One lesson to emerge from the 2014 election cycle? Despite all their vaunted mastery of online politics, digital experts on both sides of the aisle are finding themselves deluged by all the data available to them, struggling to make use of it all during the heat of a campaign.
Take the (ultimately unsuccessful) bid to unseat Senate Minority Leader Mitch McConnell (R-Ky.). Revolution Messaging, a Washington-based digital agency, delivered online ads on the minimum wage and the economy directly to the computers and cell phones of Kentucky's union members on behalf of the United Food and Commercial Workers International Union. Once delivered, the ads allowed Revolution to learn much more about its targets: Are they hovering their cursor over the ad for several seconds before clicking on it? Are they more likely to respond to the ad on a Tuesday morning or after dinner on Sunday? Did the user turn on the sound in the embedded video? Did it take five showings before the user engaged with the ad? Twenty showings? Forty?
That data was vacuumed up and sent back to Revolution, said Keegan Goudiss, a partner and head of advertising at Revolution Messaging. But, while interesting, the data ultimately proved overwhelming, he said.
"One of the strengths and weaknesses of digital advertising is the depth of the data it makes available at any one time," says Goudiss. But "if we just sent it all in a spreadsheet to our clients, their eyes would glaze over."
Republicans don't disagree that they're being inundated. But one acquired lesson for successful data wranglers is understanding what information is worth trying to make sense of and what's just noise, said Matthew Dybwad, a partner at CRAFT, a digital agency working with conservatives and Republicans.
"The potential for data overload is very real," says Dybwad. He backs up a bit: "I don't know if there's too much data, but there's certainly more than enough."
The key, he says, "is to know what you're looking for." The field may be obsessed with metrics and numbers, but learning to understand the data, says Dybwad, is "at least one part art and two parts science."
And one hard-earned revelation that Goudiss and Dybwad agree on is that there are no great universal truths buried deep within all that data, no gleaming gem of timeless insight to be claimed by the strategist just willing to hunt the hardest for it.
"Two or three cycles ago," says Dybwad, "everyone was buzzing about what's the absolute best day to send e-mails to get high click-through rates. But that's useless." Those sort of truths can change from audience to audience, even from House district to neighboring House district, he says. What generates a flood of clicks in one community might fall flat on the other side of town.
Says Dybwad, "It's about figuring it out for your little audience."
In fact, adds Goudiss, "as we have all this data flooding back to us we're learning that [online behavior] can be really very different from person to person. The future of all this is going to be adjusting in real time based on what people are reacting to."
But that takes a willingness to plunge deep into the data deluge. What digital politicos like Goudiss and Dybwad will spend the two years working on before the next election is figuring is how to dive in without drowning.