You Won’t Believe What Upworthy Is Planning Next
Here’s what the viral site is looking at as it prepares to expand.
Upworthy announced plans to expand its editorial scope Wednesday, turning to its readers for ideas about where to go next.
Speaking at a Social Media Week event, founders Eli Pariser and Peter Koechley said they’ll be conducting a reader survey over the next few weeks to determine exactly what areas they’ll cater to as they build out. Upworthy has 19 verticals right now.
While it’s gotten attention for its clicky headlines, shares matter a lot more in its viral model, Pariser says. Reaching out to its audience for suggestions on what they’d like to read about—and presumably, tell their friends about—makes sense.
"Headlines can draw people into a piece of content, but they’re not actually going to get something to go viral," Pariser says. "Sharing is a really high-bar ask. You’re asking people to go out in front of all of their friends and say: ‘This is something I believe in and care about and I want you to pay attention to it.’ People take that trust really seriously."
Clicks and shares aren’t mutually exclusive—more clicks lead to more Newsfeed slots on Facebook too—but virality clearly starts with the latter, he says pointing to site data.
The Right Way(s) to Find Out What Your Readers Want
Upworthy has gotten headlines of its own for its approach to analysis. The group has a rigorous A/B testing process for its article titles, and unveiled a new stat, attention minutes, in early February.
Ironically, directly asking readers what they want is a new way for them to gather data.
There are important distinctions between what readers do—measured by behavioral stats like unique visitors, page views and attention minutes—and what they want—tallied in audience surveys and polls—that aren’t being made, Pariser says.
Just because a story doesn’t drive traffic doesn’t mean people aren’t interested in it—it could just mean the story wasn’t presented effectively. Too often, publishers fail to distinguish between cause and effect here.
"You have to be collecting both of these signals and the key is to treat both equally," he says. "I think a lot of folks in the data-driven world like to believe that the behavioral self is the only real self. That’s not true. You have to feed both. You have to read behavior within the context of aspirations…It’s ‘Am I doing this right? Am I telling this story in an interesting way?'; not ‘Are they interested in this general topic area?"