This is how your personal data helps candidates predict your vote

Read the Full Transcript

  • JEFF GREENFIELD:

    It has been the core of political communication: speaking to the multitudes, by any and all methods: the street corner rally; the leaflets, the radio speeches…

  • FRANKLIN ROOSEVELT:

    "This generation of Americans has a rendezvous with destiny!"]

  • JEFF GREENFIELD:

    … the T-V ads.

  • TV AD:

    "It's morning, again, in America!"

  • JEFF GREENFIELD:

    But in recent years, technology, in the form of data crunching computers, has upended just about every traditional notion about how to reach voters, how do you talk to them, and how to persuade them to vote. Indeed, modern campaigns more and more are taking the mass out of mass communications, choosing instead to target voters one at a time.

    It's called "micro-targeting." And it's something almost everyone with a computer or smartphone experiences. Companies send you e-mail pitches based on what you've bought from them; Amazon knows what books to recommend based on the books you've already bought; your neighbor with different tastes gets a completely different list. It's why your computer looking at the same website will have different ads from those of your neighbor, based on the digital trails you've left.

    What's happened in the last decade-and-a-half is that the tools of consumer marketing—the ability to gather and interpret enormous amounts of data in order to know more about individuals—have been adapted and refined by political campaigns.

  • SASHA ISSENBERG:

    I think we're talking about using the reams of available new data and statistical modeling tools to find new ways to segment the electorate.

  • JEFF GREENFIELD:

    Sasha Issenberg literally wrote the book — Victory Lab — on the arrival of so-called "analytics".

  • SASHA ISSENBERG:

    What analytics is doing is collecting the hundreds if not thousands of a variables that are available about every adult in the United States — and looking for patterns that are less visible to the naked eye, and saying, 'Okay, can we come up with a better likelihood of assuming somebody's propensity to be a Democrat, or to vote in a primary, or to turn out to vote at all.

  • JEFF GREENFIELD:

    It's happening across the political landscape–Left and Right. Months before Democrat Hillary Clinton announced for President, a pro-Clinton Super PAC—"Ready for Hillary"—was gathering data on social media like Facebook to find and target potential supporters. Republican Senator Ted Cruz has explicitly acknowledged his campaign "is very much a data-driven, grassroots-driven campaign."

    It — and allied Super PACs — have already spent millions of dollars searching Facebook and other sites to find likely Cruz voters. And according to the Washington Post, the Cruz campaign explicitly credits its use of "analytics" to the Senator's recent rise in the polls.

    So—how does micro-targeting work? It all starts with "the voter file"—public information about where you live, whether you're a registered party member, and whether you've voted in primaries and general elections—and then taking it one step further.

  • SASHA ISSENBERG:

    Every magazine subscription that you've had, if you're on mailing lists from catalogs. A lot of states have their hunting licenses, or gun permits are available.

  • JEFF GREENFIELD:

    And here's the key: thanks to computing power, campaigns can look at traits that are common to their supporters—people who've donated or volunteered. By finding other people with those same traits, they are likely to find more supporters.

    That's what President George W. Bush's re-election campaign did in 2004, when it looked for supporters in heavily Democratic districts in the key swing state of Ohio.

    In earlier times, this would have been futile; but now the Bush campaign could find supporters scattered throughout hostile territory…kind of like finding raisins in a bowl of raisin bran.

  • SASHA ISSENBERG:

    And so the Bush campaign and the Republican National Committee made a– concerted effort between– George Bush's election in 2000 and his re-election in '04 to build the infrastructure to make this work in politics.

    And one of their big things was saying, "Yeah, let's– let's figure– we know that even in the– the least Republican county in America, 15%, 20% of the votes are– are ours, and– I– if we know how to go find those people– we can be targeted and lower the risk of– of inadvertently mobilizing the wrong people."

  • JEFF GREENFIELD:

    Bush won Ohio by two percent; a loss there would have cost him the Presidency. It was a lesson the underdog Barack Obama campaign took to heart in the 2008 Democratic contest.

  • MICHAEL SIMON:

    Yeah, necessity is the mother of invention.

  • JEFF GREENFIELD:

    Michael Simon headed Barack Obama's "targeting and analytics" team in 2008.

  • MICHAEL SIMON:

    In that contest of Hillary versus Biden versus Obama versus Edwards, etcetera, there was a need to do something fairly groundbreaking, because the electoral pathway for Barack Obama simply didn't exist if the electorate as it typically was configured was going to turn out.…So, what we were really trying to do was to break down the geographic barrier.

    And all these different pieces of sort of digital footprints that we leave behind, that were some telltale signs.

  • JEFF GREENFIELD:

    This information meant, for example, that an Obama canvasser would not go to every door in a neighborhood, but go only to the doors where potential supporters lived…and would go with a message shaped to the interests of that potential voter.

    And in the fall, the same approach that worked for George Bush in Ohio four years earlier proved critical for Obama in at least one battleground state.

  • SASHA ISSENBERG:

    One of the underappreciated reasons why they were able in a cycle to take a state like Virginia– where no Democrat had run a– serious campaign for the presidency in– in 30 years– and make it competitive was that they said, "Yeah, there are a lot of conservative counties here where a Democrat only gets 21% of the vote, but we now have the data to figure out how we can make that 24% of the vote just by mobilizing a few more people."

  • JEFF GREENFIELD:

    By 2012, Obama's team embedded analytics into every aspect of the re-election campaign.

    Everything from where the candidates and spouses should go, to what kind of ads should run where and at what time…was rooted in mountains of data which enabled the campaign to test competing messages scientifically.

  • SASHA ISSENBERG:

    So I randomly assign, let's say 10,000 people into two groups, one of them gets the piece of mail that says Marco Rubio's tax plan would raise taxes on middle income people, and 5,000 of them get one that says he would lower taxes on rich people.

    And I see that, 'Wow, one group of people now supports Rubio at a rate three points higher than the other.'

  • JEFF GREENFIELD:

    And because you know about those 10,000 people, you can make reasonable assumptions about a far larger people you've never polled.

  • SASHA ISSENBERG:

    Yeah, it's what they call 'look-alike modeling.' And then you're going out and maybe up to 250 million other American adults saying, "Who are the people who statistically resemble them, and how much do they resemble the type of person whom I know cares about this or will vote that way?" And then you can go out and prioritize among that group how you want to communicate with them.

  • JEFF GREENFIELD:

    In the last two Presidential campaigns, Democrats more skillfully wielded these analytic tools. But in 2016, both sides appear to be well-armed.

    Audience Partners is a firm in suburban Philadelphia that offers targeted messaging largely to Republican clients. Jeff Dittus, the firm's managing partner says micro-targeting will be even more prevalent this time out

  • JEFF DITTUS:

    There's a group of folks, and I suspect it's maybe 35 percent of the population, that you cannot reach with mass media.

    All they do is look at their phones and their computers to get news. And so I see it as a necessary thing for the democracy to be able to push messages and reach those people, because you can't reach them any other way today.

  • JEFF GREENFIELD:

    Dittus showed us how his company used the voter file in Chris Christie's 2013 re-election campaign for governor of New Jersey—a state with some 5.4 million voters—to specifically target registered Republicans with a weak voting record.

  • JEFF DITTUS:

    So I would exclude the people who let's just say in '09 who did show up. And now my audience is down to 296,000 people.

  • JEFF GREENFIELD:

    Of course, as with any new technology, there's a danger of overstating its power. Lynn Vavrek teaches political science at UCLA.

  • LYNN VAVRECK:

    I think that if you have the money to spend on building an analytics shop and sustaining it, then yes. I think you would be a little foolish not to use it.

    The thing I think that is important to remember is that that may be the icing on the cake that gets you that 538th vote in one county in Florida. But you still need the cake.

  • JEFF GREENFIELD:

    If analytics IS overrated, one candidate seems to be testing that idea. GOP frontrunner Donald Trump has apparently no analytics operation at all–relying instead on his celebrity, personality and media appearances, as well as a torrent of Tweets, to drive his campaign.

  • LYNN VAVRECK:

    The big stuff in campaigns, I think, is national context. What's going on in the country right now? What's the state of the nation's economy? Who are the candidates? Given the state of the world and who they are, how do they try to beat one another? if you know those things, you can go a pretty long way toward having a sense of how a presidential election, at least, is going to turn out.

  • JEFF GREENFIELD:

    Even former Obama analytics chief Michael Simon says analytics are less meaningful when more big issues dominate an election…as huge Democratic losses in the 2010 and 2014 mid-terms demonstrated.

  • MICHAEL SIMON:

    It's going to get you a couple of points, and it's going to help build, you know, a robust organization that can help carry the ball or the remaining few yards to the end zone. It's not going to help buttress against a tsunami.

    And, you know, when the macro-forces are as stacked or as lined up as they were against the Democrats in those two midterms, there's only so much you can do.

Listen to this Segment