Lessons from the Cave: Future Campaigns at All Levels are Sure to Tap Into Obama-Style Analytics

Feb 1, 2013

While he waited for the concession call from Mitt Romney, the president worked on his acceptance speech with Jon Favreau, director of speechwriting, and campaign adviser David Axelrod at a Chicago hotel.
Credit WUIS/Illinois Issues
Collin Corbett is a Republican whose job it is to elect Republicans. His company in Chicago’s northern suburbs, Cor Strategies, has worked with presidential contender Mitt Romney, gubernatorial hopeful Sen. Bill Brady and dozens of lesser-known politicians vying for offices such as city council or circuit clerk. 

After President Barack Obama won re-election in November, Corbett says his Republican clients immediately wanted to copy the Democrat’s campaign tactics, especially when it came to technology. 

“They think they’re going to win if they’re up on social media, if they’ve got posts all the time on Twitter and Facebook, if they’re on Instagram and Pinterest,” Corbett says. The candidates, he adds, want apps for smartphones and tablets, too. 

The tech-savvy Corbett says candidates are missing important lessons from the Obama campaign if they focus solely on the flashy technology and social media sites.

“In 2012, Romney probably had more tech than Obama did, although … half of it didn’t work on Election Day,” Corbett says. “The key was that the Obama campaign was the best we’ve ever seen at tracking and managing data.”

Many campaign professionals dubbed 2012 the “Year of Big Data,” and the Obama campaign is its undisputed poster boy. But Obama tactics and technology that sound extraordinary today will be old news by the next election cycle. Even campaigns at the state and local levels will adapt quickly. The biggest question is which of those tactics are worth adopting.

Pulling apart the factors that led to Obama’s victory is tricky business, and it is easy to overstate the role of the campaign’s fancy new data tools. Even Obama’s top technical adviser warns that people are too quick to credit tech and data innovations for the win. Few other candidates will have anywhere near the time or the resources to assemble such a sophisticated operation. 

But the fact remains that technology did give the Obama campaign an advantage, and everyone running for office or running a campaign is looking for an edge. So naturally, campaigns everywhere will try to figure out what parts of Obama’s digital strategy they can use for themselves.

One unifying theme about the Obama camp’s use of technology is that every digital tool helped the campaign achieve one of its basic goals. “We were trying to recruit volunteers, register voters, raise money, persuade people and turn them out,” Teddy Goff, Obama for America’s digital director, said a few weeks after Election Day. “That was it.”

For end users, the tools made life simple. Undecided voters saw ads that specifically talked about issues they cared about. Volunteers got straightforward calling assignments and scripts to go along with them. Potential donors got email messages that made them want to give money.

But on the back end, things were much more complex. The campaign had 54 staffers in its Chicago headquarters crunching data and building digital tools. They assembled one of the most sophisticated voter lists ever built. They modeled election results and probed public polls for weaknesses. They decided which TV channels to advertise on to reach swing voters who were not watching the nightly news. They tested the effectiveness of different fundraising pitches. In short, there was very little in the campaign that the data gurus did not have a hand in.

After Obama’s victory, his campaign’s technical prowess fascinated observers — including Romney staffers — as details of the operation emerged. In the weeks after the election, journalists wrote about their visits to “The Cave,” the nerve center of Obama’s data operations at One Prudential Plaza in Chicago. Obama’s fellow Democrats clamored for him to share his database of some 16 million voters, donors and volunteers. But it is important to remember where technology fit in the larger campaign.

Even for presidential campaigns, Obama’s re-election operation was massive. The campaign had 3,000 paid staffers, compared with 500 for Romney. The president also enlisted 30,000 volunteer “neighborhood team leaders,” who were basically precinct captains working full time coordinating other volunteers. The Boston Globe reported that in key precincts, the average Obama worker was responsible for 50 voters. Romney’s workers, by comparison, were each responsible for thousands. 

The Obama team bet that a strong ground game with lots of person-to-person contacts would lead them to victory. David Simas, Obama’s top pollster, said at a post-election debrief that the campaign wanted to make the national contest feel local. “We ran ward races throughout the country based upon trust.”

The starting point was Obama’s vaunted voter database, which was built and refined over the last six years. For every voter, it included 80 potential categories of information. Beyond standard contact information, the campaign tracked voting history, mail-in ballot requests and even magazine subscriptions. It assigned every voter a score to predict how likely the voter would be to back the president if an Obama volunteer contacted him or her. 

The campaign also measured how voters reacted to campaign developments, ads and outreach efforts by polling 9,000 voters every night and tracking the responses.

As basic as it seems, one of the biggest advantages the Obama campaign gained with its reams of data was a better picture of who was most likely to vote. Obama aides claim the campaign did a better job of identifying likely voters than even the most rigorous media polls, which is one reason they appeared so confident, even at times when Obama’s national poll numbers dipped.

Much of the campaign’s technology, though, was simply used to connect Obama supporters with uncommitted voters. Volunteers in noncompetitive states, such as Illinois, could get call lists of voters in swing states. A military veteran, for example, might get a list of fellow veterans and a script highlighting the president’s positions on veterans affairs. 

The Obama campaign used Facebook to reach undecided voters in ways that officials say was not even possible in 2008, when Obama’s campaign was roundly hailed for its use of social media. Between the campaigns, the number of Facebook users grew tenfold. So by the re-election campaign, Obama’s 33 million Facebook followers were friends with 98 percent of U.S. voters. The campaign figured the vast majority of its Facebook followers had already decided to vote for the president and did not need to be persuaded. Instead, the Chicago team focused its efforts on getting its followers to spread the campaign’s message to undecided voters by sharing photos or sending messages to specified friends in swing states.

Obama for America also bet big on online advertising. The campaign doubled the share of its budget spent on online ads from the president’s 2008 bid. This election cycle, officials say, the Obama team spent 15 percent to 20 percent of its advertising budget on Internet ads. Some of that money went to buy mobile video ads on news sites and in smartphone and tablet apps, including the Weather Channel and Pandora. 

Some parts of the president’s 2012 campaign were revolutionary, others merely evolutionary. The Obama camp had the biggest and best game in town this cycle, but it used many tools that much smaller campaigns can and do use, too. 

Take the voter file, for instance. Almost every campaign starts with a list of voters in its area, which can be purchased from vendors such as Aristotle Inc. The company culls most of its information from public records, such as voter registration rolls and driving history. For a higher price, Aristotle also lets campaigns buy email addresses, magazine subscription information and histories of political donations. In fact, says chief marketing officer Brandi Travis, Aristotle offers a total of 500 categories of information on voters.

“That’s really what the Obama campaign did, too,” Travis says. “Their way of using it was a little more innovative than what we’ve seen in the past, but the actual data being used isn’t really any different than it was in the last campaign or the one before that.”

Vendors can make data easier to get and easier to use. Aristotle, for example, sells an app for smartphones and tablets that generates walk lists for volunteers canvassing a neighborhood. The workers can use a credit card reader to take donations right at the front door. The data on their mobile devices can tell them who is most likely to donate and how much.

Just because local campaigns are capable of using the same tools as the Obama campaign did does not mean they are. Corbett, the Republican consultant from the north suburbs, says many campaigns do not even use basic filters to deliver the right messages to different voters. They send all of their mailers to all of their voters. It would be relatively simple, say, to send a postcard about women’s issues only to women or a mailer about schools only to parents. “That’s not happening,” he says. “A good start would be that.”

That type of targeting is relatively simple, and it is a long way from the “hypertargeting” that is now possible with online advertising and robust voter data. That method, Corbett says, can identify supporters who would otherwise be hard to find: a conservative woman in Highland Park, for example, or a liberal living in rural southern Illinois. Finding those pockets of support is especially valuable in larger races, such as those for Congress or statewide offices.

But the proliferation of technology could be a challenge for smaller campaigns, Corbett adds. Local candidates may not know much about the new technologies, and they are less likely than better-funded candidates in big races to have the money to hire a tech expert. But voters are getting used to the more sophisticated candidates that target voters online or during their free time. They do not want to be interrupted at dinner. “If [candidates] make phone calls and knock on doors like it used to be, people are less and less inclined to pay attention because they’re expecting you to come to them in the ways they prefer,” Corbett says. 

Larger operations also face tough choices about the amount of time and money they put toward digital tools. “Something we had to struggle with that is newer for statewide campaigns is how technical to make your campaign,” says Steve Schoeffel, who was a deputy campaign manager for Brady’s gubernatorial effort. “We had a lot of tools that we knew were available to us. The hard decisions for us to make were: [How much] do you direct resources as far as staff to dedicate to technology?” 

Staffers could be used for a long list of tech-related tasks, from keeping up a campaign’s social media presence to analyzing voter files. Some of those could have a big payoff for the campaigns, Schoeffel notes. An effective online ad campaign, for example, could reach voters for far less money than a mass mailing or radio ad. But managers must weigh whether hiring a tech staff is the most effective use of resources.

Terry Cosgrove, the head of the abortion rights group Personal PAC, is also wrestling with how much to invest in the high-tech tactics used by the Obama campaign. Personal PAC is known for its aggressiveness, and it is on a winning streak. Cosgrove has said his group was instrumental in re-electing Gov. Pat Quinn by sending out a barrage of direct mail pieces, making robocalls and producing the group’s first TV ad, all attacking Brady. The efforts were specifically directed at winning over suburban women.

In other words, Personal PAC sounds like the type of group that would be interested in more precisely identifying voters. But Cosgrove says a broad-based strategy has served the group well. He says single-issue voters who oppose abortion rights are only a small fraction of the public. The rest, he claims, are open to the group’s messages. “Strategists make a big mistake if they worry about anti-choice voters,” he argues.

Before Cosgrove jumps in to more tailored outreach efforts, he wants to know which of the myriad tactics Obama used led to measurable results. The Obama campaign, after all, employed a lot of different tools and tactics in its re-election effort. Just because Obama won does not mean that all of the methods he used worked. And Cosgrove is also leery of vendors hawking new products that may not be effective at targeting the relatively small number of people who vote in legislative races.

Still, Cosgrove plans to meet with members of Obama’s digital team to keep up with the new developments. “We’re open to learning and being engaged in new strategies,” he says.

Corbett, the Republican strategist, says he is excited that the Obama victory woke up the political world. Politicians, he says, are a decade or more behind the private sector in how they are using technology in marketing. Democrats have an edge in it now, but Corbett is hopeful that Illinois Republicans can regain it. “If we wanted to catch up … one of the ways we could do it is to start learning more about the voters we are hoping to have vote for us,” he says. “We’ve made a lot of assumptions about these voters, and they’ve turned out to be inaccurate.”

It would not take long for the state Republican Party or an individual candidate to gather the data, Corbett says. Whichever gubernatorial candidate his firm works with in the 2014 race will use that data in a sophisticated way, he promises. It is especially important, he says, in light of the fact that the last Republican candidate for governor, Brady, lost by less than 1 percent of votes cast. 

“This is the path to success,” Corbett argues. “We had a candidate in 2010 who should have won, but we weren’t able to pull across the finish line. What we’re talking about here today is the stuff that will actually allow us to win statewide as Republicans.” 

Daniel C. Vock is a reporter for Washington, D.C.-based Stateline.org.

Illinois Issues, February 2013