Search for “top app publishers” and every list ranks by revenue or downloads. Neither figure is public for any publisher on the App Store. Apple publishes no sales by developer, no install counts, and no chart formula; what those lists rank is one estimator’s model of the number, and most of them do not say which estimator, or say nothing at all about where the figure came from.
There is one publisher-level number anyone can check: the rating count on each app’s public listing, summed across every app a seller publishes. That is what this post ranks by, and it is what the top publishers report on this site computes every day. The short version: Google leads with 91.8 million US ratings across 17 apps, sixteen sellers account for half of all ratings in the dataset, and 92% of publishers have exactly one app in it. The longer version is about what that number is a measurement of, because it is narrower than it looks.
What was counted, exactly
Every figure below comes from src/data/publishers.json in this site’s repository, as committed on 1 September 2026. It is derived from the bundled app dataset like this:
- The apps are drawn from Apple’s US top charts — free, grossing and paid — plus apps that answer a fixed list of subscription-intent searches with at least 100 ratings, plus apps that appeared on those charts in an earlier build and have since dropped off. That last group is kept so already-published pages do not vanish, so the pool is “apps that have charted or matched at some point”, not “apps on the chart this morning”. On 1 September, 277 of the 1,155 bundled apps carried a chart position; the other 878 did not.
- Each app’s
userRatingCountandaverageUserRatingcome from Apple’s lookup endpoint withcountry=us, so the counts are US storefront only and lifetime. - Apps are grouped by Apple’s
artistId— the seller record. 1,148 apps carry one; 7 do not and are excluded. - A publisher’s total is the plain sum of its apps’ rating counts. The rating column is the unweighted mean of each app’s star average, so a seller with one app at 4.9 and one at 3.0 shows 4.0 whatever their sizes. It is a mean of means and it is reported here as one.
That produces 998 publishers holding 820.5 million US ratings between them.
The top 15
| # | Publisher | Bundled apps | US ratings | Mean star rating |
|---|---|---|---|---|
| 1 | 17 | 91,807,953 | 4.5 | |
| 2 | Spotify | 1 | 41,842,066 | 4.8 |
| 3 | Meta Platforms, Inc. | 3 | 41,786,647 | 4.6 |
| 4 | Instagram, Inc. | 3 | 31,710,319 | 4.7 |
| 5 | Uber Technologies, Inc. | 3 | 28,037,705 | 4.7 |
| 6 | DoorDash, Inc. | 2 | 25,517,387 | 4.8 |
| 7 | WhatsApp Inc. | 1 | 18,508,982 | 4.7 |
| 8 | TikTok Ltd. | 3 | 18,419,243 | 4.8 |
| 9 | Walmart | 1 | 18,274,316 | 4.8 |
| 10 | Lyft, Inc. | 1 | 17,610,357 | 4.9 |
| 11 | Venmo | 1 | 15,469,588 | 4.8 |
| 12 | AMZN Mobile LLC | 3 | 14,805,971 | 4.8 |
| 13 | Microsoft Corporation | 4 | 14,568,226 | 4.8 |
| 14 | Apple | 6 | 14,032,707 | 4.4 |
| 15 | X Corp. | 2 | 12,358,320 | 4.8 |
Five of the fifteen are single-app sellers. Spotify sits second on one listing with 41.8 million ratings — although YouTube alone, at 48.8 million, is a larger app than Spotify, which is why Google’s total is not really a story about seventeen apps. Apple is fourteenth with six apps and the lowest mean rating in the table, and none of its six held a chart position on the snapshot day.
A publisher is a seller record, and that decides the ranking
Rows 3, 4 and 7 are the same company. Apple’s enrolment page says of an organisation that “your organization’s name will be displayed as the seller name of your apps on the App Store”, and Facebook, Instagram and WhatsApp each ship under a different one. Grouped by artistId they are three publishers. Added together they are 92,005,948 ratings across 7 apps — 198,000 more than Google’s 91.8 million across 17, a gap of about 0.2%.
So the answer to “who is the biggest App Store publisher” depends on whether you mean the corporate parent or the entity Apple lists as the seller. This site groups by the seller record because that is the only grouping Apple exposes, and because merging by hand means deciding, for 998 sellers, which ones belong to whom. The report does not do that, and this post does not either. Where it matters — and at the top it matters — the ranking above is a ranking of seller records, and you should read Meta’s position as “three rows, together the largest”.
The same caution applies further down: nothing here asserts which of the 998 seller names share a parent.
How concentrated it is
Rating counts are heavily skewed and the publisher totals inherit it.
| Cumulative share of all 820.5M ratings | Publishers needed |
|---|---|
| 11.2% | 1 (Google) |
| 28.7% | 5 |
| 40.6% | 10 |
| 50% | 16 |
| 78.2% | 50 |
| 90% | 89 |
| 92.0% | 100 |
Sixteen seller records hold half of every US rating in the dataset. Eighty-nine hold nine-tenths. The remaining 909 publishers share the last 10%.
The other side of that distribution matters more if you publish apps. The median publisher has 5,350 US ratings. 310 of the 998 (31%) are under 1,000; 562 (56%) are under 10,000. Only 110 clear one million, and 17 clear ten million. This is the same shape the charting-apps distribution has at the app level, and for the same reason: the sellers you have heard of are on top, and the field is mostly sellers you have not.
Most publishers here have one app
918 of the 998 publishers — 92% — have exactly one app in the dataset. Eighty have two or more; seven have five or more; two have ten or more. Those two are Google, with 17, and Flipline Studios, with 14 apps that together hold 10,487 ratings and sit at position 433.
Two things follow. First, a portfolio is rare among the sellers who reach these charts, and having one is not what puts the big sellers on top — four of the top ten are single-app sellers, and Google’s total is dominated by YouTube. Second, “apps in the dataset” is not “apps the seller publishes”. Flipline’s 14 are in the bundle because they answered a search or charted at some point. The count measures presence in this pool, not catalogue size. For a seller’s full catalogue, the publisher search reads it live from Apple’s lookup-by-artist endpoint.
Where the ratings sit, by category
Each publisher is filed under the category most of its bundled apps belong to. By publisher count, Productivity leads with 185 sellers (18.5%), then Games with 157 (15.7%), Finance 86, Entertainment 84 and Utilities 69. By summed ratings the order changes: Productivity still leads at 15.0% of all ratings, but Travel (10.3%), Shopping (9.8%), Social Networking (9.1%), Finance (8.5%) and Food & Drink (8.4%) all sit above Games, which holds 6.9% of ratings from 15.7% of publishers.
That gap is not evidence that games are rated less often — a seller’s category here is the category of its most common bundled app, a Games seller with one charting title counts the same as Google, and the games chart is one of three feeds being sampled. It is a description of this pool.
Publisher country, and why it is not in the table
The report shows a country beside publishers where one is known, and the site’s search pages filter on it. That field is not Apple’s. It is publisher_country from a third-party estimate provider, fetched alongside download estimates, and it has two properties worth knowing before you trust it:
- Sampled across 600 apps on this site,
USwas the only two-letter code ever returned; every other value was already a full country name (“Sweden”, “South Korea”). The site normalises the code to a name so the largest bucket is not split in two. - Roughly 19% of apps carry no value at all, and the site shows those as unknown rather than guessing.
Because the field is third-party, per-app rather than per-seller, and missing for a fifth of apps, it is not part of the committed publisher dataset and there is no country breakdown in this post. Apple’s lookup response carries a sellerName and no seller location, so there is nothing first-party to fall back on.
What is not knowable from public data
- Revenue. Apple publishes no sales figures for any developer. Every revenue ranking of App Store publishers is a third-party estimate, and this site labels them as such wherever it shows one. One widely circulated list names neither its metric nor its source; another ranks by estimated 2025 revenue attributed to an estimation vendor. There is no revenue ranking here.
- Downloads. Same situation. The site shows download estimates on publisher pages, batched from the same third-party source, and treats a figure returned as a floor (
< 5k) as unmeasured rather than as a number. - Worldwide ratings. Apple reports rating counts per storefront, so these are US-only totals. A seller whose users are mostly outside the US is under-counted here relative to one whose users are mostly inside it, and the ranking cannot tell you which is which. The worldwide review sweep reads the other 57 storefronts for one app at a time.
- Users. Ratings are not users. How often an app’s users rate it varies by category, by age of the app and by whether the app asks, and none of that is visible here.
- Whether being big causes anything. Apple’s App Store search page says ratings and reviews “can influence how your app ranks in App Store search”; it does not say by how much. A publisher’s summed count is a description of accumulated ratings, not of anything the publisher did to earn them.
What this ranking is good for
It is good for one thing: a public, exact, reproducible answer to “which sellers have accumulated the most US ratings among the apps that have reached these charts”. The number is Apple’s, the sum is arithmetic, and anyone can rerun it. Where an estimate would give you a plausible figure with no way to check it, this gives you a narrower figure with every input on a public listing.
The ranking updates daily as the dataset is rebuilt — the working copy being built as this was written already had Roblox Corporation entering seventh on a single app — so the table above is a dated snapshot, not a leaderboard. Read the current version on the top publishers report, and look up any seller’s full catalogue, with its estimates labelled as estimates, on the publisher search.