The GOP’s Data Surge

by Eliana Johnson
Republican strategists work to improve data analytics in advance of the midterm elections.

For two days last week, Republican strategists, political consultants, and digital gurus convened in a sleek, wood-paneled conference room on the third floor of New York City’s Standard hotel. Their mission: to reverse the fortunes of the Republican party by leveraging voter data, technology, and public opinion to win elections.

More than a year removed from an election year in which Democrats used data to gain insights that allowed them to swing a handful of races, including at the top of the ticket, those gathered were determined that the GOP do better in this year’s midterm elections. Among the attendees: National Republican Senatorial Committee executive director Rob Collins; Alex Lundry, the Romney campaign’s director of data science; GOP pollster Kristin Soltis Anderson; Johnny DeStefano, a former aide to John Boehner, now serving as the president of Data Trust, the organization providing a warehouse for Republicans’ voter files; and representatives from Google, Facebook, and AOL. 

Top party officials have readily admitted that the GOP has work to do in the realm of data collection and digital targeting. The next step, Collins says, is to “start getting better,” and that is easier said than done. The goal of last week’s meeting, in the words of one attendee, was to “get everybody on the same page” ahead of the midterm elections.

Campaign strategists and consultants were urged to stop handwringing about how to allocate resources between mail, television, and the Web and instead to use voter data to determine the best method for persuading individual voters. That, speakers explained, allows campaigns to allocate resources efficiently: to deploy volunteers to the neighborhoods where their efforts are most likely to pay off or to pay for advertisements only where persuadable voters are most likely to see them.

That is what the Obama campaign and Democrats did with ruthless effectiveness in 2012.

The Republican National Committee’s post-election report recommended creating a new data platform accessible to conservative groups and candidates, bringing data crunchers into the RNC, and ​identifying strategists and donors to build an external data-analytics institute. There are signs of progress. The RNC has moved to centralize its voter data through Data Trust and made key hires, among them Andy Barkett, formerly of Facebook, who now serves as the RNC’s chief technology officer. Barkett plugged the committee’s latest efforts to top-dollar donors at events including an American Crossroads retreat in October and a September fundraiser hosted by New York Jets owner Woody Johnson.

Multiple sources say that tension between Data Trust, effectively a subsidiary of the RNC, and Themis, a voter database funded by the Koch brothers, has dissipated under donor demands that Republicans not duplicate their efforts. Representatives from the two organizations have been in conversation over the past six months to discuss the possibility of cooperation but have yet to reach an agreement. The Koch-affiliated group, which employs just under 50 people, including in-house statisticians, has touted the superiority of its data. Data Trust, created in 2011 to shoulder the cost of building and managing the GOP’s voter file — the party was looking for ways to deal with debt left in the wake of Michael Steele’s chairmanship — has evolved more slowly but, according to a GOP strategist, is finally “coming out of its slumber.” Among donors, the strategist says, there is “no appetite” for competition between the two groups. “The idea that there’s going to be competition at the realm of raw data is a bit silly,” says Republican strategist Patrick Ruffini, the CEO of the digital-media firm Engage.

Nonetheless, the RNC’s announcement in May that Liberty Works, a group backed by Karl Rove ally and Silicon Valley venture capitalist Dick Boyce, would build the platform through which Data Trust could manage its data, effectively put Rove and the Kochs, two of the most influential forces on the right, at loggerheads. The RNC’s announcement later last spring that the short-lived partnership had come to an end has led, sources say, to some cooperation between Data Trust and Themis.

The biggest hurdle is translating these efforts into electoral victories, and there the signs are less inspiring. When numerous polls showed Democratic candidate Terry McAuliffe with a comfortable lead against Republican Ken Cuccinelli in the Virginia governor’s race, the RNC simply stopped investing in television ads and voter-mobilization efforts. “We had twelve months to get the polling right,” says Yale political scientist Luke Thompson, a consultant to the NRSC. Yet the party, relying on internal polls that had McAuliffe cruising to a seven-point victory, spent just $3 million on the race, compared with the $9 million it expended in 2009. “McAuliffe’s people knew all along it was going to be a two-point race,” Thompson says.

Indeed, Politico’s Alexander Burns chronicled how the McAuliffe team, using sophisticated data modeling that mirrored that performed by the Obama campaign the previous year, foresaw a tight race and systematically located the voters it needed to drag The Macker over the finish line. Democratic volunteers asked voters to sign postcards pledging to vote and mailed them back the week before the election. Another social-science technique employed by the campaign was to hit swing voters with experimentally refined messages; it knew which voters were persuadable and on what issues.

This sort of targeting is something that, since John Kerry’s loss to George W. Bush in 2004, the Democratic party, in a cumulative effort and over the course of eight years, worked tirelessly to perfect. Ironically, it was first deployed by GOP strategist Karl Rove, who introduced micro-targeting into politics in 2000 to sharpen the Bush campaign’s ground game, leveraging voter data to identify supporters across the country. Democrats, reeling from two consecutive losses and smarting in particular from ceding the technical advantage to Republicans, committed to redouble their efforts. Howard Dean, who served as chairman of the Democratic National Committee from 2005 until 2009, helped to usher in some of this progress. He “recognized the potential of the Internet to mobilize large sums of money through small-donor contributions,” says Thompson. “Those donations — and Internet outreach in general — provided a platform for political big data.”

The AFL-CIO, too, created the Analyst Institute in 2006, an organization housed in the labor union’s headquarters and devoted to determining the most effective ways to turn out grassroots voters, which it did largely by analyzing social-science experiments. The Obama campaign in 2012 not only used the group’s research findings, it embedded Analyst Institute scientists within its ranks.

Democrats also centralized their voter file when former Bill Clinton deputy Harold Ickes founded the firm Catalist in 2006 to serve as a data warehouse. By 2008, most of the party’s candidates were also inputting data into a network known as NGP VAN (voter action network). Democratic campaigns on every level were learning from and building off each other and providing future campaigns with an increasingly clear and evolving picture of the American electorate.

By the time the 2012 campaign rolled around, the Obama campaign had a comprehensive handle on the data available to it — so much so, Slate’s Sasha Issenberg has written, that while pundits “talked in the abstract about reassembling Obama’s 2008 coalition,” for many campaign staffers “the goal was literal.” According to Issenberg, the Obama campaign “began the election year confident it knew the name of every one of the 69,456,897 Americans whose voters had put him in the White House.”

Meanwhile, Yale’s Thompson says that Republicans simply failed to realize that data analysis and get-out-the vote efforts “were as important as they proved to be.” Thanks in part to bad data, they thought the country was, at a fundamental level, favorably disposed to their candidate. “Most of the Romney camp genuinely thought they were going to win,” Thompson says. “These were serious, smart folks, but the 2010 midterms gave them an inaccurate picture of the electorate.”

The task before the GOP now is similar to the one that confronted Democrats a decade ago: to effect a cultural shift within the Republican party, toward what GOP strategist Ruffini calls “data-driven decision making” all the way down the ballot. While the RNC has invested in key hires at the national level, Ruffini worries that it’s not happening at the state and local levels and that the changes being discussed in Washington may not be scalable. He and others acknowledge a “talent gap” that Republicans, for all the thinking they have done over the past year, have yet to confront. “We are vastly outnumbered in the number of field operatives and the number of people doing data analytics,” Ruffini says, “and we always have been.” Solving that problem, more than anything else, may prove crucial to the party’s future.

— Eliana Johnson is media editor of National Review Online.

Did you like this?