If you translate a store with hundreds of products, you end up translating the same sentences many times. Translation memory is the tool that stops that. This guide explains what it is, how it works, why it matters and the mistakes to avoid.
What translation memory is
A translation memory is a database of text you have already translated, stored as pairs: a source sentence and its approved translation. When the same sentence, or one very close to it, turns up again, the stored translation is suggested or reused instead of being written from scratch. It remembers finished sentences. It does not translate anything new by itself.
A concrete example
Your store sells twenty mugs, and every description ends with the same care line. Once that line has been translated into French and approved, the memory holds the pair.
| Source text | What the memory does |
|---|---|
| "Dishwasher safe. Do not microwave." | Exact match: reuses the approved French |
| "Dishwasher safe. Do not use in the oven." | Close match: suggests the stored version to edit |
| "Hand-thrown in small batches." | No match: needs a new translation |
The twentieth mug gets the same care line as the others, word for word.
Why it matters for a store
Product catalogs are full of repeated text: care instructions, shipping notes, size guidance, material descriptions. Reusing approved translations keeps that text consistent in every language you sell in.
- Consistency: The same sentence reads the same way on every product
- Effort: Repeated text is reviewed once, not on every page
- Updates: When a product changes, the unchanged sentences are already done
A translation memory is different from a glossary. A glossary fixes how individual terms, such as your product names, are translated. A memory stores whole sentences. They work well together.
Common mistakes
A memory repeats whatever you put into it, so errors spread as easily as good translations.
- ❌ Saving unreviewed machine output as an approved translation
- ❌ Reusing a sentence where the context has changed, such as a different product type
- ❌ Accepting a close match without reading the part that differs
- ❌ Keeping old entries after you change your brand terms or tone
- ✅ Reviewing a translation before it becomes something you reuse
For term-level control, see how to build a translation glossary for your brand. To handle changes over time, see how to keep Shopify translations in sync.
Doing this at scale
LinguaFlow drafts AI translations for each language you add and applies your brand glossary and tone rules. You review and edit before publishing, and it publishes through Shopify's native Translations API with no theme edits, with an audit log and rollback. See LinguaFlow.