Which languages to localize your indie game into, measured on Steam

Indie games on Steam that support Simplified Chinese drew a median 17.5% of their reviews in it. Games without it drew 1.8%, and no other language came close to that gap.

If your game is out and you’re deciding which translation to pay for first, this is what 1,651 indie games’ reviews show, and how to read the same signal in your own.

Most of an indie game’s reviews aren’t in English

We took every paid indie game that launched on Steam from 2020 to 2022 and found at least 100 players who reviewed it in year one, 1,651 games. Then we counted every Steam-purchase review each one has, language by language, on September 23, 2026.

The median game had 55% of its reviews in a language other than English. Across all of them, English was 46% of reviews and Simplified Chinese 20%. Everything else was a long tail.

1. English46.4%
2. Simplified Chinese19.8%
3. Russian8.8%
4. Portuguese (Brazil)3.5%
5. Spanish (Spain)3.4%
6. German3.2%
7. Korean3.1%
8. French2.2%
9. Turkish2.0%
10. Japanese0.9%

Paid indie games launched on Steam 2020 to 2022 with 100+ Steam-purchase reviews in year one. Share of all their Steam-purchase reviews, by the language each review is filed under. Read September 23, 2026.

Simplified Chinese was the biggest non-English language for 841 of the 1,651 games, just over half. Russian was the biggest for 375.

Valve’s own survey counts differently. In August 2026, 38% of Steam users ran Steam in English and 24% in Simplified Chinese. Reviews lean further toward English because many players who speak another language keep Steam in English, and because people review at different rates.

Japanese is 2.5% of users but 0.9% of these reviews. Read reviews as who shows up for indie games, not as a census.

Simplified Chinese is the gap that matters

1,167 of the games support Simplified Chinese today. They drew a median 17.5% of their reviews in it. The 484 that don’t drew 1.8%.

Bigger games are more likely to be translated, so we also compared each game with untranslated games of the same size from the same launch year. The gap held: 18.3% against 1.9%. It held again when we set aside games from studios likely based in China, Japan or Korea: 15% against 1.8%.

1. Simplified Chinese, supported18.3%
2. Simplified Chinese, not supported1.9%
3. Russian, supported7.9%
4. Russian, not supported1.5%
5. Portuguese (Brazil), supported3.2%
6. Portuguese (Brazil), not supported1.0%
7. German, supported2.9%
8. German, not supported0.7%
9. Korean, supported2.8%
10. Korean, not supported0.5%

1,651 paid indie games launched on Steam 2020 to 2022 with 100+ reviews in year one. Each game that supports the language against games that don't, launched the same year with total reviews within 25% of its own. Languages as listed on each store page, read September 22, 2026. Reviews read September 23, 2026.

Russian is the clear second. After that, every language had a median of 3.3% or less of a game’s reviews, even when supported. Those languages can still be worth it for the right game. They just aren’t where the volume is.

This is not proof that translating causes the gap. Developers add languages where players already are, and games with Chinese-speaking audiences get translated first. But the pattern fits what Valve tells developers: “The less a language looks like English, the higher the expected return.”

Valve also says “Steam favors products that are localized.” Players choose their languages in Steam, and games in those languages are “more likely to appear in discovery features” like the Discovery Queue. By Valve’s account, then, a game without Chinese is less likely to be put in front of Chinese players, as well as being harder for them to play.

Your reviews already show who’s missing

For a game that’s already out, the most useful data is your own. Steam’s review filters let you read reviews by language, and your Steamworks wishlist reports break down by region.

Our numbers give you a benchmark. Among games without Simplified Chinese, a quarter still had 4.3% or more of their reviews in it, and a tenth had 8.8% or more. Those players bought a game they would have to play in another language. If your Chinese share is in that range, it is a strong signal.

English-only games show the same thing. Of 256 games that support only English, 142 already had at least 5% of their reviews in a single other language. Across all 256, the biggest non-English language was most often Russian (80 games), then Simplified Chinese (49) and German (48).

Valve suggests a cheaper test before a full translation: translate the store page into the languages you’re considering, then watch which regions start wishlisting.

Translating isn’t the end of the work

Chinese and Japanese reviews run cooler than English ones for the same game. Among games that support Simplified Chinese, the median game’s Chinese reviews were 3.4 points less positive than its English reviews. For Japanese, 4.3 points. Brazilian Portuguese and Spanish reviews ran about 5 points warmer.

That now shows on the store. Since August 2025, Steam shows shoppers by default a review score from reviews in their own language, once a game has more than 2,000 reviews and 200 in that language.

Of the 323 games in our set with Chinese support and enough reviews for a Chinese score, 101 (31%) come out with a lower rating label in Chinese than overall, by our calculation from each game’s share of positive reviews: Very Positive becoming Mostly Positive, for example. For Japanese it was 23 of 52.

Some of that is taste, and some is translation quality, which Valve names as one reason the scores can differ. A careless translation can cost you in exactly the market you paid to reach. Budget for testing in the game, not just translation, and read your reviews in that language after you ship.

When localization is the wrong move

• If almost nobody is playing in English yet, a translation extends reach you don’t have. It rarely rescues a launch on its own. Look at how often quiet launches recover, and what came before it first.
• If your game is mostly wordplay, jokes or dense text, cost rises with every word and the result is harder to get right. A mistranslated joke costs more than a missing one.
• If you can’t update in every language you add, each patch note, new line of dialogue and store update now needs translating too. A language you can’t keep up turns into bad reviews.
• If your European audience already reads English, German, French or Spanish may add little. English-only games still drew them. Chinese players mostly didn’t come.

What to do next with a game that’s already out

1. Check your review languages and regional wishlists. If Simplified Chinese or Russian already shows up without support, that is your first candidate.
2. Price it by the word. Allcorrect Games, a localization studio, lists English to Simplified Chinese at €0.09 to €0.12 a word, Russian at €0.06 to €0.09, and German, French, Spanish (Spain) and Japanese at €0.10 to €0.14. A 20,000-word game comes to about €1,800 to €2,400 for Simplified Chinese before testing, which it charges by the hour.
3. Ship it as an update and announce it. Valve’s advice is to “release additional languages as an update” and use an update visibility round. A Steam update visibility round is worth spending on it.
4. Translate the store page too. Valve treats store and in-game localization separately, and players judge the page first.
5. Read the new language’s reviews for a month. Fix what they flag before adding the next language.

Publishers already do this. In our study of what publishers visibly do, 87% of indie games with a publisher supported Simplified Chinese, against 60% of self-published games. It’s also something you can do yourself for a few thousand euros. Strike Two partners on games that are already built, so read this knowing where it comes from.

How we measured it

The games are the 1,651 paid titles under Steam’s Indie tag that launched from January 2020 to December 2022 and drew 100 or more Steam-purchase reviews in their first year. They are the same games as our study of what publishers do, drawn from the listing behind our study of reviews after launch.

For each game we read Steam’s review summary in all languages, then in each of 13 languages, counting Steam-purchase reviews only, the ones Steam’s review score uses. Read September 23, 2026. Twelve random counts rechecked the same day matched exactly.

A review’s language is the one Steam files it under, which usually follows the player’s Steam language. Supported languages come from each game’s store record, read September 22, 2026. The positive gap compares each game’s reviews in a language with its own English reviews, for games with 20 or more of each.

The limits: supported languages are today’s, so a game that added Chinese in year three counts as supporting it all along. Review shares aren’t sales shares, since communities review at different rates. And because developers choose where to translate, none of this measures what translating your game would do.