Can a strong US rating hide a weaker rating in another country?

Most country-rating gaps were small.
One app stood out.

We compared 35 highly rated US apps with their ratings in other countries. Only KICK had a rating at least half a star lower in another qualifying market.

The theory we tested: a strong US rating can hide substantially weaker ratings elsewhere. The repeated pattern we expected did not appear under the original rule.

We required at least three different apps with a gap of 0.50 stars or more. We found one. Across 233 qualifying US-to-country comparisons, 231 gaps were smaller than half a star in either direction.

KICK provides a concrete example of a country gap. It does not establish that large gaps are common across highly rated apps.

A country-specific rating is worth measuring, not assuming.

Apple gives each app a separate summary rating for each App Store territory. An ASO report using only the US rating therefore cannot tell you the exact rating people see in Japan, India or Australia. Apple’s ratings guidance.

But different does not automatically mean dramatically worse. In this study, 34 of 35 qualifying apps had no foreign rating at least half a star below their US rating. Of the 233 paired comparisons, 147 differed by less than 0.10 stars and 227 by less than 0.30 stars.

The useful ASO finding is the distinction between a small local variation and a substantial gap. Our half-star cutoff was a study choice, not an official Apple threshold or a proven conversion threshold.

These results describe a selected group of prominent apps. They do not prove that the US score is an accurate substitute for every app’s international ratings, or that smaller gaps have no business impact.

KICK: 4.7 in the US, 4.1 in India and Japan.

United States

4.7 stars

146,355 ratings

Unrounded API value: 4.67171.

India and Japan

4.1 stars

India: 1,431 ratings
Japan: 3,234 ratings

Unrounded values: 4.08106 in India and 4.12276 in Japan.

The differences were 0.59 stars in India and 0.55 stars in Japan. Both passed the original half-star rule and the minimum of 1,000 ratings in each market.

The reported app version was the same in all three markets: 40.33.0. That rules out a different version string in these listings. It does not establish that pricing, features, service or users’ experiences were identical.

We checked the public listing’s structured rating data for all three markets. The rounded ratings and exact rating counts matched the lookup data.

The smaller gaps are shown too

Temu’s rounded public ratings can appear as 4.7 in the US and 4.2 in Japan, suggesting a 0.5 gap. The unrounded gap is 0.44494, so it does not pass the main rule. This is why the analysis uses the underlying values.

Small differences went in both directions.

A positive number in the last column means the US rating was higher. A negative number means the other country’s rating was higher. The median is the middle difference among the qualifying apps in that country.

Japan had the largest median US advantage in these comparisons, at about 0.18 stars. The median difference went the other way in Brazil and the Philippines.

The countries contain different sets of eligible apps. This table does not establish that one country’s users are more critical than another’s. It also cannot show whether language, translation or culture caused a difference.

No foreign rating was at least half a star above the US rating in the primary group. That reverse test has little room to vary: every US score already starts at 4.5 or above on a five-star scale. It must not be treated as evidence that US ratings are generally the highest.

The half-star finding depended on one app.

With the smaller 0.30-star cutoff, four apps qualified: KICK, Temu, Freecash and Google. That does not replace the original half-star test.

Requiring 10,000 ratings in both markets left no half-star cases. KICK’s India and Japan ratings fell below that count requirement. The gap did not disappear; those observations were excluded from that stricter comparison.

Among the 19 apps with qualifying data in all nine foreign markets, none had a half-star US advantage. Requiring the same version made no difference because all 247 eligible pairs already had matching reported version strings.

570 requested app-and-country records.

We kept the same 57 app IDs from the earlier research. Each appeared in the US top-free apps chart’s top 20 at least once in the preserved 19 September to 25 October 2026 series. We did not add apps after seeing the results.

The ten storefronts were fixed before collection: United States, United Kingdom, Canada, Australia, Germany, France, Brazil, India, Japan and the Philippines.

We made 30 new metadata requests on 27 October 2026, between 20:11 and 20:12 Manila time (12:11:03 to 12:12:50 UTC). These produced 375 app-and-country records. Matched US and foreign requests were no more than 95 seconds apart.

Apps were matched using their numeric App Store ID. A translated name did not create a new app. Every usable comparison needed a valid rating and at least 1,000 ratings in both markets.

All 30 requests completed without a request error. “No record returned” means this source supplied no usable record for that app and storefront. It is not proof that the app is unavailable or delisted. Missing ratings were never treated as zero.

Inspect every app and every requested market.

The full data and calculations are included.

The download contains all 570 requested records with their status, 375 returned metadata records, every eligible pair, app summaries, six extra checks, source URLs and capture times, the analysis rules and independent verification code.

We checked all 30 raw response hashes and matched all 375 extracted records to those responses. A separate implementation passed 2,321 agreement checks. Ten public listing pages, covering all six pairs in the smaller-gap check, matched both the rounded rating and exact rating count.

Extract the ZIP and run python analyse.py, then python verify.py. Both work offline using Python’s standard library. The listing checks use structured page data, not visual inspection of the App Store on a device.

1. Apple: ratings, reviews and responses. Explains territory-specific summary ratings and rating resets.
2. Apple: ratings and reviews overview. Explains country attribution and edits.
3. Apple: lookup API. App metadata was retrieved separately for each specified storefront; exact requests are in the acquisition log.
4. AppOrbit pinned source, by ivershuo. Earlier source for the fixed panel selection, under ODbL 1.0. The derived panel attribution and license are included.