3 Tips From BuzzFeed on Interpreting Short-Form Video Analytics

By: Justin Joffe

March 15, 2018

Figuring out what video content performs best on each platform can cause even the most level of heads to spin.

While an overnight change in a social platform’s algorithm might leave some startups in ruin, Chuck Lee, director of video growth analytics at BuzzFeed, says that short-form video provides brands with a chance to read the digital tea leaves and see what content resonates best.

“The media landscape is not getting easier in America,” says Lee. “Specifically, we are really invested in developing mid- and long-form IP. Short-form, for us, is a way to test and validate IP that we’re developing in-house. The opportunity for us is to use these platforms to get a very dynamic signal around what creative [content] speaks to our audience.”

Lee shares three tips for communicators hoping to make the most out of short-form video analytics.

Chuck Lee, Director of Video Growth Analytics, BuzzFeed

There’s no Silver-Bullet Metric

Content creators are always looking for the lone measurement hack that can propel their work to go viral and ‘break the Internet,’ but this strategy is hardly the best way to produce successful content.

“I wish there was one key metric,” Lee says. “We are a truly cross-platform company, so we’re always trying to make sense of all the platforms on a macro level and see what the tradeoffs are as one rises and one falls.”

The best metrics reveal a platform’s specificity: Each platform has different key metrics creators should be paying attention to. With that data, BuzzFeed can then figure out how to optimize a piece of content across platforms.

Watch time is a good metric, he says, but misleading. If a longer piece of content has a higher watch time, for example, it doesn’t necessarily indicate that the piece is better received than a shorter video.

“It’s one thing if I abstractly say watch time is the key way to growing organic reach. But watch time is super hard for content creators to control; it’s a composite of retention and reach,” Lee says. “We have to find metrics that creators and people making the content can focus on. In reality, that key metric may not be controllable.”

Focus on What Is Controllable

Each platform has its own native metrics, and you can use old archived content to figure out what resonates the most. But the key is to focus on metrics you can optimize, such as shares.

“Pay attention to key metrics that are eminently controllable,” Lee says. “For us, we realized well before Facebook’s move to MSI (meaningful social interactions), our content was differentiated from downstream and even upstream competitors, because it was eminently shareable.”

While shares may not be a perfect metric—especially now, as Facebook de-emphasizes branded content in News Feeds—it’s a more quantifiable metric for video creators than other pieces of data, Lee says, such as watch time or views.

“If they were [focused on] watch time, they might just make a 40-minute video,” he says. “And if they were focused on views, they might just focus on scandalous titles.”

Make Your Video Do the Most Jobs Possible

BuzzFeed firmly believes in the “Jobs to Be Done” paradigm developed by Clayton Christensen, a professor at Harvard Business School. Christensen’s disruptive-innovation theory says that the more jobs a product can help a customer complete, the more value that product has.

Figuring out what those jobs are begins with focusing on the human need your product (or in this case, your short-form video) is trying to fulfill, rather than casting a wide, abstract net.

“Validate what the audience is looking for,” says Lee. “‘Cultural Cartography’ is a fancy way of [understanding] how we map and categorize all the content we produce. How do we laser-focus on those jobs? We collect a lot of metadata around that, in addition to all the key metrics.”

Part of this data collection involves an understanding that the needs of customers are seldom monolithic. Still, he data BuzzFeed collects around its short-form video informs how it approaches mid- and long-form video as well.

“People might look to our content to connect with other people, go someplace they haven’t gone before, look up how to cook something,” says Lee. “In the early days for us, it was generally [about] formats that are easy to reproduce, like taste tests. People react to eating Ethiopian food or Swedish canned fish.”

Follow Chuck: @ChuckMLee

Follow Justin: @Joffaloff