A consistent world, across every language
A mobile game publisher needed to coordinate a global RPG release across nine target languages. Its existing localized content had inconsistent skill names, interface text that did not fit, damaged variables and dialogue that drifted from the characters’ intended voices.
Those issues crossed several disciplines. A translation could read well in a spreadsheet yet fail inside the game, while a corrected skill name could remain out of date in the store description. The project needed a shared review process for game text, engineering constraints and marketing assets.
Project scope
Game localization and linguistic QA
- Source language
- English
- Platforms tested
- iOS, Android and PC emulators
- Interface
- 11,800 UI strings
- Narrative
- 7,600 dialogue lines
- Content
- 3,200 item and skill entries
- Supporting media
- 90 minutes of cutscene subtitles
Languages in this projectSimplified Chinese, Traditional Chinese, Japanese, Korean, German, French, Latin American Spanish, Brazilian Portuguese and Russian.
How we approached the work
Prepare context and terminology
We analyzed the asset formats and collected context from the existing translation memory, game documentation and reference material. Language teams established terminology for characters, locations, skills, items and interface controls, with client approval tracked alongside each entry.
Adapt text to the game
Native-language translators and reviewers worked with a client-approved AI pretranslation workflow, translation memory and terminology checks. Separate guidance covered concise UI copy, character dialogue, item naming and marketing language. Review focused on meaning, tone, cultural references and length constraints.
Test in the build
Engineering checks covered variables, tags, file formats and text limits before in-game LQA. Testers then examined real builds for clipped text, font support, terminology, dialogue, subtitle timing and tutorial clarity. Issues were classified by severity and passed through correction and regression testing.
From preparation to delivery
- PrepareWeeks 1-2
Asset analysis, terminology and style guidance
- LocalizeWeeks 3-6
Translation, review, file integration and engineering checks
- ValidateWeeks 7-9
In-game LQA, correction, regression and handoff
What the client received
- Nine localized language packages
- Language-specific terminology resources
- Content-specific style guides
- LQA issue reports
- Regression test reports
- Final build validation records
A traceable handoff for the release team
The nine-week project delivered nine language packages with reusable terminology and style resources. Translation, engineering checks and in-game testing were managed within the same delivery process.
The project record lists 1,180 issues identified and 1,150 corrected. This is a 97.5% issue correction ratio, calculated from those counts; it does not measure the proportion of test cases passed. The report does not specify the disposition of the remaining 30 issues.
- 1,180
- Issues identified
- 1,150
- Issues corrected
- 97.5%
- Issue correction ratio
1,150 ÷ 1,180, rounded to one decimal place. This figure describes the recorded issue correction ratio, not a regression test pass rate.
Project questions
What did in-game LQA cover?
The project covered text display, font support, variables, terminology, dialogue, cultural context, subtitle timing and tutorial clarity inside the supplied builds.
Were marketing assets included?
Yes. The scope included store-page and marketing copy so that customer-facing language could be checked against the terminology used inside the game.
