Client Talk: Buzz Feed

Welcome to Client Talk, where we chat with the people who buy (or should buy) language services. When do they say professional translation is worth it? 

By talking with clients outside the sales environment, we hope to uncover what really drives purchasing. Each month offers a different profile to learn from. The challenge is to find patterns from one issue to the next. Together, what do these interviews tell us about how clients see our industry as a whole?

Gabi D’Addario is based in Los Angeles, California, but spent time in Brazil growing up. She “learned Portuguese as a result” and, thanks to a year spent living in Italy, speaks Italian as well. Russian, Chinese and Arabic are on her hope-to-learn list.

The client

As senior manager of international publishing, D’Addario oversees translation for BuzzFeed, a cross-platform news and entertainment network. Based in the United States, the company markets to readers in Canada, Australia, the United Kingdom, India, Brazil, Mexico, Spain, France, Germany, Russia, Japan and the Philippines. “The content published by these editions is in the primary language or dialect of the market,” she says. “Internal communications within these offices takes place in the local language.”

The client’s need and how they meet it

According to D’Addario, “BuzzFeed has an in-house global adaptation team that translates content across all verticals and formats across the company: entertainment and news, long and short-form video, articles, illustrations and memes.”

Bilingual adaptation editors in BuzzFeed’s major international markets make up the team. D’Addario says they curate and edit all translation for the country where they’re located and “approach curation through a combination of in-market editorial expertise and analytical models that suggest translations based on previous successes and what is trending on domestic channels.”

During a December 2017 presentation at AI Summit in New York, vice president of data science Gilad Lotan broke the model down: it’s a machine learning algorithm that uses social network, site traffic and other data points to choose which news articles to localize. “We’re obsessed with giving every piece of content a fair opportunity to reach the right audience,” he said.

So what gets translated?

Once the algorithm gods shine down on certain content, BuzzFeed sends it to an outside provider, has bilingual staff perform the translation or starts from scratch with subject matter in-language. D’Addario says it’s usually the latter: “The majority of content published to our non-English, international editions is created by in-house teams in their primary language or dialect.” She does add that for certain types of content, BuzzFeed outsources “an initial translation from a vendor or freelance translator,” but wouldn’t answer when we asked which type.

Once translation is complete, D’Addario says, “An in-market, adaptation editor edits it for tone, style and accuracy. In other cases the translation is completed from scratch by an in-house adaptation editor. Our adaptation editors review any translation that we outsource before publication.”

So 1-5, how important does BuzzFeed think professional translation is?

A full 5. Despite outsourcing relatively little work to pros, D’Addario claims translation is critical to BuzzFeed’s international branding. “Translation allows us to make the best BuzzFeed content available around the world. It’s a fast and cost effective way for us to expand our reach and take advantage of all content created by BuzzFeed teams, no matter what language it’s in. BuzzFeed’s business team also relies on translation to globally distribute branded content for certain advertisers.”

Is there an emerging pattern?

Over recent months, Client Talk has seen its share of buyers who could solve the vast majority of their translation difficulties with existing industry tech — like translation memory or translation management systems (TMS).

But BuzzFeed presents a new first: the company is aware of industry tech, but decided to develop its own. Lotan’s algorithm is not a TMS, but Buzzfeed did consider at least one existing industry solution before his team designed it. And since the program predicts a piece of content’s ROI — an offering a top TMS would include — the company has essentially built something it could have bought off the shelf. Maybe these system vendors should ask themselves why BuzzFeed chose to invest time and money into building something that data-driven solutions could give them for a low monthly fee.