Localized/Copy & language

The 12 Most Common Mistakes in Japanese App Localization

12 specific mistakes that mark an app as foreign-made to Japanese users — punctuation, register, script choice, layout — with real examples and fixes.

9 min read·2,000 words·Updated May 2026

This is the list we use as a QA checklist before signing off on a localization. Each item is a mistake we see weekly in apps shipped by foreign teams. Each one is a small thing that, on its own, might not break the app — but stack two or three of them and the user concludes "this app wasn't made for me."

For background context, see the complete guide to Japanese app localization.

1. Half-width punctuation in Japanese sentences

Japanese punctuation is full-width: 「、」「。」「！」「？」. Half-width Latin punctuation , . ! ? looks visibly wrong when it appears in a Japanese sentence.

❌ ファイルを保存しました. ✅ ファイルを保存しました。

This is one of the easiest tells that a localization was done by a non-native or by automated tooling that didn't normalize punctuation. Set a CI check that flags Latin punctuation in Japanese strings.

2. Spaces between Japanese words

Japanese does not space-separate words. Translation pipelines that go through English-segmented intermediate steps sometimes leave word-level spaces.

❌ 写真 を 編集 する ✅ 写真を編集する

There are limited cases where a half-width space is acceptable (between a katakana phrase and a kanji phrase for visual breathing room, e.g., 「ライブ 配信」), but these are stylistic choices made deliberately. Random word-spacing is just wrong.

3. Quote marks: " instead of 「」

Japanese uses corner brackets 「」 for quotation. Latin double quotes are wrong.

❌ "ログイン" を タップ ✅ 「ログイン」をタップ

For nested quotes, the convention is 『』 inside 「」.

4. Mixed registers within a screen

Covered in detail in keigo in app copy, but worth restating: a single screen with both 「保存します」 and 「保存する」 reads as written by two different people. Pick one register and stay.

5. Romaji where katakana is expected

Foreign personal names in romaji break the visual rhythm of Japanese text.

❌ John さんからメッセージ ✅ ジョンさんからメッセージ

Brand names in romaji are fine — Japanese readers expect "Apple" "Google" "Nike" in Latin letters. But personal names, common nouns, and any non-trademark proper noun should be in katakana.

6. Untranslated English in error messages

Error messages are the strings most likely to be missed by a translation pipeline because they're often defined deep in error-handling code rather than in the main strings file.

❌ Network connection error. Please try again later. ✅ ネットワークに接続できませんでした。しばらくしてから再度お試しください。

Untranslated English in an otherwise-Japanese app is the single fastest credibility killer. Audit every catch block and error-mapping table.

7. Western date formats

The default MM/DD/YYYY is alien in Japan. Local conventions:

  • 2026年5月10日 (formal)
  • 2026/05/10 (most common in UI)
  • 2026.05.10 (occasional in date-stamped UI)

Never use 5/10/2026. Either it parses as ambiguous or it parses correctly and feels American.

The longer treatment is in date, time, currency, and number formats in Japan.

8. Wrong currency display

Yen has no decimals. ¥1,000.00 is wrong; ¥1,000 is right. Do not use JPY as a prefix in user-facing UI; use ¥ or 円.

❌ JPY 1,000.00 ❌ ¥1000 ✅ ¥1,000 ✅ 1,000円

The 「円」 suffix and 「¥」 prefix are interchangeable; pick one and stay. Comma separators are standard at every 3 digits.

9. Truncated buttons due to text overflow

English "Save" is 4 characters. Japanese 「保存」 is 2 characters in practice but takes ~4 characters of width because each Japanese character is wider. "Continue" → 「続ける」 is 3 characters but visually similar in width.

The trap: longer English strings translate to longer Japanese strings, but the conversion is not linear. "Submit" → 「送信」 (shorter). "Sign up with Email" → 「メールアドレスで登録」 (longer than expected). Always review buttons in context — see handling Japanese text expansion.

10. Wrong line breaking

Japanese line-breaking rules differ from English. A few key conventions:

  • Avoid breaking immediately before particles 「を」「が」「は」「に」「で」.
  • Avoid orphaning a single character on its own line.
  • Prefer breaking at clause boundaries.

Modern iOS and Android handle most of this with their text engines. But if you implement custom text layout (e.g., in a game), you need to apply Japanese line-break rules explicitly. See Japanese typography for mobile apps.

11. Onboarding images with English embedded text

Strings in strings.xml get translated. Strings inside PNG/SVG assets do not.

A typical onboarding flow has 3–5 illustration cards, each with a tagline burned into the image. If the localization brief did not call out these assets, they ship with English text inside an otherwise-Japanese onboarding flow. Catch this by walking through every screen during QA, not by trusting the strings file.

12. Receipt and email templates in English

The app is in Japanese. The user pays. The receipt arrives in English from Stripe / Apple / Google's default template. The transactional emails about subscription renewal arrive in English. The push notifications from your CRM tool arrive in English.

This is a coordination failure, not a translation failure. Every system that sends user-facing text needs to be plugged into the same localization pipeline, with locale routing based on the user's account locale.

The systems most often missed:

  • Stripe / Braintree receipts (Stripe supports JP locale; flip the flag).
  • Twilio SMS templates.
  • SendGrid / Postmark transactional emails.
  • Marketing automation (Braze, Iterable, OneSignal): each campaign needs JP variants.
  • App-side receipts shown after IAP — these are yours to localize; Apple's are Apple's.
  • Customer support chat widgets (Intercom, Zendesk): chat-bot welcome messages need JP versions.

Honorable Mentions

A few mistakes that didn't make the top 12 but are worth knowing:

  • Using 「お客様」 in a casual app. It's the right "you" pronoun in formal contexts but is suffocating in casual ones. See keigo in app copy.
  • 「シェア」 instead of 「共有」. Both technically mean "share" — the former is conversational, the latter is what apps say.
  • Calling a Japanese user 「ユーザー様」. This title doesn't exist in natural Japanese. Use 「お客様」 (formal) or drop the title entirely.
  • Using 「アカウント様」 or 「メール様」. Don't add 様 to non-people. We have seen this in real apps — it's bizarre to natives.
  • Calling support 「サポート」. Acceptable but 「お問い合わせ」 is more natural for the menu item.
  • Ranking emoji literally. "Top user 🥇" → 「トップユーザー🥇」 is fine in social apps, weird in finance.
  • Mistranslating "free" as 「自由」 instead of 「無料」. "Free trial" is 「無料トライアル」, not 「自由トライアル」.

A Self-Audit You Can Run in 30 Minutes

If you have an existing Japanese localization and want to know how clean it is, do this:

  1. Pick 10 random screens.
  2. Screenshot each in Japanese.
  3. For each screenshot, check:
    • Punctuation is full-width.
    • No spaces between Japanese words.
    • Quote marks are 「」.
    • All visible English is intentional (brand names, etc.).
    • Date/time/currency formats match Japanese conventions.
    • Register is consistent within the screen.
    • Buttons fit without truncation.

If 8/10 screens pass cleanly, the localization is in good shape. If 4/10 pass, you have a real problem and the app is leaking trust.

Where to Go Next

If you want us to run this audit against your shipping Japanese app and return a written report with screenshots and line-numbered fixes, get in touch.