The good news: What CAT can do for you.

The good news: What CAT can do for you.

With “The Good News”, I intend to start an English language series that deviates from the widespread negative tone that marks many translator’s blog posts. Is our “industry” (more fittingly: our craft, our art) challenged by low-grade dump work from Asia, by large corporate “language service providers” who don’t have a clue about the product they are selling or by “colleagues” who I would rather not call “translators”? Sure, but a lot of other business sectors face similar problems and likewise dance their own Limbo under the bars of quality and price.

I’d like to start with CAT, or computer aided translation. For translation professionals, this is a no-brainer, because the vast majority of us don’t remember the time when we were without it. Our customers, on the other hand, either don’t know about CAT or have a very hazy notion of what it does and what it is good for. So let’s take the quick tour of why CAT is good for translators and translation buyers alike, and who profits from it in which way.

Limbo crazy bambou

 What does the acronym “CAT” encompass?

Computer Aided Translation, in its wider sense, encompasses anything a computer can do to help a human translator do his job. This broad definition starts with simple things like the spell checkers and grammar/style checkers provided by MS Word (or the lovely open-source LanguageTool), it includes tools that count text or working time, electronic dictionaries and a lot of other applications which ease a translator’s work. In a more narrow definition, it refers to applications that have specifically created for the core translation process: Translation Memories, multilingual Terminology Databases and language Quality Assurance tools. These three are often sold together as one “Translation Environment Tool(TEnT) – much like an Integrated Development Environment (IDE) integrates several tools for programming.

What do TM, TB and QA actually do?

A Translation Memory (TM) splits any text into small segments (often sentences, sometimes whole paragraphs, sometimes parts of sentences called phrases). During translation, it stores each original segment together with the translated segment. And whenever the translator comes across the same (“100% match”) or a very similar (“partial match”) segment again, it will offer the corresponding translated text to the human user. Today’s TM systems can automatically adjust numbers and units when doing this, and often colour those few words which differ from the saved version. Thus, a TM differs from Machine Translation (MT) in that it returns the work the translator himself (or another human) has already done, it does not spit out something an algorithm has cobbled together. This also means that TMs are most effective if texts are repetitive or if the current text is very similar to prior texts. On the other hand, if texts are very unique and creative, TM technology is mostly useless and can even hinder the creativity of the translator (especially marketing texts often need to be written from scratch in the target language, because the message itself has to be adapted to the new audience).

A Term Base (TB) is a bit like a multilingual dictionary, which offers target language entries for easy inclusion whenever it detects a source entry in the current working segment. The difference between a dictionary and a term base is that dictionaries collect words or expressions while terminologies collect meanings or concepts. If done the right way, the terms in a term base have been carefully selected and verified to represent the standard, normed expression for an object or idea inside a specific context. Often, Term Bases include part of speech and gender information, usage examples in context, authoritative sources which have coined or used the term, images or multimedia files to clarify, and an assessment of the authority of the term, e.g. a range from “mandatory” to “recommended”, “alternative”, “unassessed”, “discouraged” and “forbidden” use of a term. A translator always hopes that his client has an existing terminology he can use when translating, because terms for the exact same meaning may vary by country, industry, company or even department (e.g. Engineering calls the product by its technical term, “juice squeezer JS-2452”, while Marketing only knows it as the “Fruizer®”). If the client doesn’t have his own multilingual terminology management, the translator examines (or “mines”) the text before beginning translation and researches the best equivalent in the target language(s). For smaller project, translators often add terms on the fly while translating.

A Quality Assurance Tool (QA Tool) assures the quality of the translation (duh!). More specifically, it warns the translator if the target text is much longer or shorter than the original, if numbers, units, variables or code tags are missing or incorrect, if terms have been translated as mandated by the Term Base, if spelling, grammar and style are correct, if all ©, ™ and ® signs have been included and if proper names have been reproduced correctly. Depending on the project, QA tools can also check for other requirements like controlled language.

So what are the benefits?

CAT technology has several benefits which are being marketed quite differently to translators and translation buyers.

Speed is, in my opinion, the most important benefit for translators and their clients alike, because Time is Money. In many scenarios (repetitive texts!), the technologies described above can raise and even multiply the hourly output of a translator. For you, this means you get your translation faster and reduce time to market – and the translator can take on higher-volume projects or more projects. And now read carefully: Given the usual ratio of translation costs (three to four digits) to project budgets (five digits or more), the time gained is more valuable to the customer than to the translator by a factor of 10x!

Quality is almost equally important for both the translator and his client: As the translator’s TM, TB and QA rule-sets grow and become ever more refined (and adapted to his clients’ needs), the savvy language-weaver will raise customer loyalty and grow his reputation for quality work, attracting new customers. But the same can be said for his client, too: If you can impress your partners and customers abroad with flawless marketing copy, accurate manuals and readily understandable communication of any type, you will be able to measure your ROI on translation costs by the number of foreign partners and clients you can inspire to trust in you instead of your competition, who might have a product that is on par with yours – but who delivers a horrible manual with it and advertises, by mistake, with a local swearword.

Consistency is another effect of CAT technology. When similar phrases are translated in a similar way due to Translation Memories, you get consistent texts, which augments readability. And here’s the benefit: Easy texts are more likely to be read, and only texts which are actually read can boost your business objectives. On a smaller but equally important scale, Term Bases and QA tools assure consistent use of your proper names, product names and technical terms, which makes it absolutely clear what you are talking about.

Client: “Neat, you have made an investment in technology, you deliver more swiftly and more accurately, can I have your ROI now?”

With those three benefits already working for you, you can immediately stop falling for the big lie of large language service providers and CAT tool vendors that CAT technology would lower translation costs. You might expect lower translation rates if you provided your translator with a CAT tool license and delivered useful, high quality translation memories and term bases to him. In this case, yes, you have provided your translator with advantages that you paid for, and thus you might as well reap the fruits of this advantage. In the vast majority of cases, though, the translator is the one who has made the investment in his costly professional software, and it is his own hard work which has created and filled his TM and TB databases.

So please forget about the “lower rates due to CAT tools” meme immediately. Just think about your answer if your client came up and said: “Neat, you have made an investment in technology, you deliver more swiftly and more accurately, can I have your ROI now?” I bet your answer would lie somewhere between a polite “No, thanks. May I show you the door?” and something considerably more rude, because this client would already have considerable free added value to his service without taking your increased profit. So don’t try to haggle on the basis of CAT use, all right?

Do other benefits come to mind? Just post them below, I’ll be happy to comment on them or to include them here!

Christopher Köbel (Portrait)
Christopher Köbel

Inhaber von DeFrEnT Christopher Köbel. Fachübersetzer für Deutsch, Französisch und Englisch für die Branchen IT, Web, Maschinen- und Anlagenbau, Kunststoffe, Industrie 4.0. Allgemein ermächtigter Übersetzer für Französisch und Englisch. Mitglied im Bundesverband der Dolmetscher und Übersetzer (BDÜ).


The good news: What CAT can do for you. | DeFrE… Posted on6:02 - 24. July 2014

[…] I'd like to start with CAT, or computer aided translation. For translation professionals, this is a no-brainer, because the vast majority of us don't remember the time when we were without it. Our customers, on the other hand, …  […]

The good news: What CAT can do for you. | CAT |… Posted on6:12 - 30. July 2014

[…] I'd like to start with CAT, or computer aided translation. For translation professionals, this is a no-brainer, because the vast majority of us don't remember the time when we were without it. Our customers, on the other hand, …  […]

Colin Mansell Posted on17:32 - 30. July 2014

So how much do you produce using CAT and when translating “by hand”?

Conservatively, I can do 5,000 words a day without any CAT tool, and I find Trados or Wordfast can not compete with this level of production.

I would love to know what your experience is, and of others who may be interested in this topic.

    Christopher Köbel (Portrait)
    Christopher Köbel Posted on18:02 - 30. July 2014

    Hello Colin,
    I am reluctant to throw around productivity figures, both because I am not used to the “per word” counting scheme and because words/characters per hour vary widely with each project: True, some jobs can be written down effortlessly, but for others, I need to research some obscure terminology or I might spend an hour or two to optimise a marketing text’s key line – such activities quickly spoil any “top speed”. With some jobs (marketing!), CAT can be nigh-useless for leverage and only help with the QA, with other jobs (product catalogues, I am looking at you!), they can take care of 40% of the text with all those price tags, article numbers, similar product descriptions and recurring “Buy now!” and “-x% off!” yells. With one 120 page technical text that I would have spent 2-3 weeks of work with, I could align prior versions into a TM and deliver in just 3 days. I still hit the keyboard, but I know colleagues who use speech-to-text technology to dictate their translations, and they achieve 3-5 times their typing speed in certain circumstances.
    My average speed when translating, with CAT, on a low-match (0-10%) text? 2500-3000 words per hour, I might guess. But as I said, this varies widely. Hope that helps!

3 Tipps: Office / Word-Dokumente optimal für die Übersetzung vorbereiten | DeFrEnT Posted on12:31 - 23. February 2016

[…] wie XLIFF. Über die Vorteile von CAT-Tools für Sie und Ihren Übersetzer können Sie hier mehr erfahren (Artikel in Englisch), aber die Essenz ist: Sie erhalten in der Regel schnellere […]

Leave a Reply