The Round Number Effect

Look at the raw counts and one pattern immediately stands out: out of the 483,983 rating totals studied, 113,580 end in the digit 1. That's 23.5% of the whole set, more than four times the share ending in 0 (27,222, or 5.6%). If publishers or app store mechanics were pushing counts toward round numbers, you'd expect the opposite, with 0 winning by a mile instead of trailing near the bottom. Instead the digits decline in a fairly steady slope from 2 (72,790, 15%) down to 9 (28,558, 5.9%), and 1 sits well above where that slope would otherwise put it. The likely explanation isn't rounding at all: huge numbers of apps barely get rated, so counts like 1, 11, 21, and 31 turn up constantly. Near the bottom of the rating scale, small integers cluster on 1 far more than chance alone would predict. So the biggest pattern in this dataset has nothing to do with tidy rounding habits. It's really just a signal of how many apps never pick up more than a handful of ratings. If you're using rating counts as a stand-in for popularity, keep in mind that the most common ending digit reflects scarcity at the bottom, not neatness at the top.

Look at the 109,396 apps that have crossed the 100-rating threshold, and only 1,322 of them land on an exact multiple of 100. That works out to 1.21% of the group, which is a surprisingly thin slice given how obvious a round hundred becomes once a count hits three digits. If publishers or reviewers were quietly steering rating counts toward tidy numbers, this is exactly where you'd catch it, apps large enough to matter but still small enough that someone could plausibly round them for show. The 1.21% figure tells you that just isn't happening in any real way. Rating counts past 100 pile up naturally rather than getting massaged into place. So next time you're eyeing a listed rating count and wondering if it's been dressed up, keep this in mind: at this scale, the odds of landing on a round hundred by chance alone are roughly 1 in 83, barely above what randomness alone would produce.

Push the threshold up to apps with 1,000 or more ratings and the numbers drift even farther from anything round: just 73 of 46,411 qualifying apps sit on an exact multiple of 1,000, which works out to 0.16%. That's roughly a seventh of the already-slim 1.21% rate seen at the hundreds level. The reason isn't complicated. Ask for more digits to line up as zeros and the target window shrinks fast, and nothing about how ratings accumulate naturally would cause growth to stop right on a four-digit round number. If anything, this is the cleanest signal in the whole study that rating counts aren't being rounded artificially or dumped in tidy batches. It's also a handy baseline for anyone building a fraud or manipulation detector on App Store data: exact-thousand landings are rare enough (0.16%) that even a small cluster of them around one publisher or category would be worth a second look, since so few should turn up by chance.

Track the roundness score as it moves from bucket to bucket and something odd shows up. It opens at 1.3 in the 100-999 range, drops to 1.1 for 1k-9.9k, hits its low point of 1.0 in the 10k-99k range, climbs back to 1.1 for 100k-999k, then jumps to 1.9 once an app crosses 1M+ ratings. That dip-then-spike shape is the interesting bit. Apps sitting in the middle, roughly tens of thousands to hundreds of thousands of ratings, are the ones least likely to land on a clean number. Makes sense when you think about it: at that volume, ratings accumulate too quickly for a running total to happen to stop on something round. But the 1M+ group breaks that pattern, and I doubt it's a statistical quirk. Apps at that scale get watched, screenshotted, written up, so a round million-plus number is precisely the kind of figure that ends up quoted, whether that's in a headline or tucked into a store listing. So really there are two separate things happening: small numbers cluster naturally at the low end, while milestone visibility takes over at the top.

Travel comes out on top with a roundness score of 1.6. Weather isn't far off at 1.5, and Health & Fitness sits tied with Medical at 1.4. At the other end, Shopping, Food & Drink, and Photo & Video bring up the rear among the twelve categories, each stuck at 1.2. The spread from top to bottom isn't huge, but it's telling. Travel and Weather apps tend to pick up reviews in short, seasonal bursts, whereas categories people rate constantly, like Photo & Video and Shopping, rack up so many reviews that the sheer volume smooths out any tendency toward a tidy number. When there are fewer total rating events, the running count has more room to settle near a round figure, particularly when usage bunches up around travel seasons or weather events instead of staying steady all year. Shopping, Food & Drink, and Photo & Video see the busiest, most continuous rating activity of the bunch, and their distributions come out flattest and least round. That points to low, irregular volume as the real driver behind the roundness, not anything about the category itself.

Roundness by launch year bounces between 0.9 (2010) and 1.4 (2009, 2013, 2017), and there's no real climb or dip across the 2008 to 2019 stretch. 2010 and 2015 land at the bottom with 0.9 and 1.0, while 2009, 2013, and 2017 all reach the same peak of 1.4. None of these highs or lows line up with the App Store's growth, its maturity, or any platform shift worth mentioning. So a tempting story just falls apart here: round numbers didn't get more or less common as the App Store aged, as review systems evolved, or as apps got sharper at gaming their metrics over the years. The numbers simply don't back a trend narrative. Whatever pushes roundness in rating counts, it never built up or tapered off across launch cohorts. It looks like noise, sitting in a tight band from 0.9 to 1.4 no matter when an app first showed up.