How to Estimate Any App's Downloads and Revenue (2026 Guide)

You can estimate any app's downloads and revenue for free — as long as you accept that you're getting a range, not a number. Google Play shows a rounded public install count on every listing, iOS shows nothing, and every estimate tool you've ever seen fills the gap with a model built on chart ranks and review counts. Those models are useful: the best-case error is around 5%, per Appfigures' own accuracy documentation. They're also capable of being off by 2x or more on a specific app. This guide walks through where the numbers actually come from, the free routes that get you 80% of the answer, and how to sanity-check whatever figure you land on before you make a decision with it.

Why nobody has the real numbers

Start with the uncomfortable truth: outside of a developer's own App Store Connect or Play Console, real download and revenue figures don't exist in public.

AppTweak's explainer on how estimates are made says it plainly: "No third-party platform can provide perfect accuracy, because the App Store and Google Play don't publish raw download or revenue numbers." Everything you see in an estimate tool — every "this app makes $120K/month" screenshot on social media — is a model output, not a measurement.

The two stores leak different amounts of information. Google Play publishes a rounded install bracket on every listing (10,000+, 1,000,000+, and so on), which gives estimators a public floor to anchor on. Apple publishes nothing, so iOS downloads have to be inferred from indirect signals — most commonly review counts and chart positions, as Sonar's breakdown of how revenue estimates work explains. That's why, tool for tool, Android estimates tend to be steadier than iOS ones.

How the estimate tools actually do it

Once you know the mechanics, the numbers get much easier to interpret.

Downloads: chart rank in, install count out

The core insight behind every download estimator is that category rank and daily downloads are tightly correlated. AppTweak describes its model as a deep-learning system trained on years of historical store data plus anonymized real download data from thousands of apps, taking four main inputs: category, rank position, country, and date. Feed in "ranked #12 in Health & Fitness, US, a Tuesday in September" and the model returns a download estimate for that day.

The date matters more than you'd think. AppTweak's data shows US football apps spiking as much as 51% in September and Indian sports apps peaking around 50% during IPL season — the same rank means different download volumes in different weeks.

Revenue: a second estimate stacked on the first

Revenue is where things get wobbly. As Sonar lays out, the basic skeleton most tools use is estimated downloads × estimated revenue per user. Both halves are estimates, and errors multiply: in Sonar's worked example, shifting individual inputs by a plausible 30–40% moved the final revenue figure by more than 10x. Tools compensate by leaning on Top Grossing chart behavior — which reflects monetization directly rather than installs — but a revenue estimate is still a model built on top of another model.

One developer case Sonar documents: an app listed at $40K MRR by a major estimator was actually doing $90K. Not a rounding error — a 2.25x miss.

The free routes

If you're an indie sizing up a competitor or a market, you can get surprisingly far without a subscription.

Check Google Play first. The public install bracket is the only real number either store gives you, and for Android-first apps it immediately puts you in the right order of magnitude. Android Rank offers free install estimates built on that public Play data, as GoPractice's tool roundup notes.

Use the free tiers of the estimate tools. GoPractice's comparison points out that Sensor Tower shows current-month download and revenue estimates without a signup, and MobileAction shows the previous month by country and platform. Sensor Tower's own guide confirms full historical Store Intelligence is enterprise-only — but for "is this a $5K/month app or a $500K/month app?", a free lookup answers the question.

Read the review count as a rough proxy. Review volume correlates with download volume, which is exactly why estimators use it as an iOS input. Comparing review counts between two similar apps in the same category is a legitimate way to gauge relative scale. Just keep it relative: the review-per-download rate varies wildly with how aggressively an app prompts for ratings, so it can rank apps against each other but can't give you an absolute count.

Calibrate with the one app whose numbers you do know: yours. This is the indie superpower. If an estimator says your app did 3,000 downloads last month and Connect says 1,900, you've just measured that tool's error for your category and size — roughly ×0.63 — and you can apply that correction to its estimate of your competitor. Sensor Tower itself invites this comparison, and Appfigures lets you line up your real numbers against its estimates inside the product.

How wrong should you expect estimates to be?

Appfigures is unusually transparent here, and its published figures are a good mental baseline for the whole category: a mean absolute percentage error between 5% (best case) and 25% (worst case), varying with app popularity and day of week.

The fine print matters as much as the headline. Per Appfigures' documentation: paid app downloads aren't estimated at all, revenue estimates show developer proceeds after Apple's and Google's 30% (or 15% subscription) cut rather than gross consumer spend, iOS estimates cover iPhone but not iPad, and small apps below per-country download thresholds simply get no estimate. Every tool has a version of this fine print; most bury it.

And per Sonar's analysis, estimates are weakest exactly where indies live: mid-ranked apps (roughly #300–#1000 in category), apps with unusual retention, and iOS revenue generally. They're strongest for big apps, for trends over time, and for comparisons made within a single tool.

So the practical rules: treat any single figure as the middle of a wide range, never compare App A from one tool with App B from another, trust the trend line more than the level, and never use an estimate as the load-bearing number in a financial decision. If you're pricing an acquisition, ask for console screenshots — estimates are how buyers find targets, not how they verify them.

We'll say it about our own numbers too: Kintsu (currently in private beta, built for indie developers) works with public store data, and ranges-not-points is the honest way to read anyone's estimates, ours included.

For the paid-tool landscape once you outgrow the free routes, see our guides to Sensor Tower's pricing, Appfigures, and what replaced data.ai.

FAQ

Can I find out exactly how much money an app makes? No — unless the developer publishes it or you're in a due-diligence process with access to their console. Apple and Google don't release per-app revenue, so every public figure is a model estimate with meaningful error. Appfigures' own published error range is 5–25% in the cases it covers, and misses of 2x have been documented on specific apps.

What's the most accurate free way to estimate downloads? For Android, the Google Play listing itself: the rounded install bracket is real data from Google, and free tools like Android Rank build on it. For iOS there's no equivalent, so triangulate a free estimator lookup against review counts of comparable apps — and expect a range, not a number.

Why do two tools show completely different revenue for the same app? Because each runs its own model with its own panel data, retention assumptions, and revenue-per-user benchmarks — and small input differences compound into large output differences. Some also report net developer proceeds while others model gross spend. Compare apps within one tool; never across tools.

Are download estimates more reliable than revenue estimates? Generally yes. Downloads are modeled fairly directly from chart ranks (and on Android, from public install data), while revenue stacks a monetization model on top of the download model, multiplying the error. If a decision hinges on one of the two numbers, lean on downloads.

---

Sources: AppTweak — Understanding mobile app download & revenue estimates · Appfigures — Comparing your downloads and revenue with estimates · Sonar — App revenue estimates: how they work (and fail) · GoPractice — How to estimate a competitor app's revenue, downloads, and audience · Sensor Tower — How to get estimated downloads and revenue for apps