launch to build a user base, then gradual narrowing as monetisation pressure arrives. Limits get introduced. Existing users are grandfathered, then eventually not. The app you chose in year one is not the app you have in year three.
How to spot it: an upgrade or premium screen. Worth searching for the app's name plus "removed free" or "pricing change" before committing to it.
5. Data
How it works. The app collects behavioural data and licenses it, or shares it with advertising and analytics partners who combine it into profiles.
The important nuance. Outright sale of individually identifiable data is less common than people assume, partly because platform rules and privacy law have restricted it. What is extremely common is sharing for advertising purposes, which is functionally similar and does not meet the legal definition of a sale in many jurisdictions.
The mechanism is usually not the app developer setting out to be a data business. It is bundled SDKs: advertising, analytics, and attribution libraries compiled into the app that collect on their own behalf. A small developer may genuinely not know the full extent of what they gather.
What it does to the product. Permissions and collection that exceed the features. This is the most reliable signal available: an app requesting location, contacts, or device identifiers with no feature that uses them is telling you what it is. Research has found popular Android apps requesting around eleven dangerous permissions on average, which is far past what most of their features justify. Our breakdown is in app permissions explained.
How to spot it: the data safety section on the store listing, which both platforms now require. Look specifically at what is shared with third parties, not just what is collected.
6. Cross-subsidy
How it works. The app is not the business. It supports a business that earns elsewhere: a bank's app, a retailer's app, a hardware company's companion app, or a free tool that exists to sell a related paid product.
What it does to the product. Usually the most benign of the six for privacy and attention, because the app is a cost centre rather than a revenue centre and has no pressure to extract from you directly.
The risk is different: abandonment. When the parent business changes strategy, the app is a line item, and line items get cut. This is where exit cost matters most, which we cover in how to choose an app you will still use in a year.
The mixing
About 35 percent of apps now combine several models, and pure single-model apps are increasingly rare.
The common combination is ads plus an ad-removal purchase, which creates a peculiar dynamic: the free experience has to be irritating enough that removing the irritation is worth paying for. The advertising is doing double duty as both revenue and sales pressure for the upgrade.
Watch for models being added after launch too. An app that was one-time-purchase adding a subscription, or an ad-free app introducing ads, is a common and legitimate response to running out of money. It also means the app you evaluated is not the app you now have.
Working out which one you are in
Five minutes on the store listing, before installing:
- Does it say "Contains ads"? Model 1.
- Does it list in-app purchases, and at what prices? A range up to hundreds means model 2 aimed at high spenders. A single recurring price means model 3.
- Read the data safety section. Anything shared with third parties that does not map to a feature you can name points at model 5.
- Check who the developer is. A bank, retailer, or hardware company means model 6.
- If none of the above explains it, either the funding is unresolved, or it is model 5 and not fully disclosed.
That last case deserves care. An app with no visible revenue model is not necessarily sinister, but it is necessarily unresolved, and unresolved eventually resolves. Usually into advertising, a subscription, or shutdown.
Where we sit, including the awkward part
We build apps, so this is a disclosure rather than a survey.
PackPilot and PawDex are free on Google Play with no advertising, no in-app purchases, and no subscription. Which means, by the framework above, they are currently in the unresolved category, and by our own advice that is something you should weigh rather than ignore.
The honest position: the intended path is a paid tier once each app has enough users to support one, not advertising. We prefer that because advertising would require optimising for time on screen, and both apps are built to be finished with quickly. Those goals cannot both be served, and the design would lose.
That is a plan, not a guarantee, and plans change under financial pressure. The checkable part is what the apps do today: no advertising SDK in either one, which is why the permission lists are short, and no engagement mechanics, which is why neither has a streak. We wrote about why in why app streaks work and why we do not use them.
Verify it against the Play listing rather than this page. That is the correct treatment for any company describing its own incentives, ours included.
The short version
Ask how it makes money before you ask what it does. The funding model predicts the roadmap.
- Advertising is fine for entertainment and structurally wrong for utilities
- Subscriptions align better and produce feature bloat and cancellation friction
- In-app purchases optimise for a small minority of high spenders
- Freemium lines move, always in one direction
- Data shows up as permissions exceeding features
- Cross-subsidy is benign until the parent loses interest
- No visible model means unresolved, not free
Related Reading
- How to Choose an App You Will Still Use in a Year turns this into a pre-install checklist.
- App Permissions Explained covers reading the data safety section properly.
- Dark Patterns in Apps: A Field Guide covers what these incentives produce in the interface.
- What Is a Single-Purpose App? covers the shape of software with fewer of these pressures.
Sources and further reading: Mobile app monetization statistics, App Verticals, App revenue statistics, CoinLaw, FTC guidance on negative option marketing.
Frequently Asked Questions
Six main routes: in-app advertising, in-app purchases, subscriptions, freemium upgrades, data licensing, and cross-subsidy where the app supports a business that earns elsewhere. Around 35 percent of apps now combine several of these rather than relying on one.
Roughly 97 percent on both Google Play and the App Store as of 2026. Paid-up-front listings are about 3 percent, which is why the interesting question is not whether an app is free but how it is funded.
It depends heavily on how the measurement is done, and published figures disagree. In-app advertising spend was around 390 billion dollars in 2025, while in-app purchases were about 167 billion and subscriptions about 120 billion. Per user, purchases and subscriptions convert a small minority into much higher value.
Not inherently. Advertising is a legitimate way to fund software that people would not otherwise pay for. The problem is when it funds a utility, because advertising rewards time on screen while a tool should reward finishing quickly, and those two goals pull in opposite directions.
Check the store listing for in-app purchase and advertising disclosures, look at the permissions and data safety section for collection that does not match any feature, and see whether a subscription is mentioned. If none of those explain it, the answer may be data or it may be that funding is still unresolved.
Direct sale of individually identifiable data is less common than assumed and increasingly restricted by platform rules. What is very common is sharing data with advertising and analytics partners so it can be combined into profiles, which achieves a similar outcome without meeting the definition of a sale.