In an open-plan office in Durham, N.C., that feels part ad agency and part rec room, a 31-year-old EvoApp executive named Pritam Das sits rummaging through all the Twitter posts that have mentioned Mitt Romney in the previous 24 hours, sifting and sorting by various keywords and metrics.
The day before, a senior Romney advisor had likened the campaign's transition from the primary fight to the general election to shaking an Etch A Sketch. With the stroke of a key, Das reveals that of the tweets mentioning Romney, the word "Etch" ranked third among tags, behind only the candidate's first and last name. More tellingly, perhaps, Das was able to gauge that many of those judging the impact on Romney's campaign most harshly had the greatest influence in the Twittersphere--not just the most followers, but followers who in turn have influence of their own.
"These guys have come up with a way to understand and interpret unstructured data in real time," says EvoApp CEO Kip Frey of founders Joe Davy and Alexey Melnichenko, 23-year-olds who met at a specialty math-and-science high school and abandoned college careers to start the company. "Politics is only one of its uses, but it's an important one. It shows that we can solve difficult problems." EvoApp, which has landed $2.9 million in angel investment, expects to gross nearly $1 million in 2012, its first year in business.
It's part of a new wave of polling and data firms that are changing the way political campaigns gather and interpret information. Many launched after the 2008 presidential election, during which the Obama team made revolutionary use of the internet to solicit donations and communicate with supporters.
That was just four years ago--a short time in traditional political strategizing, but a generation to today's pollsters.
"The stuff we used to do even two and three years ago, we don't do anymore," says Mike Henry, a veteran Democratic political operative who is running Tim Kaine's Virginia Senate campaign. "It's dated. It's not accurate. Politics drives people to adapt, or they lose. That helps move things along."
Adapt or lose. That's what Newt Gingrich realized after he started his campaign for president in mid-2011. His staff was reacting to issues on a 24-hour news cycle, which was the way politics worked the last time he'd run for office, in the 1990s. The strategy proved disastrous in a world in which even waiting for a debate to finish before responding to it means you've already waited too long.
Witness the $10,000 bet Romney proposed to Texas Gov. Rick Perry on national television, a moment that helped pull even old-school pols into the Twitter era. "All of a sudden, within minutes, there was a '#10Kbet' Twitter hashtag. The debate hadn't even ended yet," says Rebecca Katz, a partner at Washington, D.C.-based consultants Hilltop Public Solutions.
"If you wanted to know how a certain issue was playing, it typically took you three to four days, and even that was extremely fast-track," says Zac McCrary, vice president of Anzalone Liszt Research, a technology-savvy Democratic polling firm. "Now you can do it in a matter of hours. That can change the trajectory of a campaign."
The pressure to win is far more intense in the 50-percent-plus-one-vote world of politics than in ordinary commerce. In nearly every consumer category, a customer base can be built up incrementally over years, and a product with a small market share can still be extremely profitable. Not so in politics. If you win the election or the referendum, it doesn't matter by how much. If you lose, it doesn't matter how fervently your supporters loved you.
In politics, even the smallest advantage can mean the difference. So anyone who can identify voter sentiment or political trends faster or more accurately than the rest of the world will be in great demand. "The information game is rapidly changing," says Jeffrey Plaut, a founding partner of New York-based Global Strategy Group, a strategic communications and research firm that works with companies and candidates. "And politics is acutely tied to how people get and process information."
Eli Kaplan is founder of Rising Tide Interactive, a D.C. firm that uses digital media for fundraising, information-gathering and showing ads. "We've reached a tipping point where the people who run campaigns understand that they really need to take this stuff seriously," he says. "There's still room for traditional polling and advertising. But it has become clear that if all you're doing are things the way you always did, you're doomed."
Some of the new players in the category are political insiders. Others are not, like EvoApp's Davy and Melnichenko, who are far more comfortable with programming than with public policy. "We're 23 years old," Davy says, "and our intelligence manages some of the most secret and important projects in the world. When you think about that, it's crazy."
EvoApp pulls in data from throughout the internet--Twitter, Facebook, blogs, forums, news stories--and gives it structure. "Politics was one of the most obvious uses because it's all about information," Davy says. EvoApp's tools analyze conversations and can provide insight into sentiment around a candidate in real time--sometimes a week earlier than Gallup poll results.
It's this promise of incremental insight through innovative methodologies that is leading the charge. "I see new vendors come in all the time who've never even been involved in politics but think they have the next great idea," says Katie Harbath, a former Republican National Committee (RNC) staffer who is now the manager of public policy for Facebook. "The thing is, some of them actually do."
Opportunities abound. "There's a pot of money to be had in politics for people who understand this emerging world," says Patrick Ruffini, the former eCampaign director of the RNC. "These new firms are looking to bank on the shift from the traditional to the digital."
In 2007 Ruffini founded Engage, a Republican consulting company based on the idea that the next generation of political data-collection and advertising would be wholly different from the last one. "Before us, there were a lot of firms that built websites and billed themselves as digital strategists," he says.
"We distinguish ourselves by showing how to use digital to actually win elections."
Even established names are getting into the act. In the 1980s, Jill Hanauer appeared in Time magazine and on the CBS morning news talking up the student movement. Later, she worked for Gary Hart, Patrick Leahy, Michael Dukakis and Barbara Boxer, among others. In 1994, she founded the Interfaith Alliance to protect America's church-and-state separation.
After helping Colorado's Democrats take back the legislature in 2004 even as John Kerry was getting trounced in the state's presidential race, she realized that Eastern political operatives didn't understand the West. "We were winning races on the local level that Democrats never would have won before," she says. "We wanted to figure out why."
That led to Denver-based Project New West, a research and strategy firm founded with former Clinton political director Doug Sosnik. Instead of working for candidates like a traditional pollster, Project New West hired pollsters itself in an effort to gather detailed demographic and attitudinal data. Then it analyzed and integrated the data using stringent metrics of analysis and sold the resulting information to state party organizations and other entities through annual subscriptions of up to $140,000 each. "It's a unique business model that allows millions of dollars of research to be shared with our subscribers," Hanauer says.
The model worked so well that the company has relaunched as Project New America (PNA) and started operating nationwide. "We don't describe ourselves as a polling firm," Hanauer says. "Polling only gives you one snapshot of a deeper, richer picture that you need. We pay various pollsters to do work on an issue or in a region--not one data point but many, and over time."
From a staff of two, PNA has ramped up to 16 full-timers and expects to gross $5 million this year. Next comes a private offering that will raise $1 million in investment capital. "Our secret is, we're not competing with pollsters," Hanauer says. "We're adding value to the work they do. Because of our information, campaigns can micro-target in a way that hasn't been done before."
Some 15 blocks away from PNA headquarters, amid the sports bars and trendy restaurants of downtown Denver, campaign data is being collected using even less traditional means. Scott McNealy, the co-founder of Sun Microsystems and a political enthusiast with strong Republican ties, created Wayin, an internet polling and survey tool that works with sports, consumer brands--and politics.
"This is a fun business, immediately engaging and psychotically addictive," he says. "And people are responding. We've just had our first quarter of revenue, and the results are pretty exciting." Both the Romney and Gingrich campaigns signed on early; the RNC is also a user.
Wayin polling is not anonymous (an e-mail address must be provided), so clients are able to aggregate data about users--not only making every answer more meaningful, but also letting sponsors target their follow-ups. It's the secret weapon to a tool that has enticed interest and investment from some heavy hitters. "I've been able to go to all kinds of buddies, and they're all invested in it," McNealy says. "We have a list of investors that's about as sterling as you can have. And my board of directors is a Fortune 200-caliber board."
Frank Luntz, a Republican strategist and commentator, periodically references an ongoing Wayin poll on his Fox News appearances. Then he sits back and watches the responses add up. "Thanks to Wayin, you ask a question on Monday morning and you get an answer on Monday morning," says Luntz, who has invested in the tool. "It allows you to break down an issue to the finest, cleanest demographics and geographics of any technology I've seen. Is it the wave of the future? I don't know, and I don't really care. It's tremendously exciting to me right now."
For many campaigns, the idea of hiring a firm solely to handle digital media or internet polling remains an expensive novelty. But at some point soon, many industry professionals believe, it will become a necessity, just as hiring a pollster in addition to a campaign manager became de rigueur for national campaigns during the 1960s. "Social media gives us an additional data point," says campaign manager Henry. "Smart campaigns acknowledge that, and use it to their advantage."
Despite this, says John Anzalone, founder of Anzalone Liszt Research, "I still find a lot of political people have their heads in the sand." His Montgomery, Ala.-based polling firm is considered traditional in that it actively calls people and asks them questions.
But over the years, his methodology has evolved radically. It's not so much about keeping pace with the competition, he insists, as serving clients by giving them the most accurate information possible. "The Anzalone model is the future," Henry says. "Taking both disciplines, the old way and the new way, and bringing them together."
The son of a long-haul trucker, Anzalone managed early in his career to intersect with most of the top Democratic operatives of the generation, including James Carville, Paul Begala and David Wilhelm, who managed Clinton's presidential run. So when he decided to set up shop as a pollster, he had the contacts. Five years later he added Jeff Liszt, former technology director for the Alabama Democratic Party, as a research director and eventually made him a partner. These days, Anzalone is one of the most respected information-gatherers in the business. In 2008, he worked seven states for the Obama campaign, including Florida, North Carolina and Virginia; this year, he added Arizona and Nevada.
Anzalone will bill at least $5 million in 2012 from 60 clients. "It's great that we poll for Obama, but we're also running a poll for the mayor of Elizabeth, N.J. It's repeat business like that that pays our bills," he says.
Recently, Anzalone sat in his office on a conference call about a poll he's conducting for sponsors of a California ballot initiative. The office looked almost the same as it had when he'd handled Obama's polling four years before. But in 2008, almost no polling firms in America were calling cell phones. It's far more expensive than calling land lines, since federal law prohibits calls to mobile phones using recorded messages or automatic dialing software. Yet because he's convinced it yields more accurate readings, Anzalone now targets cell phones for up to one-fifth of his responses.
He also uses innovative messaging polls on the internet that yield exponentially more data than anything he can do on the phone. "It's 40,000 data points instead of 1,000," he reports. "You're testing 16 different messages in boxes of four. And you can amalgamate the information in a way that you never could."
That's just the start of the next great information revolution, he believes. "People coming out of this campaign will start businesses around analyzing data, bringing all the strands together in ways that haven't been done before in politics," he says.
Some of that methodology will be cribbed from the corporate sector, which is several iterations ahead of political campaigns when it comes to testing advertising and branding messages online. "Most corporate strategic research is now done digitally using rapidly developing tools," says Global Strategy Group's Plaut. "There's the ability to get immediate feedback, to change things based on consumers' attitudes." Companies such as his that can engineer a successful crossover have a unique product to pitch to campaign managers. "We're seeing increasingly that the political and corporate work feed on and inform each other," he adds.
Campaigns aren't the only ones doing the testing. On Fox Radio one April morning, conservative strategist Luntz and commentator Sean Hannity fell into a discussion about which issues mean the most to GOP voters. They decided to create an instant poll on Wayin.com and asked listeners to ... well, weigh in.
Sarah Uhran, Wayin's director of unusual projects, sat in a small conference room back in Denver tallying results on her cell phone. The poll had no connection with any campaign or political party. It was for entertainment purposes only; a media personality and a consultant asking a question, and their listeners responding. But since many of those who voted were frequent visitors to the site and had logged in so their views could be recorded, Uhran was able to "slice and dice" the votes in a variety of ways, aggregating their opinions with views they'd expressed in the past.
"That's the part that Stephen Colbert was so interested in when we sat down with him," Uhran says. "That if you have this data, eventually you can say that people who like Obama, for example, also tend toward these other attributes that you can list."
The Luntz/Hannity poll generated thousands of responses. "Whenever Frank [Luntz] does a poll with us, we always get at least 2,000 votes, and sometimes as many as 30,000," Uhran notes--numbers that far exceed the 1,000 or so that most traditional pollsters use to create a snapshot of current opinion.
These were Hannity listeners, so they hardly represented a cross section of the American public. They hadn't been screened for gender, ethnicity, geography or anything else. Still, it was hard not to be intrigued that, say, only 3 percent of these hard-core Republicans blame Bush for the economic crunch.
"Wait, that number is now down to 1 percent," Uhran points out, peering at her phone. "It may not be scientific," she acknowledges. "But it is interesting."
More than that, it's a business.