Linguistic quality assurance (LQA) is a structured check of translated content by a native speaker, ideally in its final layout, that catches errors in meaning, grammar, terminology, formatting and consistency before the content is published. It is the standard answer to “how do we catch translation errors before they reach production?” Back translation, where a second translator renders the text back into the source language for comparison, is a narrower check that confirms the meaning survived.
Neither tells you whether the copy persuades, which for a hotel, resort, attraction or travel brand is often the question that matters most.
This piece explains LQA and back translation in plain words, what each catches and misses, and when a marketing team needs more than either.
Key Takeaways
- LQA is a check for errors: wrong meaning, wrong terms, broken formatting, inconsistent names. It is essential before anything goes live, and it works best when the reviewer sees the text in place.
- Back translation proves meaning came back from a round trip. It is useful for safety, legal and policy text, and a poor test for marketing copy, which is often rewritten on purpose.
- Neither LQA nor back translation measures tone, brand voice or persuasion. Content that has to sell an experience needs a brand-level review by native writers as well.
What LQA is, in plain words
LQA is a second, independent native reader checking translated content against the source, the glossary and the style guide, and logging what is wrong. Good LQA is done in context: on the staged web page, the app screen or the print proof, not in a spreadsheet of sentences. Text that reads well in a document can truncate on a button, break across a line badly or render the wrong characters once it is in place.
Reviewers usually log issues by category and by severity, so the team can see at a glance what must be fixed before launch and what can wait. Typical categories are:
- Accuracy. Mistranslations, omissions and additions.
- Language. Grammar, spelling and punctuation.
- Terminology. Whether property names, room categories and programme names match the approved glossary.
- Style. Whether the text follows the brand’s style guide for that language.
- Locale conventions. Dates, times, currencies, units, addresses and phone formats.
- Layout and function. Truncated text, broken characters, untranslated strings, links and buttons that go to the wrong place.
What back translation is, and where it helps
Back translation means asking a second translator, who has not seen the original, to translate the target text back into the source language. You then compare the two source-language versions. Where they differ in meaning, something may have gone wrong.
It is useful where literal fidelity is the point: safety and emergency information, terms and conditions, cancellation and privacy policies, and anything a regulator or legal team will read.
It is a poor test for marketing. A strong hotel tagline in Japanese or Korean has often been rewritten so that it works for that reader, so it back-translates as something quite different from the English. A flat, literal line will back-translate perfectly. Used as a quality gate for brand copy, back translation can reward the weaker version.
Back translation proves the words came back. It cannot prove the brand arrived.
Translation checkers and automated QA
Automated QA tools, often searched as a “translation checker”, are a fast first pass, not a reviewer. They are good at mechanical checks: numbers and prices that do not match the source, missing tags or placeholders, glossary terms not used, untranslated segments, double spaces and inconsistent punctuation.
Run them before human LQA so the native reviewer’s time goes on what only a person can judge. They cannot tell you whether a sentence sounds natural, whether the register suits the market or whether a description makes anyone want to visit.
What LQA catches, and what it does not
| Usually caught by LQA | Usually not caught by LQA |
|---|---|
| Mistranslated facts, policies and prices | Copy that is correct but flat |
| Inconsistent property, room and outlet names | A register that is accurate but wrong for the brand |
| Wrong date, currency and unit formats | Taglines and wordplay that no longer do their job |
| Truncated or broken text in the layout | Pages ordered for the source reader, not this one |
| Grammar, spelling and punctuation errors | Missing information this market expects |
| Departures from the glossary and style guide | Search wording locals do not actually use |
The left column is about errors. The right column is about effect. A page can pass every check on the left and still fail with the guest it was written for.
When a marketing team needs more than LQA
When the content’s job is to persuade, LQA is necessary but not enough. Hero copy, room and destination descriptions, campaigns, menus, CRM emails and booking-flow text all depend on tone and emotion that an error check does not measure.
IPPWORLD defines transcreation as creative translation: a process that adapts the emotional expressions and contextual relevance of the source message, beyond the literal meaning of words. For content like this, the quality question shifts from “is it correct?” to “is it ours, here?”, and the reviewer needs to be a native writer who knows how brands in that market speak.
There are two common situations:
- New content. Build the brand question into production: a brief with the tone of voice, a writer working in the target language, LQA in context, then a review by someone who markets to that audience. Our guide on how to check hotel translation quality sets out a seven-point checklist.
- Content already live that feels off. If pages passed QA but are not landing, more LQA will not find the problem. A translation audit looks at register, brand voice, cultural fit, what is missing and search wording as well as errors.
Machine-translated content raises the stakes on both. If a tool produced the draft, LQA can confirm it is accurate, but somebody native still has to judge how it reads. Our guide to AI translation or human writers covers where machine output fits.
How LQA fits into a localisation workflow
LQA works best as a scheduled step before go-live, not a rescue after it. IPPWORLD runs six steps on every engagement: Understand the Project Brief, Pre-Transcreation Study, Transcreation, Client Review, Update and Approval, Final Delivery. A dedicated Translation Memory per client, built and reused across projects, means the terms and decisions settled in one round carry into the next.
For FWD, the Omne mobile app was localised into seven Asian languages, and linguistic quality assurance is what allowed a simultaneous launch across its Asia-Pacific markets. For Millennium Hotels & Resorts, brand content for China, Taiwan, Japan, the UAE, Spain, Italy and France went through quality assurance before launch.
Conclusion
LQA is how you stop errors reaching guests. Back translation is a useful extra check where literal meaning is the point. Neither is designed to tell you whether your copy sounds like your brand or makes a guest want to book.
Use LQA on everything before it goes live. For the content that sells the experience, add a native, brand-level read as well.
Next step
If your translated content is accurate but not landing with the guests it targets, an audit is the sensible first step: our guide to when your translated marketing feels off explains what one looks at. When you are ready, get in touch with IPPWORLD to explore options.
Frequently Asked Questions
What does LQA stand for in translation?
LQA stands for linguistic quality assurance. It is a structured check by a native reviewer of translated content against the source, glossary and style guide, ideally in its final layout, with issues logged by category and severity before publication.
How do we catch translation errors before they reach production?
Run automated QA checks first for numbers, tags and glossary terms, then human LQA in context on the staged page, app screen or proof. Record fixes in the glossary or Translation Memory so they do not recur.
Is back translation a good way to check marketing translation?
Not on its own. Back translation confirms meaning survived a round trip, which suits safety, legal and policy text. Good marketing copy is often rewritten for the target reader, so it back-translates differently from the source, while a flat literal version back-translates perfectly.
Can a translation checker tool replace a human reviewer?
No. Automated checkers are good at mechanical issues such as mismatched numbers, missing placeholders and unused glossary terms. They cannot judge whether a sentence sounds natural, whether the register suits the brand, or whether a description persuades.
What is the difference between LQA and a translation audit?
LQA is usually part of production and checks for errors before content is published. A translation audit usually looks at content already live and asks whether it works for the market, covering tone, brand voice, cultural fit, missing information and search wording as well as accuracy.