If political races have become data wars, conventional wisdom has it, the Democrats clearly have the advantage in 2014 and 2016. After all, the stockpiles of data from President Barack Obama's two campaigns have been deposited in the party's armory alongside the software to put it to good use.
But it's not so simple. While the party as a whole navigates a newly treacherous political landscape -- none other than Nate Silver predicted the Democrats could actually lose control of the Senate -- individual campaigns across the country may struggle to use something as big and complex as Obama's data trove, which was built for a nationwide campaign. Think of taking a fire hose to your flower garden, or asking the local marina's security guy to dock a submarine.
The fact is, even if the political topics had stayed the same, most state legislative or U.S. House candidates can't possibly use all the data that's been given to the party. And, just as important, a single candidate simply doesn't have the resources to hire more than one internal data handler, much less replicate the 50-plus crew that steered the Obama analytics ship.
"People read about the Obama juggernaut," said Tom Bonier, co-founder and partner of Clarity Campaign Labs, a data-analytics firm that works with Democratic campaigns. "If you're a state legislative caucus director … you look at that and say, 'In no way is that anything that we could possibly do.'"
An embarrassment of riches
Despite the party's mission to provide unified, fresh data and underlying standardized technology platforms for all, there's a limited pool of practitioners who know how to put all this stuff to use for the more than 6,000 races this year.
And there's an awful lot of stuff.
The Democratic National Committee serves as the central hub for the party's voter-file database, known as VoteBuilder. With regularly updated voter data from secretaries of state at its core, the database has been enhanced during the past decade with commercially available consumer and lifestyle data, geographic information and survey responses from volunteer interactions. Any Democratic campaign can tap into VoteBuilder using an interface provided by data-management and software firm NGP VAN, nicknamed in party circles "The Van."
During elections, supplementary data flows in daily from the field via mobile tools carried by canvassers or through pipelines transporting information from social media and digital-ad systems.
The Obama campaign has handed over its volunteer information to the DNC, in addition to its highly prized "IDs." The IDs identify which voters campaigns should contact and contain information such as how much money a voter donated online or what he said in response to questions about issues like immigration while chatting with a volunteer in his neighborhood. Voters are allotted "party points," a number displayed in their DNC voter file. That information, updated each night in the heat of election season, is particularly important in places like Virginia, a swing state where people don't designate party affiliations when registering to vote. More than 80 party points, for example, is a strong indicator that a voter will support Democratic candidates.
That's the good news.
"The bad news is we need to keep on training more and more," said Stu Trevelyan, CEO and president of NGP VAN.
The party says it's working on that. The DNC held a training event in March attended by around 80 state party and campaign staffers. Other groups operating in support of progressive campaigns and organizations -- most prominently the New Organizing Institute through its RootsCamp tech-training pilgrimages -- also have helped with that goal in recent years.
Over 1,000 state party staff and activists were trained to use the database platform by the DNC last year, both in person and through webinars, learning how to do things like run queries using SQL, a database-management-programming language. In many cases, those students are learning to teach others in their regions how to do the same.
The high price of smart data
But even if a small campaign has the money and wherewithal to hire a voter-file manager, most will have limited analytical resources.
Building custom data models to gauge voter sentiment and craft media plans isn't cheap. The handful of firms that offer analytics services to Democrats, including Clarity, TargetSmart Communications, Civis Analytics and BlueLabs, did not provide detailed pricing for custom modeling; however, Clarity said custom data services cost around $25,000 for a U.S. Senate or House race. Modeling is less cost-efficient for smaller state-legislative efforts; often a state-legislative caucus will invest in such analysis for the whole state and oversample in key districts. Custom models for New Jersey's 2013 election cost the party around $35,000, according to Clarity.
Clarity has built models that decipher how likely voters are to side with a specific party or fall a certain way on issues such as gun control or abortion rights, said Mr. Bonier.
"The challenge for some of these small races is sometimes there can be high upfront costs to entering into a data project," said Robby Mook, a Democratic consultant and campaign manager for Virginia Gov. Terry McAuliffe in 2013.
Through relationships with companies like TargetSmart and Civis, which has contracted with the Democratic Senatorial Committee to provide analytics services to campaigns, the party aims to create economies of scale that benefit its most financially-strapped candidates.
"We're still working through the exact economics of it," said BlueLabs co-founder Chris Wegrzyn, "but we're looking at orders of magnitude less."
Tossing out the Obama models
Most down-ballot campaigns will need to rely on a small set of relatively generic state-specific algorithmic models provided by DNC data-consultancy partners like TargetSmart.
Indeed, the multitude of customized equations used in 2012 to categorize and target voters are considered obsolete, in part because they evaluate voter sentiment in relation to then-Republican presidential hopeful Mitt Romney, and reflect a 2-year-old political climate that is now irrelevant (consider the Obamacare effect).
"We're not using the same models … but the same techniques we used to develop them do still apply," said Mr. Wegrzyn, who served as the DNC's director of data architecture from 2011 to 2013.
Voter scoring will be used in "just about every single targeted campaign across the country," said Michael Sargeant, executive director of the Democratic Legislative Campaign Committee.
The Republicans and Democrats each launched tech-development and data-related initiatives this year looking ahead to November and the 2016 presidential race. The DNC named its overarching mission to improve its voter data and analytics and train campaign staff on related tools "Project Ivy."
The Hillary effect
Inside the DNC's Washington headquarters, which face Ivy Street in the shadow of the Washington Monument and Capitol Hill, several rooms full of young tech developers, analytics wonks and digital-media staff sit at computers. In one particularly dark room, a pantsuit-costumed Hillary Clinton doll is perched on a cubicle ledge, overseeing an operation that will be very important to her if she makes another presidential run.
The former New York senator and secretary of state is fueling another data-rich initiative for Democrats. The Ready for Hillary PAC launched in early 2013 to gear up for Ms. Clinton's anticipated 2016 campaign. The project essentially is a grassroots database-building effort that sits on NGP VAN's platform for fundraising and organizing data. The ReadyForHillary.com site created by 270 Strategies includes social-organizing features that will give the organization insight into the social connections of supporters.
"We're not just building a list -- we're building an infrastructure of people who have relationships," said Lauren Kidwell, partner at 270 Strategies. The group already has amassed data on more than 2 million supporters and expects that information to help identify and categorize voters for outreach during this year's midterms.
The commonly-accepted Democratic data advantage "is not going to be a given moving forward," said Mr. Mook. "It's incumbent on Democratic campaigns to continue to innovate. I definitely think the Republicans are taking this very seriously."