**This article is updated regularly. The latest update was made in April 2023.**
What is translation memory and how might its use impact the speed at which you translate? These are important questions if you want to translate as efficiently and consistently as possible. I’ll answer both below, plus look in detail at the advantages of using translation memory software – and the potential downsides too.
What is translation memory software? In short, a translation memory (TM) is a database packed with words, phrases, sentences, and sometimes entire paragraphs that have been translated before. It pairs the original and translated texts together as ‘translation units’. The TM then suggests these (or automatically inserts them, depending on the TM you’re using) each time that particular phrase (or word or sentence or paragraph) comes up again.
According to the Common Sense Advisory (CSA), 70% of LSPs acknowledge using machine translation and memory translation software in their operations. With the advancement of AI and machine learning technology, more language professionals will be implementing these tools in their work. In another study conducted by the CSA Research, most businesses are aware that if you're relying on English-only content, you are leaving 47% of your total addressable market to your competitors. The increase demand for multilingual content has lead LSPs to make global communication and commerce language-focused while finding internal and external resources to make their services more accessible to the wider public.
Currently, translation memory is not the same as machine translation, which involves a computer translating a document in its entirety. Rather, translation memory software is something that aids translators in their work, enabling them to translate faster and with greater consistency. A TM is usually one element of a computer-assisted translation (CAT) tool. CAT software also provides a range of other functions, in addition to the TM, such as glossaries of technical terms (which act more like a dictionary rather than a translation memory system).
Translation memory software works by logging phrases as you use them, and recording them in both the source language and the target language. The software then uses these translation units to aid your translation as you work on the rest of the document.
One of the interesting things about using a translation memory system is that it can be shared among translators. When multiple teams are working on providing document translation services remotely, they can use a translation management system to access the translator memory of a specific client. This introduces consistency across the translation work, despite various different people contributing to the document.
Different memory translation software approaches exist, naturally, and some memory translation tools are better than others. As such, it’s important to do your homework and compare different products before committing to one. Some translation memories will automatically replace source text with target text drawn from the database, for example, while others will make suggestions.’
One handy feature that can crop up in a translation memory is a percentage to show how accurate the match between the source and target phrases is. Another is tool the use of color to highlight differences. Features such as these can make a major difference to how effective you find the tool to be. Like I say, it’s worth exploring different options.
Best Practices When Using Translation Memory Software
Since we have discussed what is translation memory and its functions, I’ve listed a couple of the best practices for using translation memory to make your work more efficient and generate consistent output, as follows:
Build a database that’s clear and concise: Many people don’t realize that translation memory servers heavily depend on how you collect and organize data. So it’s vital that you set out to assess and select translated terms and phrases and arrange them accordingly.
Regularly update your memory translation software: Part of assessing and categorizing data will be ensuring that it’s up to date with the latest terms and phrases. Any changes in language use or terminology should be reflected in your TM’s database.
Use high-quality source texts: Subsequently, the quality of your TM will depend on the quality of the source texts you use. Make sure to use high-quality source texts to ensure that your translations are accurate.
Be more consistent when using terminology: Consistent terminology is key to effective TM use. Try to use the same terminology throughout your translations to ensure that your translation memory can be used to its full potential.
Customize your translation memory settings: Different translation memory software will have different settings and options available. Customize your settings to meet your specific needs and preferences.
Take advantage of concordance searches: Concordance searches allow you to search your translation memory system for specific words or phrases. Use this feature to find translations that you have used before quickly.
Review your translated output: Always review your translations to ensure that they are accurate and appropriate for the context. Don't rely solely on your translation memory to provide you with the right phrase or term.
If you haven’t used a translation memory system before and are considering trying one, it’s time to weigh up the benefits. And, indeed, there are plenty of those. I’ve included some examples below to help highlight how valuable a decent TM can be.
Using a translation memory means you have to contribute fewer human translation hours to a project. Plus, the more work you do for the same client, the larger the translation memory grows and so the more efficiently you can translate for them. As such, using a translator memory can mean a lower cost for the client, as you’ve had to bill them for less of your time. It also frees you up to take on additional work, as you can fit more translations into your working day. Happy clients, happy you.
Having a database crammed full of suitable phrases and sentences means it is much easier to be consistent when working on large or complex projects. The translation memory system essentially automates consistency in your work. This can be a great way to quality assure your translations – and again, to keep your clients happy.
As I mentioned above, translation memories are also ideal for projects where multiple linguists, and other teams, are working on the same project. Take a large-scale software translation project, for example. You might have multilingual developers working on the software, as well as translators, project managers, and others. Having a translation memory server in place means that no matter who is translating, the work can be consistent.
If you use a translator memory to speed up the pace of your translation work, you can deliver your clients’ projects faster. In doing so, you support your clients to achieve their business goals faster. And in today’s economic climate, time is money…
Downsides of Using a Translation Memory
I’ve talked a bit about the benefits of translation memories, but there is also a downside that you need to be aware of. Like any database, a TM will need cleaning up from time to time. This is essential if the translation memory system is going to remain true to the client’s evolving content needs and quality standards.
According to CSA Research's report, generative AI can customize emails and content across different languages a lot more accurately and quickly compared to earlier forms of technology. It has led to drastic transformations in the language industry and will continue to expand the total available markets wherein English doesn't cut it. How LSPs will go about implementing these tools and technologies will depend on them and the long discussions that come after.
Like with any technology, translation memory tool isn’t perfect. Language professionals will always be needed in ensuring that the generated translated content and cultural context are accurate. Language professionals will still be needed to manage over machine translation quality assessments, which is just something to bear in mind.
Translation memory software comes on its own when you have multiple projects from the same client. In this situation, you can save plenty of time by using translation memory.
Another good time to use a translation memory system is when you’re working on a project with other linguists. In that case, the project manager will provide what you need to ensure that all of the translators can be consistent in their approach and choice of phrases.
When discussing what is the difference between machine translation and translation memor, it boils down to translation memory as a software tool that is built upon translators adding new content to its database. Meanwhile, machine translation requires no human input when translating a document.
So when you say a professional translator post-edits machine-translated content, it often refers to content that is generated from machine translation software, which is then edited afterward. A translation memory tool can be used to make the post-editing process more efficient, as they don't need to get a glossary and examine previous works as it is all in the database.
Translation memory (TM) and translation glossary (TG) are two different tools used in translation. While they may seem similar, there are some key differences between the two.
TM is a database of previously translated segments of text, such as sentences or paragraphs. When a translator encounters a segment of text that has been previously translated, the translation memory tool will suggest the previous translation, which the translator can then confirm, modify, or reject.
The translation glossary is a list of key terms and their corresponding translations or definitions. Glossaries are often used to ensure consistency in the translation of specific terms, especially those that have multiple possible translations. Some memory translation software has in-built glossaries which can be used as references as the language professionals translate content.
The major main difference between a TM and a TG is their scope, as TM focuses on previously translated segments of text, it offers suggestions for reusing them. Glossaries are more on the specific terms and provide consistent definitions and alternative translated words for them.
I hope the details above have shed some light on what a translation memory is, how it works and what the benefits are. Using translation memory software as a part of your work can save you time and support you to translate more efficiently and consistently. It means you can help your clients get to market faster and translate a higher volume of work.
Tomedes works with linguists around the globe, providing professional translation services that span a huge range of industries and professions. Many of our translators use memories daily. If you’re one of them, why not leave a comment below to share your views on how it helps you translate?