Skip to Content
Uncategorized

Facebook: The Real Presidential Swing State

The outcome of the 2012 campaign could have less to do with grand vision than with online data analytics and peer-to-peer voter targeting.
July 20, 2012

Facebook and Internet campaign strategies grew up at the same time. In 2003 and early 2004, when Facebook was a new dorm-room plaything, Howard Dean’s presidential campaign pioneered Internet fund-­raising. By 2008, Facebook had crossed the 100-million-user mark and was coming to dominate online social networking; that year, Barack Obama’s campaign wielded a custom social-networking site that helped win the White House (see “How Obama Really Did It”). A Facebook cofounder, Chris Hughes, helped build that site.

Now, in 2012, Facebook is central to the upcoming presidential election. Both Obama and his Republican opponent, Mitt Romney, are well aware that half or more of the electorate is on Facebook. Both campaigns’ websites are entwined with Facebook pages; visitors are encouraged to log in with their Facebook accounts and then post messages supporting the candidates for their friends to see. What Facebook also gives the candidates is an arena for testing, analyzing, and distributing precisely targeted political advertising. Both campaigns can also use Facebook to urge their supporters to vote and, potentially, to lobby their undecided friends in swing states. That means this is where the 2012 election might be won or lost—even if far more money will be spent elsewhere, especially on TV ads.

Things reviewed:

President Obama’s Facebook campaign app

Mitt Romney’s website and Facebook page

Social organizing tool
a Democratic Facebook app from NGP VAN

Making use of social connections can lead to the ideal form of marketing: individual messages of persuasion delivered by trusted friends. You can see the president’s campaign reaching for this goal with Obama 2012, an app that his supporters can use to integrate their Facebook accounts with the campaign’s website. The app’s avowed task is to give people a quick and easy way to access the volunteering and organizing functions that worked so well for Obama in 2008. But the permission screen that comes with the app makes clear that it has another purpose as well. When I installed the app, I noticed that it said it would grab information about my friends: their birthdates, locations, and “likes.”

Facebook’s policies require that such data be used in only the context of the app itself, but even so, the campaign should be able to create tools that prompt supporters to approach voting-age friends in swing states and craft personalized appeals based on what the campaign can infer about those friends’ interests and views. Similar tools are coming from other quarters, too. In July NGP VAN, a company in Somerville, Massachusetts, that maintains a database on all registered U.S. voters and helps Democratic candidates access the data, released a Facebook app called Social Organizing. The app lets Democratic volunteers log in with Facebook and match their friends with voters in the database. Like the Obama app, NGP VAN’s makes it possible for candidates to execute a peer-to-peer persuasion strategy using Facebook.

So don’t be surprised—especially if you live in a state that is considered up for grabs, such as Ohio or Florida—if you hear from an old college friend with a political pitch based on what the campaign thinks is important to you, as suggested by your Facebook data. If you’ve “liked” a page blaming Obama for high gas prices, you might be reminded about his pro-drilling positions.

Don’t be surprised if you hear from an old friend with a pitch based on what the campaign thinks is important to you. If you’ve “liked” a page ­blaming Obama for high gas prices, you might be reminded about his pro-drilling positions.

The Obama campaign didn’t respond to requests for an interview about its plans, but Joe Trippi, the Democratic strategist who pioneered Internet fund-raising for Dean in 2003 and 2004, expects that the campaign will use sophisticated methods to determine how and when to encourage peer-to-peer appeals in the final weeks of the race. “What’s most important in terms of being able to reach people is to know not only that the voter is undecided—and also what issues, what is holding them up from crossing the line—but who their friends are in the network that might be able to talk to them,” he says. “And then get those friends the information that says, ‘We need you to talk to your friend in Pennsylvania about these three issues that matter most to them.’ This is a field organizer’s dream.” Certainly it is more than Trippi could have dreamed of as a $15-a-day campaign worker knocking on doors in Jones County, Iowa, for Senator Edward Kennedy in 1979, carrying shoeboxes of index cards indicating whether voters said they supported ­Kennedy for the next year’s Democratic presidential nomination.

The Romney campaign’s website also encourages supporters to log in using Facebook, but it requests permission only to view the individual user’s information—not information about the user’s friends “right now,” says the Romney campaign’s digital director, Zac Moffatt. The same is true for the Republican National Committee’s Facebook app. This may change, though, because the Republicans share Trippi’s view. “I think you will start to see, on our side, that app permissions will get changed,” says a Republican digital strategist who spoke on condition of anonymity. “Republicans are working on apps that take advantage of all the things in the Facebook social graph.”

How much information can the campaigns glean this way? Consider that the average friend count on Facebook is 190. As of early August, more than 150,000 people were using the Obama 2012 app. Multiply those numbers and you get more than 28 million people. Now, surely many friend lists overlap, and many of those people aren’t even voters. And some users block the ability of apps like Obama’s to gather information about them when their friends install the programs (a Consumer Reports study, however, found that only 37 percent of users touch app settings). But even if these factors make 90 percent of Obama supporters’ friends useless to the campaign, the president’s campaign app would still have intelligence on 2.8 million American voters who didn’t necessarily take any explicit action to share it.

Persuading just a small percentage of those people could be crucial. In 2000, the contested election that put George W. Bush in office was determined largely by 537 votes in Florida, out of six million cast in that state. And in 2004 Bush beat John Kerry by fewer than 120,000 people out of 5.6 million who voted in Ohio. (Facebook is the virtual battleground within that battleground state. In 2012, just over five million account holders of voting age lived in Ohio—out of a total voting-age population of 8.8 million, according to Well & Lighthouse, a Democratic consulting firm.) Given math like that, the right peer-to-peer and message targeting strategies “could be the difference in swing states,” Trippi says.

Fast and on target

In addition to any peer-to-peer strategies they might employ, the candidates are already waging online advertising campaigns that are more scientifically designed and demographically precise than the ones Obama and John McCain deployed in 2008. Political operatives can now rapidly test ad copy across multiple demographics, getting strategic insights within hours. They can even keep track of exactly which ads individual computer owners have clicked on.

These abilities were brought to bear in an ad campaign that rolled out in March of 2010, when President Obama signed the Patient Protection and Affordable Care Act—so-called Obamacare.

The midterm elections were just eight months away, and the president was concerned for a vulnerable ally, Harry Reid of Nevada, the Senate majority leader. On the health care issue alone, Reid’s online strategist, Jon-David Schlough, developed 18 sets of targeted advertisements for people in different demographic groups. For example, the version geared to students pointed out that the legislation would let them keep their parents’ insurance until age 26; the one for the elderly focused on what it would do to close a Medicaid benefit gap known as the doughnut hole.

Then, for each of the 18 campaigns, different versions were tested on Facebook. Schlough says the site gave him access to a wide range of demographic groups, made it possible to place small ads at low rates, and offered easy ways to experiment rapidly with different combinations of headline, image, and text. The versions that generated the most clicks would get wider distribution on multiple websites.

Eventually, the campaign could be sure that, say, an ad about being able to stay on parental insurance plans would be shown to a specific 25-year-old four times a day for two weeks. It’s called nanotargeting, and “it’s now a component of all campaigns,” says Schlough, the founder of Well & Lighthouse. “Political types are used to large data-set analysis on things like polling data and turnout data. But the fact that so much more data is available, so much faster, is allowing us to innovate a lot quicker.”

For Reid, such innovations might have been decisive. Consider that his opponent, Tea Party favorite Sharron Angle, spent about as much as Reid and was ahead in the polls in the weeks leading up to the election. In the end, Reid won by more than 5 percentage points.

David Talbot is Technology Review’s chief correspondent.

This article was revised on August 15, 2012.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.