Do App Store Ratings Lie?

Look at the numbers: 55.3% of rated apps score 4.5 stars or higher, while only 12.6% drop below 3.0. Grade the App Store like a classroom and almost nobody fails. There's a reason for that. Ratings here are self-selecting and heavily prompted, so the sample is skewed before a single star gets tapped. Developers fire off review requests right after a user hits a win or finishes a level, while anyone who's annoyed just deletes the app and disappears without a word. The result is a scale crammed into its top half, where a 4.6 and a 4.9 might come down to nothing more than luck and when the prompt happened to pop up.

A star average only tells you something once enough people have actually weighed in, and most of the time that threshold isn't met. 86.4% of apps have fewer than 50 ratings, 76.8% have fewer than 10, and 50.7% have no rating at all. A "5.0" sitting on eight reviews doesn't carry much weight, three friends and a good week is all it takes. A "4.6" sitting on two million reviews is a completely different animal, a verdict that's survived contact with every kind of user and every kind of mood.

That's really the whole trick to reading App Store ratings: the rating count is the confidence level, and the average is the score. A high score paired with a low count is just a guess dressed up as data. A high score paired with a high count is closer to a fact. Read them backward, count first, stars second, and you stop falling for it.

Here's a number worth sitting with: among every app with at least one rating, the median app has just 8 ratings. Eight. Climb to the 90th percentile and you still land at 806 ratings; the figure doesn't jump to 59,226 until the top 1%. App Store ratings are wildly top-heavy. A small slice of apps hold almost all the review volume, and the rest run on sample sizes in the single and double digits.

That distribution is why the average star score is so easy to misread. For the median app, one happy fan or one angry user can swing the number by themselves. The score isn't wrong, exactly. It just barely exists.

A flat 5.0 is the most tempting score on the store, and it's almost always a mirage. 26% of all rated apps are a "perfect" 5.0 built on fewer than ten ratings, roughly one in four. These aren't the best apps in the store, they're the newest and least tested, sitting at 5.0 because barely anyone has weighed in yet. One genuinely disappointed user can knock them straight off their pedestal.

The same thinness hides inside the scores that look respectable. Of every app rated 4.5 or higher, 69.7% are backed by fewer than 50 ratings. More than half of the store's "excellent" apps earn that label from a few dozen people. A high average doesn't prove quality until a crowd has actually stress tested it.

Rating volume isn't spread evenly across categories, and that changes what a given score actually tells you. In Games, 16.1% of rated apps break past 1,000 ratings, with Weather (15.7%) and Finance (13.8%) not far behind. These are big, crowded markets where people rate in bulk, so a 4.6 in one of them actually means something.

Flip to the other end and a high score barely registers. Just 0.6% of Stickers apps hit 1,000 ratings, and Developer Tools (2.6%) and Magazines & Newspapers (3.6%) aren't much better off. Most ratings in these corners come from a small handful of reviewers, so a 4.8 that would settle an argument in Games is closer to a shrug here.

Check out the apps in the store with the most reviews, the ones racking up tens of millions of ratings apiece, and not a single one gets close to 5.0. YouTube sits at 4.67 on 49,563,628 ratings; Spotify: Music and Podcasts holds 4.77 on 42,751,452; Instagram comes in at 4.69. Look across the twenty most-rated apps overall and the scores range from 3.25 to 4.92 stars, with nothing hitting a perfect 5.0.

Once an app racks up millions of ratings, it tops out at excellent, not flawless. So when some small app shows off a spotless 5.0, it hasn't actually topped YouTube. It just means nowhere near enough people have used it yet.

Line those two numbers up and a pattern jumps out right away. High average, high count: trust it, the app has earned that score. High average, low count: promising, sure, but not proven, so it's worth checking the developer's history and when they last shipped an update. Low average, high count: believe the crowd here, something is genuinely broken. Low average, low count: there's no real signal, so skip the rating entirely and judge the app on its own terms.

The rarest combination, and the most valuable one, is that first spot: a high average that holds up across a huge audience. That's not something you can fake, which is exactly why our "highest rated with real volume" board is built around it, and why it beats sorting by stars alone.

Before you dig into these numbers, two limits are worth flagging. First, a rating count tells you how many people rated an app, not how good that app actually is. A mediocre app with a big marketing budget can out-vote a brilliant niche one, and our numbers can't tell the two apart. They only tell you how much weight a score can bear. Second, Apple publishes a single lifetime average, so a 4.6 earned years ago can hide a recent drop in quality that newer reviews haven't dragged down yet. Even so, every figure here, the 8-rating median, the 26% perfect-five share, the 69.7% of top-rated apps that are thinly rated, comes from a direct count across 515,707 rated apps. None of it is modelled.