Fattened voter databases will be among the 2008 presidential race’s most enduring legacies.
If you’re a Democrat (although Republicans will have similar experiences), don’t be surprised if a canvasser knocks on your door in 2012 and, glancing at his iPhone, says, “Ms. Smith, thanks for your $50 donation four years ago–and for attending the Joneses’ party on the environment. Care to call voters in Ohio to help reëlect President Obama? Oh, and your neighbor Mrs. Jensen couldn’t get to the polls in 2008. Think you could give her a lift on Election Day?”
As you answer his questions, the canvasser will stroke his iPhone, and a campaign server will squirrel away your answers. Minutes later, you’ll get an e-mail from the campaign: “Thanks, Ms. Smith, for promising to make calls in Ohio.” If you click the link, up will pop a list of 10 Ohio voters whose answers to phone-bank callers in 2008 suggest that they wavered between McCain and Obama and were concerned about the environment. You might call them and–following a provided script–explain Obama’s environmental record, ask for their views on several issues and candidates, and record their answers with mouse clicks on a Web interface.
The next day, one of the Ohioans you’d spoken to–the one who professed strong support for Obama and a willingness to volunteer–will receive an e-mail with the names and addresses of 10 unregistered but voting-age people within a half-mile who, according to a party algorithm that analyzes demographic and consumer data, are likely Democrats. She’ll hit the streets, armed with Ohio voter registration forms she’s downloaded from a link in the e-mail. Another Ohioan–who’d told you he worked as a nurse and strongly supports Obama–will get a different e-mail. It will contain a list of licensed nurses in swing states, and a script for calling them to enlist their support.
On it will go. The night before Election Day, your phone will ring; an Obama volunteer somewhere will have logged in to the campaign’s website to make get-out-the-vote calls. “Just wanted to remind you to vote,” she will say, following a script based on the latest data about you, “and to check in with Mrs. Jensen to see if she needs that ride.”
The next morning, when you drive Mrs. Jensen to the firehouse, you will notice a college student standing near the sign-in desk, quietly tapping her BlackBerry. She is a poll watcher. She will note your arrival and–with a stroke of her finger–erase your name and Mrs. Jensen’s, in real time, from the Democrats’ Election Day call list.
This is a plausible scene from the next presidential election, which will be an unprecedentedly personal campaign. In 2007 and 2008, the vaunted Obama Web operation produced a powerful by-product: hundreds of millions of new pieces of data on voters. That information is now held principally by the Obama campaign, the Democratic National Committee, and a private database company called Catalist.
Both Democrats and Republicans have long maintained databases with whatever intelligence they can muster on individual voters. Party officials obtain lists of registered voters (about 170 million people) from the secretaries of state in the 50 states, plus the names of those who are eligible to vote but aren’t registered (another 50 million people). Then they use phone banks and shoe-leather canvassing to learn which candidates these citizens prefer and what issues they care about. Other sources provide new details: who has been a campaign donor or volunteer; who shows up in public databases of people holding, say, nursing licenses or hunting licenses. During the last election cycle, the Web tools for accessing and adding to these databases got better, and far more volunteers used them. “We went from a bolt-action rifle to a machine gun, but we also trained lots of people how to use it,” says a senior Democratic Party insider.
The Republicans are no slouches, either. Five years ago, the big story was their push toward microtargeting–identifying niche groups on the basis of hobbies, group membership, and even consumer data purchased from commercial vendors. “If somebody gets Field and Stream, they’re much more likely to be a Republican voter than a Democratic voter,” Matthew Dowd, a Bush-Cheney 2004 strategist, observed in a 2006 interview with Frontline. “If somebody gets Mother Jones, they’re much more likely to be a Democratic voter.” Democrats had fallen behind in the quest to target voters, partly because each state party kept its own data.
But that changed in 2007, when DNC chairman Howard Dean centralized data collection and management. Dean hired Voter Activation Network, a company based in Somerville, MA, to combine data from the 50 states and create an Internet interface for the resulting database, VoteBuilder. Any Democratic candidate anywhere in the nation could log in, download customized lists, and contact voters. Any detail gleaned from these voter contacts was fed back into the database for the benefit of all future Democratic candidates.
Then came Barack Obama, whose Web campaign operation turbocharged the collection of data on individual voters (see “How Obama Really Did It,” September/October 2008). In early 2007, Obama’s campaign established a social-networking site called my.barackobama.com, or MyBO; it included a custom interface for VoteBuilder. This allowed any Web volunteer visiting the Obama site to obtain lists of voters from the DNC database. Volunteers phoned these voters, asked questions about their politics, and recorded their answers through the MyBO interface, pumping more data into the DNC’s servers. (Three million such calls were made in just the final four days of the campaign.) The senior Democratic source says it’s safe to say that at least some information was recorded about the opinions of 200 million Americans. Thanks to Obama’s Web operation and those of other Democratic candidates, the DNC’s database is now 10 times the size it was in 2004, according to Voter Activation Network founder Mark Sullivan. At the same time, Obama built an e-mail list of 13 million, supplemented with information about the activities of these supporters within MyBO. Beyond the formidable Obama and DNC databases, a third database serves Democrats and progressive organizations. Built by Catalist, a company headed by Bill Clinton’s former deputy chief of staff Harold Ickes, it started with the national voter list and added supporter data from progressive groups like the Sierra Club.
With these three databases, notes Vijay Ravindran, who until recently was Catalist’s chief technology officer, “the Democratic side has a lot of information that it could bring to bear. It has enriched the slate of volunteers who can be ‘turned on’ for future grassroots activities. It has information on people’s positions that can be used to lobby congressmen.” Thomas Gensemer, managing partner of Blue State Digital, which built the Obama Web tools, says the databases will probably be used to mobilize Obama voters in support of the president’s agenda. “Think back to 1992, when President Clinton put through health-care reform [and it failed],” he says. “If you’d had a list of demonstrated activists, online or off, who could help fight back, that is a powerful use for an informed surrogate.”
While it remains to be seen whether President Obama will try to deploy his supporters in that way, it’s clear that the databases are useful instruments. The Obama campaign tried many interesting strategies in 2008. Its “neighbor to neighbor” tool, for example, let volunteers plug in their addresses and pull up the names of 25 neighbors to call; veterans were recruited to call other veterans in swing states. But such strategies are now possible in unprecedented variety, and at an unprecedented scale. “You are getting potentially all the way to a one-to-one message,” Ravindran says. “In terms of how things move in the future, that is the Holy Grail.”
What’s a bonanza for campaigning also has its potentially intrusive side, of course. “The availability of huge amounts of personal information is new, and we don’t agree as a culture on what’s useful and [what] we’re in favor of, and what’s plain creepy,” says David Weinberger, an Internet advisor to the 2004 Howard Dean campaign, who’s now a fellow at the Berkman Center for Internet and Society at Harvard. “It’s exactly the same question that marketers are facing, but this is in the sphere of citizenship, and it’s important to get this right. There isn’t an obvious path forward yet.”
Indeed, political campaigns know much more about you than they used to. They always knew whether you voted, because this is public information. Now they know the details of your campaign activities and responses to campaign messages, right down to which e-mail messages led you to click open a link. They can trace any social networks you constructed inside a campaign site. When they buy consumer databases, they know even more about you.
Somehow, under President Obama, all this seems fairly benign–at least to his legions of well-wishers. To them, the idea that his DNC or his campaign might use knowledge of someone’s environmentalism to help fight for tougher clean-air laws will appear laudable. The idea that distributed volunteers might dilute some of the power of big donors and pressure groups in the next campaign will seem to many like a step in the right direction. But in another context, such activities might take on another color altogether–just as another kind of huge gathering on the Washington Mall might seem more ominous than uplifting. Information has always been power. National leaders of all kinds have always wanted both.
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today