ATA

Featured Articles

Find a Translator or Interpreter
Search for:

Featured Article from The ATA Chronicle (January 2015)

Product-Centric Translation: What It Is and How Best to Handle It
By Uwe Muegge

Anyone who has worked in the language services profession long enough and on complex and large projects is probably familiar with the following scenario. You have a multi-document, high-volume, high-priority technical translation project. A team of highly trained and experienced linguists and project managers works on that project, adhering to the latest industry standards. And yet, after delivery, client feedback is mostly negative. Wherever they look, the in-country reviewers find errors in meaning, terminology, stylistic inconsistencies, as well as other types of errors. How is it possible that when the right people do “all of the right things” the outcome is still wrong? While there is a wide variety of reasons technical translation projects fail, I would argue that there is one common root cause for many of these failures. And that root cause is ignoring the special needs of what is probably the most common type of commercial translation project today: the product-centric translation.

What Is Product-Centric Translation?
I first used the term “product-centric translation” in 2014 in an article in which I described how my approach to teaching translation technology differs from translation teaching in the literary tradition. Of course, my teaching at that point was informed by my work as a member of the operational leadership team of a large global language services provider. That was when I realized that practically all of the projects with which I was involved had several very basic characteristics in common.

1. The text to be translated is linked to a product. Unlike with “general” technical translations, where the text stands on its own, in product-centric translation projects there is a direct link between the source text and a product or service. A translation of an article describing the pros and cons of a product in a trade publication would be a good example of a “general” technical translation. While the content of the article might be highly technical and involved, there is typically no direct connection between the article and the product it discusses. The manufacturer of the product usually does not commission an author or translator to write this type of material. A user manual, on the other hand, is a perfect example of a product-centric translation. The defining characteristic of a product-centric translation project is that it is tied to a product launch or the release of a new version of a product. Depending on the product and the international market where that product is to be sold, the translation may even be considered a part of the product.

2. The text to be translated is linked to other texts that are linked to the same product. Typically, the launch of a commercial product in foreign markets involves not only multiple documents but also multiple document types. In addition to the user manual, there can be other manuals, such as installation and service manuals. There might also be marketing collateral, such as brochures, catalogs, and a website that require translation. And if the product has a software component and a display, which more and more do these days, there will also be software strings and most likely online help and tutorials that need translation/localization.

3. The text to be translated is linked to previous versions of that text. As if product-centric translation were not already challenging enough, here is another complication general technical translation projects typically do not have: versioning! Over time, products evolve: into new versions, variants, or entire product families. As the product evolves, so must the texts associated with that product--and their translations.

Understanding the Business Requirements of Product-Centric Translation
Introducing the concept of “product-centric translation” is not an academic exercise. This type of project is actually very common in the real world of commercial translation. At the same time, the implications of product-centric translation projects do not seem to be understood fully by both the client and the language services provider.

The sections on the following pages provide some of the requirements that typical product-centric translations have. By the way, certain general translation projects have some of the same requirements. However, the consequences if these requirements are not met are much less severe in general stand-alone translation projects than they are in product-centric translation.

Requires Consistent Use of Correct Terminology
Since users experience a product through its features and functions, it is of critical importance that the correct terms be used consistently in all translations describing the product. If a term on a label on a localized device does not match the term that is used in the translated user manual, the user experience is negatively impacted. To ensure the best possible user experience, terminology must be used consistently within and across all documents associated with a given product.

Terminology Challenges: In many product-centric translation projects, ensuring terminological consistency is a big, and often unresolved, issue. It can be hard for individual translators to use terminology consistently in a single document if that document is large enough and the linguists are working under a tight deadline. But more often than not, product launches involve many different documents that are typically translated by multiple linguists who often work independently of each other. Since every linguist has his or her own personal preferred terms (think USB stick, USB drive, Flash drive, etc.), inconsistencies are an inescapable reality. And I am not even talking about incorrect terms that are introduced due to a lack of subject-matter expertise.

Can translation editors fix terminology-related problems? If there is a budget for an editor, and if there is enough time for a complete edit, and if the project is small--maybe. But in a major launch with a tight deadline and budget, it can be next to impossible for a human reviewer to identify and eliminate variants and synonyms for possibly thousands of terms.

Terminology Solutions: For terminology to be used consistently across multiple documents that are being translated by multiple linguists, comprehensive, project-specific multilingual terminology must be available at the beginning of the project. That means that either the client provides terminology data with the source text, or the language services provider creates project-specific term bases immediately after the receipt of the source. Today, however, building large multilingual glossaries within a short period of time is not the challenge it used to be. (My team regularly creates multilingual term bases containing thousands of terms in a matter of days).

The other part of ensuring terminological consistency in large, complex, multilingual translation and localization projects may seem trivial, but it is not: making sure that the (freelance) linguists actually use the terminology data in their translation memory environment.

Translation memory systems automatically look up and provide the translation for terms that are available in the translation memory system’s terminology management module. However, even today, many professional linguists are used to managing terminology in spreadsheets, where terminology lookup is a time-consuming and error-prone manual process. That is why it is not enough just to provide linguists with terminology data. It is equally important to educate linguists on terminology management best practices and to make the submission of terminology-related quality assurance records a mandatory part of every delivery.

Requires Text that Can Be Recycled
Product-centric translations typically involve a lot of text. The authoring departments that create the source documents for translation go to a lot of trouble to reuse text. Sentences, paragraphs, and even entire pages will be recycled across different documents and different versions of the same document. There are two important motivators for having translators reuse previously translated full and partial (fuzzy) matches when working on a new project.

1. Recycling drives down cost. One of the main features of translation memory systems is the automatic identification of previously translated sentences and the suggestion of corresponding translation proposals. Reusing an existing translation typically requires less effort on the part of linguists than translating from scratch. This is why many language services providers offer staggered discounts for 100% matches, high fuzzy matches, and possibly low fuzzy matches, as well as internal repetitions in a new source text. Recycling text in source texts, in combination with the use of the proper tools and processes during translation, can result in dramatically lower average costs per page for product-centric translations than for other types of technical translations.

2. Recycling ensures that reviewed and locked content stays that way. Because of the direct link between the translated text and the product, the creation of product-centric translations often involves comprehensive quality assurance measures, including multiple edit and review steps. In other words, in order to ensure the accuracy of product-centric translations, both the language services provider and the buyer often put a lot of effort into each translated sentence.
In this type of business environment, once a translated sentence receives final approval, that translation is “locked.” “Locked” translations must be reused exactly as reviewed to avoid having to undergo the time-consuming and expensive review process again.

Text Recycling Challenges: Since reusing text is so pervasive in product-centric translation, linguists often receive only the new or changed parts of a document to translate instead of the document in its entirety. The practice of sending only isolated segments for translation can make it very difficult for linguists to do their job. Translating disconnected segments is particularly difficult if the language services provider or translation buyer does all of the leveraging in-house and does not send a translation memory to the linguist because there are no good matches. In this type of scenario, many linguists are tempted to work outside of a translation memory system. (“If there are no matches, why bother?”)

While using a word processor for translation generally has many drawbacks, this tool is particularly unsuitable for translating product-centric texts because it makes it very difficult to be consistent. In a translation memory, source segments and target segments are stored together, which makes finding previous translations easy. When translating in a word processing program without involving a translation memory, there is no connection between the source and target. In fact, the source and target are typically two separate documents. While not impossible, it is very difficult in a word processing program to reference previous translations, especially
those created outside of the current project, for consistency purposes.

Text Recycling Solutions: For most language services providers, using translation memory systems to process large projects that have a lot of matches and/or repetitions is a no-brainer.

But in order to ensure maximum reuse of translations with consistent terminology, every single part of a product-centric translation project must be processed in a translation memory system. That rule applies to even the smallest project (think just a few sentences), and projects that show no matches during analysis. If product-centric translations are created outside of a translation memory system, inconsistencies in style--and terminology--are all but unavoidable.

Requires Short Turnaround Times/“Can’t-Miss” Deadlines
Technical translations in general are deadline-driven, but product-centric translations even more so. Just picture a product launch on an international market that has to be postponed because the legally required user manual in the local language is not available. Delays in product-centric translations can cost the manufacturer millions of dollars in missed opportunities.

Turnaround Challenges: As companies small and large aspire to release their products on the global market at or shortly after their launch on the domestic market, the pressure on turnaround times increases. One common strategy to deliver large product-related translations sooner is to distribute the work to more language services providers and linguists. However, doing so often results in inconsistent translation, and to have those inconsistencies fixed by an editor takes time--time that is in scarce supply to begin with.

Adverse effect reports and product recalls are other examples of product-centric translation where time is absolutely of the essence. In these types of projects, linguists are faced with the task of creating a translation within the shortest period of time that meets the highest quality standards.

Turnaround Solutions: When the pressure to do it faster is high, so is the need to do it right the first time. Put in a team translation context, this means: a) developing project-specific terminology as early as possible, b) translating in a translation memory system, and c) using style guides. When all translators involved in a product-centric translation have access to the same project-specific terminology and translation memory, and use the same style guide, review and re-work should be minimal. And if editors and reviewers do not have a lot to change, translations can be delivered faster. This is particularly true if all contributors to a project, such as translators, editors, and reviewers, use the shared resources of a server-based translation memory or translation management system.

Summary
The translation of materials that are connected to a product, or what I call “product-centric translation,” is different from generic technical translation. Product-centric translation involves the translation of dozens, if not hundreds, of interrelated texts, and they all support the international launch of a product or service. As product-centric translation projects are often not only large but complex, not to mention highly time-sensitive, following industry best practices is essential for the success of this type of project.

To achieve maximum output at consistently high quality (i.e., minimal re-work after the initial translation phase), linguists must work within a translation memory environment, even for the smallest job. Having all linguists use translation memory systems not only ensures stylistic consistency and the protection of “locked” content, translation memories also automate the lookup and use of terminology.

If there is one thing that is critically important in product-centric translation projects, it is the availability of comprehensive multilingual terminology at the beginning of these projects. Equally important is the use of project-specific terminology by all members of the linguistic team--and the automatic recognition of terminology during translation (and review!)--as is only possible in a translation memory system.

References and Resources

Muegge, Uwe. “Teaching Localization: 6 Practices that Make a Difference
(Part I),” GALA Blog (February 6, 2014), http://owl.li/DuQfg.

Muegge, Uwe. “Ten Good Reasons for Using a Translation Memory,” tcworld
Magazine (January 2010), http://owl.li/Ez0Vp.

Muegge, Uwe. “Ten Good Reasons Why You Should Validate Translated
Terminology,” tcworld Magazine (July 2011), http://owl.li/Ez2jg.

Muegge, Uwe. “Ten Things You Should Know About Automatic Terminology
Extraction,” The ATA Chronicle (September 2012), http://owl.li/EyZAy.

About the Author
Uwe Muegge is vice-president of strategic technology solutions for OmniLingua Worldwide, LLC. He has more than 15 years of experience in translation and localization, having worked in leadership functions on both the vendor and buyer sides of the industry. He has published numerous articles on translation tools and processes and taught computer-assisted translation and terminology management courses at the college level in both the U.S. and Europe. In addition, he has been active in several standardization efforts; most recently as chair of ASTM F43.03 subcommittee on language translation. Contact: umuegge@omnilingua.com.



 

>