The One-Percent Club

Food & Drink sits at the top of the inequality list with a p99/p50 ratio of 34,736.7, and Travel isn't far off at 29,866.6, with Shopping close behind at 27,514. These categories tend to revolve around just one or two delivery, booking, or marketplace apps that people keep on their phone permanently, and that small handful sits on top of a long tail of restaurant-branded or single-city apps that nobody bothers to open. Finance (17,215), Lifestyle (15,089.6), and Entertainment (14,415.1) make up a second tier: still heavily skewed, just not as extreme as the top three. Even Book, the lowest of this top twelve, comes in at a ratio of 9,188.3, which would look extreme in almost any other kind of consumer data. Categories built around daily habits, ordering food, booking a flight, buying something online, tend to produce a handful of apps that millions of people default to, while the median app in that same category might just be some regional business's listing that barely gets touched. So if you're benchmarking a Food & Drink or Travel app against "the category average," that average won't tell you much. You're either near the handful of giants, or you're basically invisible.

Look at Stickers: its p99/p50 ratio sits at just 215, and Magazines & Newspapers isn't far behind at 801. Both numbers sit well under everything else in the top twelve. Next come Developer Tools (1,465.7) and Graphics & Design (2,446.3), and a ratio in the low thousands would set off alarm bells almost anywhere else, but in this crowd it's practically unremarkable. What ties these categories together is a smaller, more niche audience: nobody's building the one sticker pack or dev tool that the entire planet needs, so ratings get distributed across a bunch of competitors instead of piling up behind a single champion. Rounding out the group are Reference (4,602.1), Music (4,963.8), Weather (5,123.2), and Games (6,184.2), and that last one is worth pausing on. Games is one of the biggest categories out there by sheer app count, and yet its ratio stays fairly contained. That tells you category size and category inequality aren't the same thing at all. Even Health & Fitness, which tops out this calmer cluster at 7,292.1, still falls an order of magnitude short of Food & Drink's 34,736.7. So "equal" is the wrong word here, or at least it only works loosely. None of this is flat. It's just less lopsided than the extreme cases.

Shopping is by far the hardest category to crack: you need 44,773 ratings just to hit the top 10%, and that's 35.5% of the 126,024-rating total across these twelve categories combined. Games comes in second, requiring 30,882 ratings (24.5% of that same total), so between Shopping and Games you're looking at roughly 60% of the combined threshold weight for all twelve categories. Everything else drops off fast after that. Social Networking sits at 9,229 ratings (7.3%), Finance at 7,520 (6%), and News, the easiest category on the list, needs only 3,194 (2.5%). Shopping's threshold is 14x higher than News.

Saying an app is in the "top 10% of your category" doesn't mean the same thing everywhere, and the gap is bigger than most people assume. Take an app with 5,000 ratings: in Shopping it wouldn't come close to the top decile, but put that same app in News, Food & Drink, or Lifestyle and it sails past the bar without much trouble. So next time someone waves around "top 10% in category" as proof their app is good, check which category they're actually talking about before you buy it.

Business has the widest gap between its 90th and 10th percentile ratings of any category, a full 2.3 points. Entertainment and Medical trail behind that at 1.8 each. Utilities, Sports, and Lifestyle sit at 1.6, and even the tightest categories here, Social Networking, Weather, and Shopping, still come in at 1.4. What a 2.3 point spread actually means is that the best and worst reviewed Business apps land more than two full stars apart on average. That's a huge quality range for apps that are, in theory, solving the same kinds of workplace problems. Part of the reason is that Business functions as a dumping ground of a category. Genuinely polished enterprise tools with loyal users sit right next to abandoned utilities and barely-disguised spam, and that collision is exactly what drives the gap so wide. Worth noting: concentration of ratings and spread of quality don't track each other at all. Shopping is one of the most concentrated categories by volume, with a 27,514 ratio, yet its quality gap is one of the narrower ones at 1.4. Business, meanwhile, sits mid pack on concentration at 13,290.8 but tops the list on quality spread. Concentration tells you where the ratings pile up. The quality gap tells you something different: whether the category's ratings can actually be trusted as a signal of good versus bad. These are two separate risks, and they shouldn't be confused for one another.

Look across the 26 genres measured and two numbers stand out. The average ratings concentration figure sits at 10,619.12. The average rating quality gap comes in at 1.31, which happens to equal exactly 5% of the 26 genres measured. Worth flagging, though probably nothing more than a coincidence. Food & Drink and Travel drag that 10,619.12 concentration figure up into the tens of thousands, while Stickers and Magazines pull it back down into the hundreds. So think of that number as a center of gravity rather than a typical case, because no single category actually looks like the average. The 1.31 rating gap tells a different story. Even Business, which posts the widest individual gap in the dataset at 2.3, still doesn't reach double that figure. Set the two numbers next to each other and the asymmetry jumps out: rating volume swings wildly from category to category, almost like a lottery, while rating quality stays comparatively steady even as it drifts around some. That's the part worth holding onto. Which category an app lands in says a lot about how many ratings it'll pile up. It says far less about how good those ratings actually turn out to be.

All 26 genres measured land in the extreme concentration tier, each one posting a p99/p50 ratio of 50x or higher. That's 26 out of 26 categories, a flat 100%. Not a single exception. So the question of which categories are concentrated versus which are balanced doesn't really apply here, because the actual spread runs from extremely concentrated to something closer to catastrophic. Stickers comes in at the mild end of the group, with a ratio of 215, which still means the top app there is 215 times more skewed than the median one. Most statisticians would flag that as a severe imbalance on its own, and it only reads as small once you set it next to Food & Drink's 34,736.7. Whatever hope you had that some corner of the App Store stayed naturally egalitarian, you can drop it. No category puts the median app and the top app anywhere near the same game. For developers, the takeaway is blunt: you don't get to choose a fair category over an unfair one. You only get to choose how unfair the one you pick will be.