+34 93 244 08 80

WooRank assistant

What is a translation memory?

What is a translation memory

What is a translation memory?

Translation memories (TM) are databases that store original texts and their corresponding translations.

How does a translation memory work?

Translation memories are integrated in CAT tools. These tools compare the source text to be translated with the units saved in the translation memory. The aim is to identify the parts of the source text that have already been translated and saved in the memory. This allows the translator to retrieve or reuse segments of the source text that have already been translated in the past. When a new text must be translated, the CAT tool consults the database, i.e. the translation memory, to check whether there are any matches between the new text and segments that have already been translated and saved in the memory.

Types of match

The CAT tool identifies the existing matches between a new source text and the translation units saved in the translation memory. There are different types of match.

1. No Match (NM)

Translation units that are completely new and do not have any match in the translation memory. The translator must translate them from scratch.

Segment with no match in the translation memory

2. Internal Repetition (IR)

Translation units that are repeated throughout the text that is to be translated.

3. Fuzzy Match (FM)

Translation units saved in the translation memory that partially match new units. Different ranges are applied to determine the match percentage of the new unit in relation to the already-translated unit. The translator must adapt the segment to reflect the differences between the two units. A good CAT tool, such as TransitNXT, will highlight the differences between the segments of the source text and the translation memory in order to alert the translator to the parts of the sentence that need to be modified before the translation is integrated into the target text.

Segment with a partial match in the translation memory 

4. Perfect Match (PM)

Translation units that are identical to those that have been translated previously. Please note that CAT tools only class segments as perfect matches when they are identical in terms of content, punctuation, numbers and format.

Segment with a perfect match in the translation memory

The advantages of using a translation memory

When choosing your translation service provider, it is important to take the tools with which they work into account, as they can provide important benefits and help you to reduce your translation costs. Below is a list of the main advantages of working with translation memories:

  • Reduction in translation costs. The main advantage of working with translation memories is that a percentage of the standard price is applied to the different types of match. This means that the larger the translation memory and the more controlled the language used in the texts, the greater the reduction in translation costs for you.
  • Time savings. The use of translation memories implies significant time savings in terms of translation processes. The translator no longer has to search constantly for the terminology used or translate the same text from scratch, but can instead consult the translations saved in the memory.
  • Guaranteed consistency. The use of translation memories ensures consistency between the translation of a new document and previous texts (especially relevant in technical translation). The translation memory enables the translator to use the same terminology and the same translations as in previous projects. The fact that the CAT tool identifies the repetitions within a document also allows these matches to be translated in the same way and thus ensures the consistency of the translation.

Tips for reducing translation costs

It is important to take into account the fact that the yield of a translation memory depends on the size and quality of the existing database. As the database expands with new translations, the matches between texts increase and the translation costs are thus reduced. In order for this to occur, the following tips must be taken into account:

  • Use controlled and simplified language when writing source texts to optimise the use of the translation memory and increase the number of matches. It is important to use the same sentences and formulations in all of your documents.
  • Use consistent terminology in all of your documents. We recommend that you establish standard terminology and that you circulate it to all those responsible for multilingual communication in your company to ensure that the same terminology is used in all documents.
  • Use the same format in all documents. As previously mentioned, CAT tools take the format into account when identifying matches between a source text and the translations in the memory. It is therefore important for all of your documents to have the same format and style. We recommend that you create templates and style or corporate design guides to ensure the same format is used.

Has this clarified what a translation memory is for you? If you still have queries, please contact us, we would be more than happy to help. Check out other posts where we discuss practical cases that illustrate the theoretical explanations on our blog.



Leave a comment