In neighborhoods across Minnesota, campaign volunteers clutching smartphones and iPads are fanning out, targeting specific houses while skipping others in what may seem like a random game of Whac-A-Mole.
But with just over five weeks before Election Day, they're meshing consumer data they bought, the state's public voter registration data and political party surveys to track down who's likely to vote for their candidate but less likely to actually show up at the polls. The goal: Convince them to turn out.
The same is happening on Facebook, YouTube and streaming sites such as Hulu, where campaigns are increasingly using digital ads to target certain demographics — for instance, by making sure a baby boomer sees a Facebook ad stating a candidate's position on health care.
Politics is even invading personal text messages, as campaigns text volunteers to mobilize or remind eligible voters to register.
More and more, campaigns are relying on hyperlocal, data-driven ways to drum up a victory. Most are measures that voters don't know about and will never see, even as they change the way campaigns are run in Minnesota and around the country.
"It makes a big difference in a state like Minnesota," said Eitan Hersh, an associate professor of political science at Tufts University, who wrote a book on the use of big data in campaigns. "These tools make it easier to pinpoint the people who are responsive to these messages."
While big data has been used throughout the 2000s, namely by former President Barack Obama, Minnesota political parties and campaigns say the efforts are getting more accurate and may be more important than ever this year because Minnesota is in play nationally. Several high-profile, competitive races are on the Nov. 6 ballot — from the governor's job and control of the Legislature to several competitive congressional seats.
"Companies and corporations have been doing this for years," said Christiana Purves, the Republican National Committee's regional communications director. "Republicans and Democrats being able to tap into this has just allowed us to be more precise."