POLITICS
11/10/2016 10:10 pm ET | Updated Nov 11, 2016

Hillary Clinton's Vaunted GOTV Operation May Have Turned Out Trump Voters

A focus on big data over people may have backfired.

“Why aren’t I 50 points ahead?”

That’s the question that Hillary Clinton asked a group of labor organizers in late September when she was up by 7 points in the national polls.

The outcome of the 2016 election stunned the members of the Democratic establishment,  many of whom left their “get out the vote” turf early to don cocktail dresses and ties and be first in line at the bar at the victory party that would surely go on all night. There they expected to toast the first woman president and her much vaunted “third term” for President Obama’s administration. But maybe leaving turf early wasn’t such a bad idea. Because out in the GOTV effort, something may have gone very, very wrong.

The media have made much ado about the absence of a real Trump organization on the ground in the states, but has largely failed to interrogate the hype around the Clinton campaign’s ground operations. As the post-election day hangover wears off, an examination of the mechanics behind the Clinton’s get out the vote efforts ― reaching out to Clinton voters in key states at the door, on the phone or by text messages ― reveals evidence of what appears to be a pretty shocking truth. Clinton volunteers were inadvertently turning out Trump voters. Possibly in significant numbers.

Volunteers for the Clinton campaign in Pennsylvania, Ohio and North Carolina have reported that when reminding people to vote, they encountered a significant number of Trump voters. Anecdotal evidence points to anywhere from five to 25 percent of contacts were inadvertently targeted to Trump supporters.

 

Voter targeting is not a new idea, but over the past few cycles electoral field organizing has become intoxicated by the concept of using “big data” to microtarget voters.

 

This is a big deal because when voters are engaged by a volunteer they are significantly more likely to cast a ballot in an election. To make matters worse, because Republicans had a non-existent ground game in many areas this cycle, this powerful reminder from a Clinton volunteer to get out and vote might have been the only personalized GOTV communication these Trump voters received.

The campaign’s text messaging GOTV effort may have been the worst offender. Volunteers reported as many as 30% of the replies they received from voters they were urging to get out were Trump supporters.

Voter targeting is not a new idea, but over the past few cycles electoral field organizing has become intoxicated by the concept of using “big data” to microtarget voters. Just like Amazon knows what to show you on their front page based on your past purchases, the idea goes, campaigns should be able to predict who their voters are based on past voting behavior and other commercially available data that can be matched on the vote file. In the avalanche of stories and books about Obama’s two victories, commentators have credited microtargeting as a major factor in his success ― in our opinion, mostly unjustifiably.

The problem is the lack of actual data. General election voting choices are of course secret. Only a relatively small number of primary voters’ partisan choices are public record, and not in all states. The much-hyped commercial data, upon closer examination, is either not useful or simply not available for most voters. Most targeting choices are not micro, but macro, such as targeting African Americans or young people. When you’re winning by a large margin, as Obama did both times, you can’t go wrong with such choices. When you’re losing, however, certain macro-targeting choices amount to doing your opponent’s GOTV work for them.

How could the Clinton campaign have gotten the targeting so wrong? In the rust belt states, there were many registered Democrats voting for Trump, not Clinton. The extent of their errors probably had a lot to do with the fact that party registration ― normally one of the best predictors of political behavior ― would have wreaked havoc with their models in many geographies. And most of us were surprised to learn the day after the election that 53 percent of white women cast their ballots for Trump. While the campaign’s failures may not have been unique, the deeply damaging outcome of this election casts the impacts of the Democratic campaign industrial complex’s possibly quite mistaken assumptions in stark relief.

The rise of something we call “small organizing” is a big part of the problem. Small organizing is the product of a complex and ultimately toxic stew that includes professionalization of politics, attempts by the liberal establishment to channel radical impulses of working class people and people of color into incrementalist politics, and the ascendancy of a bipartisan technocratic elite in both parties that has been accelerating the concentration of power in the hands of an increasingly small number of mega corporations and institutions.  

In small organizing, a campaign invests in a “big data” approach to narrow the number of people that must be engaged. They do this usually by engaging an expensive consulting firm that relies on predictive modeling, micro targeting, and message testing and segmentation to design a get out the vote program that staff and volunteers then execute.

The opposite of this model is a “big organizing” approach which we write about in our new book Rules for Revolutionaries: How Big Organizing Can Change Everything. Big organizing is how the Bernie Sanders campaign and other grassroots focused efforts tackle turn out. Instead of using technology to guess at who should be mobilized on Election Day, Bernie mobilized hundreds of thousands of volunteers to build a voter contact machine that could call, text or knock on the door of voters and identify who was for Bernie, who was against him and who was undecided. This was done in the months before Election Day so that during the crucial week of GOTV prior to voting, volunteers were contacting voters they had affirmatively identified as supporting the campaign, not voters an expensive stable of consultants guessed would be for the candidate.

We’ll never know the extent to which earnest volunteers helping the Clinton campaign get out the vote helped put the Trump campaign over the top in states like Wisconsin, Pennsylvania and Michigan. But we do know that it’s time for Democrats to take a hard look at their small organizing approach and plot a return to big organizing. When campaigns stop over-relying on predictive models and start talking to voters on a large scale, they won’t only discover who is for their candidate and who is not. They also might learn a few things about how their messages and policies are resonating with voters or gain the courage they need to forge a new way forward.

While Clinton edged Trump out in the popular vote, she lost the election decisively. Her defeat heralds the end of Clintonism and the neoliberal policies that were so widely rejected by the American people. But it also should mark the end of the small organizing approach to winning elections that has alienated the Democratic base and serves to obscure for an out-of-touch elite the painful realities that most Americans are facing.

In this political moment with Trump taking power and so much at stake, it’s time to get back to big organizing. The Bernie Sanders campaign provided a glimpse of what is possible ― a coming together of working class whites, working class Latinos and working class African Americans as a potent political force ― but it’s up to us to undertake the hard work of making this possibility a reality.

This story is written by Becky Bond and Zack Exley, who were senior advisers to Bernie Sanders’ 2016 presidential campaign. They are the authors of  Rules for Revolutionariess: How Big Organizing Can Change Everything.

CONVERSATIONS