Startups

How Comcast decides what you’ll watch next (hint: it’s not algorithms)

Big data matters, but it's actual humans who drive what's featured on Comcast's on-demand screen. A team that includes a former People Magazine editor explains their approach to “playing to the zeitgeist.”

Photo: Holly Quinn
Todd Gold and Julie Zied watch TV like it’s their job — because it is.

They follow all the network TV hits, like The Blacklist, Gotham and Jane the Virgin (“I’m obsessed with it,” Zied said) on TVs in their office or through online screening rooms. And when they’re not watching TV, they’re talking about it.

Gold and Zied are on a small team responsible for curating the on-demand selections for Comcast’s X1 TV guide platform. It’s essentially Comcast’s answer to the Netflix browsing screen. When deciding what to watch next, Netflix users browse through titles produced by the company’s algorithms. If you’ve been on a “suspenseful buddy TV show” kick (and really, who hasn’t?), Netflix recommends more of the same.

Putting your finger in the air and feeling the buzz.

Comcast forgoes the big data approach for a more human touch.

Instead of using algorithms, a team led by VP of Programming Andy Hunter chooses which TV shows to highlight on the on-demand screen. The cream of the crop goes on the “Featured” section at the top of the on-demand screen, which only spotlights four TV shows at a time. Other sections include TV shows driven by days (“Sunday Night Shows”) and special, timely ones like Veterans Day picks.

It’s updated every day, sometimes multiple times a day. (That’s something Comcast can do because of its cloud-based X1 platform. With the old platform, Comcast only updated on-demand once a month.)

andy hunter team

Left to right: Comcast’s Todd Gold, Andy Hunter and Julie Zied. (Photo by Juliana Reyes)

So what’s the end game?

It’s not really about converting cord-cutters, Hunter said. It’s about getting more people to watch on demand and getting those who are already watching on demand to, well, watch more TV.

“If we can grow reach and grow consumption, we’re doing the right thing,” Hunter said during a recent interview at the Comcast Center.

Seventy percent of Comcast cable subscribers use on-demand, he said, and subscribers are on track to have watched 2.8 billion hours of on-demand TV by the end of the year.

Gold, who’s based in Los Angeles, and Zied, who’s based in Philly, are tasked with featuring shows that Comcast subscribers will “react to,” as Hunter put it. They also choose the photos associated with featured TV shows and write copy that will hopefully lure viewers (“Disappear into the full miniseries” is one tagline they used for Houdini).

They use subscribers’ viewing habits, Nielsen ratings, social media and big events like season finales and plot twists to drive their decisions. Personal taste is a factor, Gold said, but it’s more about “playing to the zeitgeist.”

Gold, who spent two decades as an editor at People Magazine, described his job as the “old-fashioned putting your finger in the air and feeling the buzz.” In other words, it’s not something an algorithm can do.

Still, the personalized, algorithm-driven Netflix approach has its upsides. Not every Comcast subscriber digs The Blacklist. Hunter said the team is moving toward incorporating some personalization into the X1 platform.

Companies: Comcast
41% to our goal! $25,000

Before you go...

To keep our site paywall-free, we’re launching a campaign to raise $25,000 by the end of the year. We believe information about entrepreneurs and tech should be accessible to everyone and your support helps make that happen, because journalism costs money.

Can we count on you? Your contribution to the Technical.ly Journalism Fund is tax-deductible.

Donate Today
Engagement

Join our growing Slack community

Join 5,000 tech professionals and entrepreneurs in our community Slack today!

Trending

West Philly residents can get a free laptop by completing a digital skills training

Nerd Street founder on what’s next after near bankruptcy: ‘It’d be naive to say we’re out of the woods’

This exec spent 30 years at one company, and thinks more people should do the same

AI in action: How InsightFinder AI and Robin AI transform IT and legal workflows at major organizations

Technically Media