Translation Quality Assurance Shouldn’t Start With Translators

The surprising truth is that translation quality is less about talent acquisition and more about talent enablement.August 11th, 2016

Yext

The difference between the best Italian translator in the world and the second-best Italian translator in the world is exceedingly small. But what about the difference between the best Italian translator in the world and the thousandth-best Italian translator in the world?

As it turns out, that margin is still barely perceptible.

The surprising truth is that translation quality is less about talent acquisition and more about talent enablement. Companies that surround qualified translators with the resources they need to succeed are almost always assured a suitable final product.

So for businesses struggling under the weight of ballooning review costs and lingering quality concerns, the response shouldn’t be researching how to replace their translators. The real question is what could their localization teams be doing to put translators in a better position?

Start With Style Guides & Glossaries Ambiguity is the enemy of efficient translation. The more a business leaves open to interpretation at the start, the more likely they are to find themselves correcting costly misinterpretations down the road. Creating comprehensive glossaries and styles guides is the best way for brands to standardize their stance on key linguistic scenarios and keep all of their translators on a consistent script.

A glossary is a catalog of the unique terms and preferred phrases that are routinely included in company collateral. Each entry in this list of brand names, service features, and relevant acronyms must be defined, translated, and associated with a part of speech as well. Learning that a client favors the word “alert,” for example, is of limited value to a translator until they also know whether it’s intended as a verb or noun.

If glossaries define what language should be used, then style guides dictate how language should be used. For instance, how does your company feel about addressing the audience with a second-person point-of-view? What side of the oxford comma debate did you settle on? Do your landing page headlines appear in title case?

These seemingly small distinctions add up over time and ultimately form the foundation of brand voice. Compiling, defining, and translating these stylistic preferences ensures that voice remains the same regardless of the translator recruited to employ it.

Lastly, it’s important to remember that these guides and glossaries need to be easily accessible and continuously updated if they are going to yield practical results. Cloud-based translation management systems are especially advantageous in this scenario, as they provide centralized storage and facilitate real-time version control.

Take Advantage of Translation Memory The potential for error, however small, is introduced each time translators are given an entirely new string with no prior context. By that same logic, error rates will decrease when the number of instances in which translators have to start from scratch is reduced.

Translation memory brings the latter principle to life by serving as a database of previously approved translations. When translation memory tools detect an exact match between an old translation and a new string, that recommendation is instantly sent to translators. Depending on the translation management system being used, teams can even implement workflows where exact matches bypass translators altogether and skip directly to review.

When translation memory tools detect a partial (or “fuzzy”) match between an old translation and a new string, a similar recommendation is routed to translators to ensure they aren’t starting from scratch.

If the phrase “20% Off Men’s Jackets & Jeans” were previously translated, for example, translation memory would discourage translators from treating the string “20% Off Women’s Jackets & Jeans” as entirely new.

Not all cases are so obvious, of course, and translators should use their best discretion when accepting or overriding translation memory suggestions. But the takeaway here is that consistency improves, cost decreases, and efficiency increases any time redundant work can be eliminated.

As with the guides and glossaries that came before, though, there is a certain amount of administrative hygiene required to make translation memory a true quality assurance ally. Translation memory files need to be carefully paired to relevant projects and shared with every agent, freelancer, or colleague tasked with the associated work. Until every collaborator is contributing to the expansion and maintenance of translation memory, no collaborator will experiences its full value.

Offer Clues With Visual Context Letting translators see how their target text will appear on a website or within a mobile app is one of the simplest and most powerful ways to reduce quality assurance issues. Contextual confusion suddenly gets solved, intended phrasing is immediately understood, and formatting restrictions are instantly made clear.

Unfortunately, what could be a standard practice is still a privilege for translators who are given nothing more than a spreadsheet full of text strings from their client. When they see the word “run” in Cell C24, they have no way of knowing whether that phrase appears on a website button that executes a command prompt or sits in the main navigation menu of a fitness app.

The value of visual context will grow even more apparent as we move forward in the mobile era. Optimizing content for smaller screens is challenging enough as it is, and translating blind only adds to the backlog of UX issues. As a result, in-context translation is quickly climbing onto the wishlists of translators, reviewers, developers, and multilingual marketers alike.