Translation memory tools

A translation memory is a linguistic database where all of your previous translations are stored and “remembered” for future use. If a source sentence (i.e. one that is in the original text) has featured in previous translation projects and appears to be identical or similar to a new source sentence, the translator can use this stored translation rather than translate it from scratch. It works in tandem with other translation technology to improve the consistency and overall quality of the translation. At Planet Languages we have a dedicated team of project managers who are experts in using translation memory to achieve maximum efficiency and maximum quality with every translation project.

Looking for translation tools?

Benefits of using translation memory

Reduced cost

One of the biggest benefits of using translation memory (TM) tools for the client is reduced cost. At Planet Languages we can offer you a ‘weighted’ word count, which means we analyse your text for TM matches and any repetitions, so we can work out how much of your content has been translated before or how much of it is similar.

Any translations that have been stored in a TM work 10x harder. They can be reused again and again, altered and updated as your business grows. Every translation is an investment in future projects!

Terminology

Another significant benefit of TM tools is the improved consistency of key terminology. You can rest assured that a word will be handled in the same way each time, as your translators will be able to see how key terminology has previously been translated. Consistency is important for things like your brand voice and product names.

However, we still value human judgement: we ensure that translators are still able to judge whether the TM suggestion is still the appropriate choice in any new contexts. Don’t worry – you still have the final decision on whether terminology should be altered or not, and you will be able to decide whether you want any changes to be implemented in future translations or whether they might be one-offs.

Faster turnaround

Another bonus of translation memory use is that it greatly reduces the work of the translator, along with reducing the likelihood of questions, queries or edits further down the line. Linguists and project managers can use it to answer any terminology questions that pop up. The less time spent editing and answering queries, the less time you spend waiting for a delivery. Let’s get your document back to you as quickly as possible!

Looking for translation tools?

Using translation memories

For the most part, you won’t have to worry about the intricacies of using a translation memory tool. Planet Languages’ expert team will be able to look after it for you, ensuring it’s accessible to every relevant translator and updated after each project, or according to your instructions. If you want a more in-depth look into how it factors into the workflow of a translation process, read on.

Analysis

Our computer-assisted translation software organises the source text (the original one that hasn’t been translated yet) into segments. These are normally sentences but can sometimes be shorter clauses or phrases; things like headings or taglines are going to be shorter, for example.
Your project manager will use these segments to work out whether there are any repetitions, and make sure that they aren’t included in the final word count.
The software will check for any matches with previously used source text segments. If the segments are identical, they’ll be classified as 100% matches. If the segments before and after are also identical, they get bumped up to 101% matches so that the linguists involved know that the context is likely to be the same. If the segments are similar but not identical, they are classified as a fuzzy match and given a rating between 0–99%.
Your dedicated project manager will use the number of matches above 75% to calculate a weighted word count and calculate the appropriate amount of translation memory discount.
As your translator works, any 100% or 101% matches are already filled in. They read these and decide whether they are still appropriate and make any necessary changes. Any matches that are “fuzzy” are suggested by the software, and the translator can choose the relevant parts to implement as they write the new text.
Our project managers check the translation against the translation memory and ensure consistency has been maintained where possible.
We deliver the translation back to you, in less time and at less cost!

Main obstacles

However, there’s a time and a place for translation memory and it may not be suitable for every project. Content with a more technical focus, where consistency and clarity are more important than style, is perfect for these tools. However, literary texts or shorter marketing texts might not be improved as much. This is because any repetition in these texts may be limited or even unintentional, and these texts may benefit from a more creative approach, or even transcreation.

For localization, it’s also important to make sure that separate memories for different language variants are maintained – make sure that your Castilian Spanish and your Latin American Spanish stay separate!

Types of translation memory systems

There are various ways that you can use TM tools and various versions and formats to use. You might have heard of the big names like Trados, MemoQ, Wordfast, Memsource, etc. Ask your translation memory manager which they use and what benefits each one can bring to you.

There’s also the choice between desktop-based (local) systems or server-based systems. Which you use depends on how you use it. Planet Languages uses a server-based system, which means that the relevant translators can access it while completing your work and project managers can also access it to check that everything is as it should be.

Seamless translations with translation memory tools

Our in-house, qualified and experienced translation memory specialists ensure that individual client glossaries and translation memories are carefully built and maintained, so that our clients get the best possible advantage from the relevant translation tools. We work with the leading TM packages and successfully use these tools to achieve significant cost savings, increased consistency and faster turnaround times for our clients.

As soon as we have gained a clear understanding of your requirements, we will be able to determine whether we can successfully apply TM to your specific project and come up with the best solutions for your needs.

Looking for translation tools?

FAQs

1 What is Translation Memory?

A translation memory is a linguistic database that stores source sentences and their corresponding translations as segment pairs for future use. Usually a segment is a sentence, but it could also be just a few words, like a heading. The translation memory ‘remembers’ each segment that is translated and stores the source/target segment pairs in the translation memory database. If an identical or a similar segment comes up later, it does not need to be translated from scratch. Instead, the translation can be retrieved from the translation memory, and once checked for suitability in the context, reused. Translation memory can be used to improve consistency, shorten turnaround times and reduce costs.

2 What is a CAT tool?

CAT stands for Computer-assisted translation (or Computer-aided translation) and it is a piece of software or web-based platform that provides the right environment and contains all the relevant resources to support translators with their work, such as translation memories, glossaries (termbases) and quality assurance tools. More recently some CAT tools also integrate with machine translation solutions allowing translators to use machine translation if needed and then post-edit the output directly within their CAT tool.

3 Is Translation Memory suitable for my content?

Ideally all translated content should be stored in translation memory databases as you never know when you might need to retrieve it for future projects. However, there are content types that benefit more than others from the use of Translation Memory. Large volumes of work that contain repetitions are an ideal candidate as the use of translation memory will definitely help with maintaining consistency, shortening turnaround times and reducing costs. Another great candidate is content that is being regularly updated or reused in various projects.

Highly creative short texts that you are not likely to be reusing again can certainly be processed using Translation Memory, but the advantages will not be as significant as in the previous case due to the lack of repetitions.

4 How do you know what content was previously translated?

We always analyse your content against your translation memories and generate a report illustrating exactly what segments are found in the memory. The similarity between the new content and the content in the memories is marked as a percentage, with 100% being an exact match and everything under 100% a fuzzy match, meaning that there are changes between the new content and the content in the translation memory.

5 What are exact matches?

An exact match means that the segment was translated before and that there are absolutely no differences between the segment from the content we are translating now and the segment present in the translation memory. You might think that if a segment has an exact match in the memory, it will not need to be translated again and you would be almost right. Although having an exact match in the memory is very useful, this does not necessarily mean that it will be the right translation for your new context. At Planet Languages we always ensure that exact matches are carefully checked for in-context accuracy and text coherence.

6 What are fuzzy matches?

Whenever similar but not identical matches are identified in the translation memory, these are referred to as fuzzy matches. These matches generally range from 50% to 99% in terms of similarity to previously translated content. Everything under 75% is often considered entirely new text as the level of similarity is too low to be of any relevance.

7 How is the translation memory discount calculated?

Upon receipt of your project, we will analyse the content against your translation memories and provide you with the relevant discount figure based on the figures in the translation memory report. The content repetitions and various matches – 100%, 95-99%, 85-94%, 75-84% and under 75% – are charged at different fractions of the normal word rate based on the similarity to the content in the translation memory. For example, 95-99% matches will be easier to amend than 75-84% matches so the rate will be lower.

8 How is Translation Memory different to Machine Translation?

A translation memory is a linguistic database that stores source sentences and their corresponding translations as segment pairs for future use. Whenever translators work on similar content to the one that is stored in the translation memory, they are provided with matches that help them maintain consistency and increase their productivity.

In contrast, a machine translation system is a piece of software that was trained to automatically translate text from one natural language to another. In this case the linguist is automatically provided with a translation that they then post-edit to ensure the level of quality is in line with the client’s requirements.