Skip to content

Contently: This Data Scientist is BuzzFeed’s Secret Weapon


April 21, 2014

By Sam Petulla

Ky Harlin has been called BuzzFeed’s secret weapon. The viral darling’s Director of Data Science has his finger on the pulse of the Internet like few others. And as one of the company’s original 25 employees, he’s had it there for a long time.

BuzzFeed’s successes keep coming. Their latest content experiment — quizzes — commonly generate over a million views each, with some quizzes racking up tens of millions of shares.

Those are numbers most contenky-harlin-e1389639327577t creators would kill for, but it’s not luck. Science — data science in particular — drives a lot of BuzzFeed’s content strategy.

Intrigued, we talked to Harlin to learn about his scientific game plan.

What are some of the most interesting ways BuzzFeed uses data science or machine learning.

Here are a couple examples. First, think about this: A list is in itself a mini pool of content.We treat each individual item in a list almost like its own article. So, we’ll try to really figure out what people are engaging with and turn a list of 45 items to a list of 25 items without the duds, reordered to make it most likely to share.

A more recent experiment: We’re working on clustering user characteristics and content. It’s the same idea you see on other sites, like Netflix, with a ton of content and recommendations. The idea is to cluster articles in buckets, and it’s really interesting because that reveals these latent topic interests people have. Like we’ll find clusters that show people who are interested in Jennifer Lawrence are also interested in penguins.

See full story: