Jim Tankersley reports in the Washington Post:
Hillary Clinton is running arguably the most digital presidential campaign in U.S. history. Donald Trump is running one of the most analog campaigns in recent memory. The Clinton team is bent on finding more effective ways to identify supporters and ensure they cast ballots; Trump is, unapologetically, sticking to an 1980s-era focus on courting attention and voters via television. A statistical model c(an) predict an individual vote. But that requires planning, a budget and attention to detail
Hillary Clinton is running arguably the most digital presidential campaign in U.S. history. Donald Trump is running one of the most analog campaigns in recent memory. The Clinton team is bent on finding more effective ways to identify supporters and ensure they cast ballots; Trump is, famously and unapologetically, sticking to an 1980s-era focus on courting attention and voters via television.
It is a deep contrast in how the Democratic and Republican nominees allocate their time, staff and campaign funds, and one that the entire political world would be buzzing about in almost any other election year.
But this is not an ordinary year, and so the candidates' strategies have received less attention than, for example, the Obama campaign's vaunted data-mining operations in 2008 and 2012. That's why we're lucky to have Sasha Issenberg — a Bloomberg reporter and the author of (the recently updated and reissued) "The Victory Lab: The Secret Science of Winning Campaigns" — on the beat.
This week, Issenberg pulled back the curtain on Clinton's voter-targeting operation and how it is rethinking the age-old practice of grouping states geographically, for campaign purposes. In a follow-up online chat with me, he explained that report, and Clinton's approach, and he contrasted it with Trump — who Issenberg says is running an old-school, on-brand, but probably unsuccessful, style of campaign.
"He sees his candidacy as an extension of the mechanism of becoming a celebrity: It's about using television to get in front of as large an audience as possible to get as many people as you can to like you," Issenberg told me. He added: "I think that dramatically fails to appreciate the extent to which campaigns are not just about changing people's opinions to get them to like you. "
A transcript of our chat follows, lightly edited for clarity.
Tankersley: You literally wrote the book on how candidates use technology and data to win presidential elections, and you wrote a spectacular piece this week about how the Clinton campaign is rethinking one of the most basic ways that campaigns think about how and when to target voters. Can you walk us through what they're doing?
Issenberg: A big part of a presidential campaign is breaking down the country into manageable chunks, with some sort of chain of command that allows resources to be moved among states and strategic decisions to translate down into tactical directions that can be implemented locally. As far back as I can tell, campaigns have done this by region. It's perfectly logical, the same way that the party committees assign their communications and political directors: Someone has the Northeast, someone has the Midwest, etc. The actual groupings change a bit every four years, but the logic has always held.
The Clinton campaign is doing things differently — dispensing altogether with the geographical logic for breaking down the country. They've split it into three, with each "pod" responsible for states where there will be a common strategic approach. So one pod includes big diverse states where mobilization will be essential to her victory, and one small states where persuasion will play a bigger role. Then, perhaps most interestingly, they've broken off states where a significant portion of the electorate will vote early or by mail.
It's a radically different way of looking at the country and what states have in common with one another.
Tankersley: What's the advantage of breaking it down that way, do they think?
Issenberg: One way to think of it is why the old way doesn't really make a lot of sense anymore. It might have made sense to put the same political person in charge of Florida and Georgia because if you were driving among them, they're close, or because there are media markets that spill across state lines. But really your campaign plan will be very different. The way as a Democrat that you win Georgia may have more in common with the way you win Ohio — by building a massive infrastructure for turnout in college towns and largely African American cities — than in Florida. So why not have the people in headquarters who are overseeing your field, digital, targeting efforts in Georgia also be dealing with your team in Ohio?
What they share in terms of tactics and timetables probably matters more for implementing a campaign plan, managing staff and budgets, than the fact that they have different weather patterns.
Tankersley: I see this as a natural evolution of the data-driven insights that have been changing campaign tactics the last few cycles. Is that right?
Issenberg: On one level, this structure reflects an even more basic form of progress, just a willingness to question precedent. Every institution that needs to break down the country into manageable chunks starts with geography, because it's intuitive. (Even sports leagues that don't actually divide teams by proximity use the language of "Eastern Division" or "Western Conference" to characterize them.) But if you were starting a presidential campaign from scratch and not just using a hand-me-down org. chart from 1976, is there a reason why you're grouping North Carolina with Virginia and not with Nevada?
I've spent a lot of time reporting on how data-driven innovations play out in the field — how campaigns figure out which doors to knock on, which phones to call, etc. What we're seeing in Hillary's Brooklyn headquarters is, in essence, the trickle-up effect of all those innovations: the midlevel corporate structure rearranging itself to better reflect what's going on at ground level in the field.
To me that reflects an institution that is doing a good job of thinking holistically about these innovations: not just as a series of discrete tools, but rethinking the broader structure of this billion-dollar corporation so that its tools are being deployed more efficiently.
Tankersley: That's an instructive comparison: the campaign as a corporation. Especially since the Republican campaign is led by a man who made himself famous in business. So I think it's fair to ask: What do we know about how the Trump Corporation (campaign version) is thinking about its tools and how to deploy them?
Issenberg: Broadly speaking, it seems like Trump's campaign has been overwhelmingly focused on ways to leverage media coverage: interviews, rallies, online social content. There has been hardly any concerted use of any mode of communication that isn't fundamentally amplifying the candidate's voice across free mass media. There's been very little targeted voter contact, either online or offline. Only today the campaign appears to be starting its first general-election television ad buy, although it looks to be very meager by the standards of a national campaign — and certainly compared to what the Clinton campaign and its allies have already put on the air.
Part of the reason he's able to have such a small, centralized campaign structure is with so few, barely targeted tools there are ultimately few tactical decisions that need to be made on a daily basis. Which city does the plane fly to? What's the subject of today's tweetstorm or Instagram video? Which three TV producers do we say yes to today?
He's right to mock the size of Clinton's campaign, because it would be wasteful to have 900 people overseeing just a handful of tactical decisions all day, very few of which need to be locally implemented in different ways across states.
Tankersley: I believe you just tweeted a story saying his campaign isn't even spending much on yard signs.
Issenberg: I'm multitasking.
Tankersley: Aren't we all?
Issenberg: That's some good reporting from my colleagues at Bloomberg Politics. I know of campaigns that have decided not to print or distribute yard signs as a matter of principle or efficiency. I don't think that's what's happening here.
Tankersley: What you're describing here is a Trump campaign that is running what we could charitably call a "throwback" strategy to the days when television dominated campaigns — which would be what, the early ’90s, at the very latest? — or less charitably called an "analog" campaign in a very digital era.
In a way, that strategy matches his campaign message, which is in large part a nostalgic appeal for an America that felt safer, more prosperous and more … culturally comfortable … for a particular class of blue-collar, white workers.
Issenberg: The one difference, though, is that what I would think of as those high-modernist campaigns of the late-20th century all relied on a lot of paid television, not just free exposure through the media. Up until 1988, basically, campaigns would cut three checks — one to each of the television networks — and communicate with the electorate through national ads.
In the 1980s, direct mail became useful for persuasion, and then over the last 15 years, we've seen a renaissance of highly targeted individual contact because the data and analytics are now available to segment the electorate in all sorts of previously unimaginable ways.
Trump is very much a throwback to that old mass-media world — this is a guy who seems to prize being on the cover of Time or featured in "60 Minutes" above anything else — but has also decided to run for president on the cheap. So he's still relying on the three national networks (and cable news), but since he isn't paying for airtime, he is reliant on the media to filter his message in a way that past candidates haven't been. No wonder he's in such a love-hate relationship with us.
Tankersley: Can you make an argument for that approach, strategically? Other than "it saves him money"? In what ways might it be more effective than the Clinton approach?
Issenberg: I'll say that I think Trump has a more coherent worldview about campaigns than many politicians, and his tactics actually do a pretty good job of reflecting his strategic assumptions. He considers campaigns to be purely a candidate-driven, mass-media exercise. One could also say, perhaps less charitably, that he sees his candidacy as an extension of the mechanism of becoming a celebrity: It's about using television to get in front of as large an audience as possible to get as many people as you can to like you. Even as his campaign has grown and changed, he has been remarkably disciplined at not spending much time or money on anything that doesn't reflect that approach.
Now I think that dramatically fails to appreciate the extent to which campaigns are not just about changing people's opinions to get them to like you. Now more than ever, thanks to partisan polarization, campaigns are about modifying the behavior of people who already like you — getting the unregistered to register, mobilizing infrequent voters to turn out. That is best done through targeted communications that don't involve the candidate.
We know from dozens if not hundreds of randomized field experiments that the best way to turn a non-voter into a voter is to have a well-trained volunteer from his or her neighborhood conduct a high-quality face-to-face interaction at the doorstep. The Clinton campaign is building the structure to do a lot of that, at scale, before voters they have modeled as most likely to change their behavior as a result. That doesn't fit into Trump's idea of what an election is about. To his credit, though, unlike a lot of candidates, he doesn't go through the motions of halfheartedly opening field offices — or printing up yard signs to fill them with — without understanding how they fit into his broader strategy.
Tankersley: That implies that Trump needs to be wildly successful in his quest to get people to like him, correct? He either needs to get them to like him so much that they will get themselves to the polls at a historically high rate (with essentially no knock-and-drag from the campaign). And/or he needs to get so many people to like him that he can afford massive leakage on Election Day from his lack of an advanced turnout operation?
Issenberg: Right. If polls showed him carrying such a significant share of persuadable voters we could wonder whether Clinton could offset those gains through mobilization. But if he's not winning them over at a high rate, he's basically shut himself off from finding votes elsewhere. And we're getting to the point where even if Trump decided he wanted to write a $500 million check to build a national ground organization to develop that kind of turnout capacity, he wouldn't have the time to hire staff, open offices, recruit and train volunteers, etc., before voting begins in some battlegrounds five weeks from now.
Tankersley: So the analog Trump strategy is true to his brand, and efficient in terms of spending, but risky in terms of getting the votes he actually needs to win?
Issenberg: Yes. Not just his brand, but his entire way of seeing the world. I think it will prove flawed as electoral strategy, but it is intellectually coherent.
Tankersley: Let's flip the question, then: If you changed nothing else about the Trump campaign, but you gave him the Clinton field/data operation, how much better would his chances to win look right now?
Issenberg: I'm sure he could squeeze more votes out of his coalition, but there's also less room to gain. A mobilization strategy will yield far greater rewards when your supporters are more likely to be young, people of color, transient — some of the least reliable parts of the electorate — than the older whites who are the core of the Trump coalition. He would also find it a lot more difficult to build a field organization that gets to a constituency more likely to live in the type of rural or exurban areas where you can't easily walk from door to door the way you can in big cities, inner-ring suburbs and college towns. He would probably come away with more votes, but I'm skeptical it would amount to a good return on his investment.
Especially with this much fluidity in the electorate, he may find it more efficient to win his next marginal vote through persuasion — which really nets him two votes if it comes out of Clinton's support — than through mobilization.
Tankersley: How should he be strategically allocating his money now if he wants to win? Just buying a ton of TV ad time? Or is his coalition sufficiently size-limited that none of this strategy matters if he can't broaden his appeal beyond his core?
Issenberg: He could certainly be doing more persuasion via targeted channels, particularly direct mail, online display and digital ads. One of the problems the Trump campaign has faced is that a lot of its persuasion messaging has seemed to be a zero-sum game: What he says that helps him with, to oversimplify the demographics here quite a bit, downscale whites hurts him with upscale whites.
This is a particular problem when he delivers every message to every population at the same volume: When he talks about immigration or trade in ways that appeal to blue-collar Democrats, it reaches white-collar Republicans who are likely to be turned off by it.
A modern campaign could, frankly, build a statistical model that predicts the likelihood an individual voter doesn't like Muslims — they might come up with a more genteel name for it — and just target that audience with information on his "extreme vetting" proposal and avoid mentioning it in the speeches that get wide exposure. But that requires planning, a budget and attention to detail that has been rare from Trump's campaign.
0 comments:
Post a Comment