When I started working at the White House, it was the first time my parents could explain my job to their friends. They didn’t really know how to describe what I did as a press staffer or email campaigner, but when I wrote for WhiteHouse.gov, that work was tangible. They knew how to talk about it.
For me, though, there were times when the opposite was true. When I was on a campaign, I always knew how to orient myself. I knew how to judge whether I was successful. If I was raising money, I was on target. If I was recruiting volunteers, it was a good day. At the White House, it was harder. I was supposed to amplify the president’s agenda, and I just never really knew if I was getting traction.
So right around the time their headlines started to become ubiquitous, I became obsessed with Upworthy.
It wasn’t just that their content was everywhere I looked. Or the fact that they were finding an audience for news and stories with a progressive point of view. I was fascinated by the idea that they were so intentional about measuring their output.
Their writers A/B-tested every creative choice they made to motivate more people to share their content. One friend who worked there told me they sometimes wrote dozens of potential headlines to figure out how to optimize a particular story. And it worked. At one point, articles from Upworthy were shared more frequently on Facebook than those from every other U.S. mainstream media outlet combined.
I wanted that kind of continuous experimentation — that kind of intentionality — to help guide the choices I was making, to let me know I was getting somewhere.
It’s been almost a decade. Despite the fact that Facebook doesn’t work the way it once did, Upworthy is still around — and their data is still fascinating and valuable. In 2017, a group of researchers approached the company hoping they could gain access to their archived experiments. Eventually, Upworthy agreed, and the academics have used it to build a massive dataset that’s now available to others.
So today, “a team of psychologists is looking at whether headlines that emphasize morality, curiosity or emotional appeals gain the most traction. A political scientist is studying whether clickbait innovations remain effective over time or whether users adapt to them and lose interest. A group of statisticians are doing metascientific research, using the dataset to understand the statistics of experimentation and develop advanced research methods.”
Kevin Munger, a political scientist at Penn State and co-lead of the study, describes the Upworthy dataset as a “knowledge windfall” — a place where researchers can generate a lot of insights at very low cost. And we all stand to benefit from it.
I believe progressives should be looking to open up more knowledge windfalls, as well.
The digital teams at Joe Biden for President, Hillary for America, and other presidential campaigns tested thousands of email, web page, and advertising variations. When their campaign ends, those insights often have nowhere to go. And many of the learnings live on only in the heads of those of us who did the work. I wish those kinds of data were available so more progressive organizers could refer back to them.
ActBlue has created the biggest fundraising dataset ever assembled. Mobilize has a massive collection of interactions from volunteers. It’s understandable that they don’t have the capacity to analyze all that information in real-time. But their historic data have value, researchers would be thrilled to examine it all, and increasingly, I think the platforms that we use as campaigners should make those knowledge windfalls more widely available.
Democratizing what we learn is one way we can all do better work — and understand whether or not we’re getting traction.
For more content like this in your inbox every month, sign up for our This is Not a Moment Newsletter.