Localization QA for eLearning: What to Check in a Translated Course

Questions answered by this guide

• ?What is localization QA for an eLearning course?
• ?Is localization QA the same as checking the translation?
• ?What breaks when a course is translated?
• ?What does A11yCheck check when comparing a localized course to its source?
• ?What localization problems does A11yCheck not detect automatically?
• ?How do I review a translated course against the original?

What you'll learn

• Localization QA is not translation review: it asks whether the localized course still works, not whether the wording is good
• The most common failure is text expansion: translated text is often 20 to 35 percent longer than English, so it clips or overflows
• A11yCheck compares the source and target rendered courses on 10 mechanical checks; it does not judge translation quality and cannot read quiz answer keys
• Terminology consistency, meaning drift, cultural fit, variables, and branching parity are important manual checks A11yCheck does not automate
• Run a separate, full Accessibility review on each localized version, localization comparison does not re-run accessibility checks

What's inside

Instructional designers and reviewers checking a localized Storyline or Rise course against its source

1. 01What localization QA is
2. 02Why a translated course breaks
3. 03The areas to review
4. 04What A11yCheck compares automatically
5. 05What stays a manual check
6. 06How to run a source-vs-target review

Why it matters

A course that passed review in its source language can fail in a translated version for reasons that have nothing to do with translation quality. Text that fit now overflows its box. A slide that existed in the source is missing from the target. The player's Next and Submit buttons are still in English. A font that covered Latin characters silently falls back to a system font for Cyrillic or Arabic. Voice-over that was timed to an animation is now a different length. None of these are caught by reading the translation, and none are caught by scanning the source course alone. Localization QA is the review pass that compares the two.

When you'll encounter this

• A course has been translated into one or more languages and you own the sign-off before each version is published.
• A translation vendor returns a localized Storyline or Rise export and you need to confirm it still works.
• A stakeholder reports that the German or Japanese version looks broken but the English one is fine.
• You are setting up a repeatable process for releasing a course in several languages.

Examples

Best practices

• Treat localization QA as its own pass, separate from translation review and separate from the source-language accessibility and quality reviews.
• Design the source course for the longest language you expect to ship: give buttons, tabs, headings, and text boxes room to grow, or let them auto-size.
• Translate the player interface separately from the content. In Storyline this is Settings, then the Labels tab; in Rise it is the course Labels. These strings are not part of the translated content export.
• Compare the localized course to the source, not to your memory of the source. Run them side by side.
• After localization, run a full, separate Accessibility review on each language version. Translation can introduce missing language metadata, untranslated accessible names, and reading-order problems that only exist in the target.
• Keep a short glossary of key terms per language so reviewers can check terminology consistency by hand, since no automated check verifies it.

Common mistakes

• Assuming that if the translation reads well, the course is fine, layout, navigation, and media problems are invisible to a linguistic review.
• Checking only the target course and never comparing it to the source, so missing slides and dropped media go unnoticed.
• Forgetting the player and navigation labels, which stay in the source language unless translated in their own separate place.
• Treating every difference as a defect. A localized course is supposed to differ from its source; localization QA looks for differences with a real consequence, not all differences.
• Expecting the localization comparison to also cover accessibility. It does not; accessibility is a separate review per language.

Practical checklist

• The localized course has the same number of slides or lessons as the source, or every difference is explained.
• No translated text is clipped or overlapping at 100 percent size.
• Player and navigation labels (Next, Previous, Submit, menu) are in the target language.
• Non-Latin scripts render in an embedded font, not a silent system fallback.
• Dates, numbers, and currency appear in a form appropriate to the locale.
• Right-to-left languages read right to left, with focus order matching the visual order.
• Links point to locale-appropriate destinations where those exist.
• Captions, transcripts, and voice-over are present in the target language and roughly the same length as the source.
• Quiz questions and answer options are all present and translated.
• Variables and dynamic text resolve correctly, with no %placeholder% visible to a learner.
• A full Accessibility review has been run on this language version.

Frequently asked questions

Summary

Localization QA checks that a translated course still works the way the source does, structurally, visually, and for assistive technology, without judging the wording. The dominant failure is text expansion. A11yCheck compares the source and target rendered courses on 10 mechanical checks and reports only differences that have a real consequence; it does not assess translation quality, terminology, variables, branching, cultural fit, meaning drift, or text in images, and it does not re-run accessibility checks, run those separately per language.