The Victory Lab (26 page)

As she explained these machinations to DeShong, Quinn took out a paper napkin and started diagramming. She drew a line to represent the partisan continuum, from left to right, and then laid upon it a series of bars reflecting targeted messages, annotating as she scribbled furiously. One bar symbolized “terrorism and late-term abortion for older Hispanics,” Quinn
said, while another of different length was “terrorism and guns for male union members.” DeShong sat dumbfounded by the squiggly lines, until Quinn explained her point: Democrats were polling to make messages for broad audiences, while Republicans were modeling to match messages to specific audiences. They were able to do that because they had not only better data than the Democrats to find the voters they wanted, she went on, but also the statistical tools to profile those they couldn’t reach and nonetheless predict what views they were likely to have.

“Now we have to do it,” Quinn said. “And we have to do it better.”

LAURA QUINN WAS
something of an accidental technologist: a communications aide always looking for the newest way to get a message out. While working on Capitol Hill in the mid-1990s, she helped build a new television studio for Democratic senators and hired the first full-time Web expert to help members of Congress develop an online presence. In 1999, she moved to the White House to work for Al Gore as he prepared for a presidential campaign. After his loss, newly elected DNC chairman Terry McAuliffe hired Quinn to prepare a report on the party’s tech infrastructure.

Quinn was a tall, lithe woman with long gray hair and austere Modigliani features who climbed rock faces for fun, and she prowled the party’s salmon-colored, concrete-block headquarters with the same expeditionary spirit. Encountering fleas, mice, and large concentrations of radon along the way, she found an underused television studio and a press office with no database of media contacts. There were telephone lines that had been installed to support long-forgotten election-year call centers and became themselves forgotten; in one case, Quinn discovered that the committee was somehow paying the monthly phone bill of a Florida dentist. Her report to McAuliffe recommended a thorough renovation of the building with an eye to both making it inhabitable and modernizing its technology. “It was really a terrible mess,” Quinn says.

The physical decay of the Democratic National Committee was the inevitable result of a cycle of strategic disinvestment. Every four years, the Democratic presidential nominee would inherit command of the structure, installing cronies in top jobs and filling the lower ranks with staff shifted off the campaign payroll. “The Democratic Party strategy had very much been, going into presidential elections, that you just spend everything down to the dirt,” says Quinn. “Then the DNC is in debt and everything is broken.” The DNC’s election-year job was largely to be a fund-raising vehicle for the big-dollar contributions known as “soft money”: unrestricted gifts, often from union and corporate sources, that the parties could legally raise but candidates could not.

McAuliffe had risen in national politics as a fund-raiser for Bill Clinton, and together they had helped to expand the Democrats’ reach into Hollywood and Wall Street wallets, narrowing the soft-money gap between the parties. But the DNC still lagged significantly in raising so-called hard money, small contributions from everyday people. It was a predicament that Democrats couldn’t write off as an inevitable consequence of the opposition’s close ties to big business. “How is it that the Republican Party was consistently and substantially outraising the Democratic Party in small-dollar donations?” asked Quinn. “What was allowing them to do it?”

Quinn looked enviously at the RNC headquarters, two blocks away and across a set of railroad tracks. It had always been a more stable, centralized entity than the centrifugal DNC. National Republicans effectively ran their state parties from Washington, even paying the salaries of executive directors. Democratic state chairs, on the other hand, were frequently appointed by the governor or an influential legislator and felt little fealty to national leaders. As she closely examined parties’ spending over the previous two decades, Quinn marveled at how the opposition had repeatedly made farsighted investments in new communication technologies. For a generation, conservatives distrustful of the mainstream media had made a priority of finding new channels to directly address their base. In the early 1990s, the RNC had built TV and radio studios so that party figures
could call in to sympathetic talk shows. A few years later, Republicans put money in a website and even designated supporters as Internet captains at a time when Democrats had yet to begin taking the online world seriously.

But when he became chairman in early 2001, McAuliffe felt an urgency that had moved few of his predecessors. He was already looking ahead to the ways in which life would change under the campaign finance reform bill known as McCain-Feingold, which vowed to ban soft money and would be enacted the following year. McAuliffe foresaw that the national parties would be forced to reinvent themselves as engines of small-dollar contributions. Ironically, the soft-money parity that McAuliffe had helped to achieve during Clinton’s presidency now made the party disproportionately dependent on a mode of giving that would soon become illegal. The DNC had to learn to talk to people.

In April, Quinn returned to McAuliffe with her assessment of the infrastructure, which was filled with reasons for alarm. McAuliffe fixated on one set of statistics: the DNC’s list of names—donors and volunteers, mostly—totaled two million, and for many of those Quinn feared the information on file was no longer current. The e-mail list included only seventy thousand people. The party had little capacity to hunt for other loyal Democrats who could be converted into contributors. McAuliffe decided that his goal, as he started to tell people, would be to have “more live e-mails on the file than dead voters.”

Quinn’s report to McAuliffe had not recommended investing in a unified national voter file. The RNC had begun its own project to do so in the late 1970s, hiring private vendors to collect information on voters in all fifty states, standardize its format, and develop a system that could be continuously updated with new voters, changed addresses, and expanded vote histories. The Constitution leaves responsibility for administering elections to the states, and RNC officials quickly learned what a rough-hewn patchwork of laws and protocols had emerged to simply manage voter registration practices. “I can’t overestimate how bad a shape the voter registration files were in in this country when we started,” says Eddie
Mahe, the committee’s executive director. In New Hampshire, each of the state’s 234 townships was responsible for maintaining its own registration records, many times on handwritten rolls stored in officials’ homes. In Montana, Republicans had to send someone to gather computer tapes from every courthouse in the state. One Maine town was reputed to keep its voter list on chunks of tree bark.

It took more than a decade for the RNC to complete its national file, and during that time voter data became a high-stakes business for private-sector entrepreneurs.
As a junior physics major at Princeton in 1976, John Aristotle Phillips had written a thirty-four-page paper that described how to assemble a plutonium fission bomb, like that which had been unleashed on Japan in World War II, for two thousand dollars. Phillips was approached by a Pakistani operative who tried to buy his paper, which prompted Phillips to both contact the FBI and call a press conference. The bribe had proven Phillips’s point, as he later wrote, “that a terrorist, with a background in college physics, a small amount of stolen plutonium and the wherewithal to construct the device, could pose a threat to world peace.” After graduating, Phillips ran for Congress in 1980, a Democrat looking to represent Connecticut on an antinuclear platform, but as an outsider he couldn’t win access to voter files controlled by local party officials. After two unsuccessful campaigns, Phillips stopped running for office and enlisted his younger brother, Dean, a computer engineer, to develop tools for those who did.
In 1983, their Aristotle Industries released Campaign Manager, one of the first pieces of consumer software for candidates, and started assembling voter lists to feed into the program. By 2000, the company had sixty-nine employees and was expanding worldwide. “If the parties were doing a great job then we wouldn’t have any data business,” says Phillips.

Quinn knew this history and decided that a national database would be too big an undertaking to successfully complete before the 2004 election season arrived. Along with the technological and logistical hurdles, above all there was a political one: a network of state parties ready to thwart any of Washington’s efforts to seize one of their most prized commodities.
“There’s always been a trust issue between the DNC and state parties,” says Ray Buckley, then vice chair of the New Hampshire Democratic Party and a longtime DNC official. “Nobody believed that someone in Washington would not screw them.”

In some cases, party officers used the list as a fund-raising device—presidential candidates in Iowa had to pay forty thousand dollars for its list of past caucus participants. In other states, kingmaker state chairs considered the list a plum to be shared with party-endorsed candidates. The Democratic Party of Illinois didn’t even own a voter file but relied on one built and personally controlled by the state house speaker, Michael J. Madigan. “The idea that they were never going to give their data to the national party was so ingrained,” McAuliffe says. “That was their asset, and they weren’t going to give it to me.”

But he was ready for the fight. The only path McAuliffe saw to hard-money parity ran through cycles of prospecting new donors, in the mail and online. To accomplish that on the scale he believed crucial, Democrats needed the list of 100 million new names, sortable by party registration or voting behavior, that would fill a national voter file. McAuliffe proposed a deal to the state chairs, that the DNC would effectively borrow their files, help clean them up, add new data like donor information and commercially available phone numbers, and then return them for the state party’s use. At the same time, McAuliffe went to Vinod Gupta, a major Democratic fund-raiser and Clinton friend who was the founder and CEO of InfoUSA, one of the country’s large commercial data vendors. Like many of his rivals, Gupta had been trying for years to find customers in the political world, and offered McAuliffe a good deal for his product. McAuliffe agreed, and as the state files came in, the DNC would send them out to InfoUSA’s Omaha servers, where hundreds of pieces of new information were added to each voter’s profile. A new interface was built to navigate it all. It was called Demzilla.

The so-called datamart that McAuliffe helped to assemble was riddled with errors, and many state-level campaigns did not even know how to
get their software to properly handle the information that was returned to them. But a roster of more than 150 million new targets was a great gift for the DNC’s fund-raising operation, just as liberals were getting energized to rally against Bush. At one point, McAuliffe descended to the basement of the headquarters to invite the entire technology department out to dinner at the Palm as a sign of his gratitude. Merely by having built a website that could process new contributions online, they had begun to compete with the finance department as a fund-raising source.

McAuliffe’s term was scheduled to end in early 2005, and a wide-open race for the DNC’s chairmanship loomed. Quinn knew that whichever candidate succeeded McAuliffe would set new priorities, potentially jeopardizing the datamart’s future. Howard Dean in particular worried her. His primary campaign for the presidency in 2004 had been framed as an insurgency against the party’s Clinton era leadership, whom Dean pilloried as too centrist and money-obsessed, a caricature that fit McAuliffe perfectly. While Dean’s campaign had often proven incompetent at the basics of voter contact, it had pioneered the use of the Web for small-dollar fund-raising and communicating with the party’s activist class. Quinn suspected that if Dean won he would bring in his old campaign technology team, several members of whom had started their own firm, Blue State Digital, after his loss to Kerry.

Quinn began to think that the innovation necessary for Democrats to regain their footing would have to take place outside the DNC headquarters. The politics of the moment compelled it, but Quinn saw another benefit. Although party committees could share their lists with state affiliates and candidates, campaign finance laws restricted them from coordinating with outside groups. So while party officials had counted on AFL, ACT, and EMILY’s List to assume a growing share of Democratic voter contact work, they couldn’t speak to their allies about strategy or divvy up a list of voters to avoid duplicating efforts. It was time, Quinn thought, for the left to built an institution that had no purpose but to warehouse and share voter information. “For a long time, data was something you made
valuable by keeping it in a drawer away from everyone else,” says Quinn. “Only through the last decades have people realized that data becomes much more powerful when it is combined with other data, and it is free to grow and multiply.”