AS THE SUMMER OF SANDERS turned into the eventful fall, we had succeeded in demonstrating the tremendous energy and enthusiasm there was for Bernie all across the country. And the public polling in the first two states confirmed it. The Des Moines Register poll taken at the end of May had Hillary Clinton leading by 41 points. Their poll done at the end of August had Clinton’s lead down to 7 points. The mid-July NBC News/Marist New Hampshire poll had Hillary Clinton up 10 points. The same poll at the end of August had Bernie up by 9 points.
Despite all the attempts to minimize or marginalize the huge summer crowds, Bernie’s campaign had caught fire from one end of the country to the other. Unable to dispute the thousands and in some cases tens of thousands of people coming out to hear him, his critics argued that crowds were one thing, but getting votes was quite another. On that point they were certainly right. Organizing the huge numbers of people that Bernie had mobilized was a daunting task.
But even though there was still a tremendous amount of skepticism publicly, the Clinton campaign wasn’t buying its own spin, even if the media was. They knew they were losing ground. On August 3, they began spending millions of dollars on television advertising in Iowa and New Hampshire. They would be on the air in both those states until their respective February contests—approximately six months’ worth of uninterrupted advertising in the first two states.
The Clinton campaign’s decision to go on television so early put us in a difficult position. Should we respond in kind? Or were they wasting a bunch of money that would eat into their financial advantage? We were hampered in our decision making by our lack of a pollster. Without one, it was impossible to know for sure how voters were responding, because we had no way of asking them directly, either with polls or focus groups. We were flying without instruments.
Bernie resisted hiring a pollster at that time. He was concerned that a pollster would try to change his messaging for political advantage. Bernie already knew what his message to the American people would be, so he certainly didn’t need a pollster for that—in his mind, he didn’t need one at all. I think it’s fair to describe his position as adamant.
Bernie did understand the power of television, however. We met to discuss how to respond. Bernie wanted to hear people’s views. He was of two minds on the issue. I was leaning toward going on the air. We had a problem with name ID, and television ads would help fix that (as they did later in the campaign). Television ads, for all their limitations, also would have allowed us to deliver a message straight to voters without filtering by the media or spin by our opponents. Given all the media focus on Clinton’s emails, I’m sure the Clinton campaign felt the same way.
Tad Devine was the one who ultimately prevailed, with his view that we hold off. He viewed the Clinton campaign’s move as one of weakness. Tad thought that we should keep our powder dry, given the resource imbalance between the campaigns. He preferred that we maximize the impact of our advertising when voters would be paying more attention closer to the election.
It turned out to be the right decision. I had doubts at the time. We were relying on instinct, experience, and judgment without the benefit of concrete research about the effectiveness of the Clinton ad campaign. For now, we would end up relying on public polling to monitor the race.
That would change on October 1, when we brought on Ben Tulchin to do our polling. Ben’s team included senior analyst Ben Krompak and analyst Kiel Brunner. Ben Tulchin first met Bernie back in March, in the lobby of a boutique hotel in San Francisco. Tulchin had heard that Bernie was going to be in town and had asked Tad and Mark to set up a meeting.
Amid the art deco décor, Bernie spent most of the time asking Tulchin why he should hire a pollster at all. Bernie asked, “Ben, explain to me why I should spend money from my supporters, who may be living off of Social Security, on polling?” Tulchin made the case then that I and others would make in the summer and early fall, and which Bernie ultimately found persuasive. The point of polling was not to craft Bernie’s message or measure the horse race. Rather, it was key to understanding the race in a way that would allow us to target what was becoming a substantial amount of resources. For instance, it would help us figure out what parts of Bernie’s already existing message should be highlighted in television ads. Bernie, after all, had an almost two-hour stump speech. We couldn’t put it all on television. Only Donald Trump had the luxury of having almost every speech broadcast on television in its entirety. Polling data would also let us target resources among the states. We were raising money in the summer and fall, but still we would not be able to spend with equal intensity everywhere. Hard choices had to be made, and they could only be made with research. It was this efficiency argument that really made the most sense to Bernie. It appealed to his Yankee frugality. Now that we were raising money, it had to be spent wisely.
In October, we polled the first four states and also Georgia. There was good news in the polling, but there were also some serious warning signs. Almost all the polling (except in South Carolina) was done after the CNN debate, because we wanted to make sure we captured any bump (or decline) that the debate would cause. In Iowa, our polling showed us down 18 points (53 to 35). That was a huge deficit, and it showed some erosion in support from earlier public polls. Because Vice President Biden had not dropped out yet, we also polled with him in the race. In that scenario, it was Clinton, 45; Sanders, 33; Biden, 13. Clinton’s margin was cut by a third. In the Hawkeye State, Biden was taking far more from Clinton than from us.
In New Hampshire, the poll had Bernie up by 5 points (48 to 43). This was very similar to the public polls at the end of the summer. With Biden in the race, the result was Sanders, 42; Clinton, 37; Biden, 12. Biden took about equally from each candidate.
Nevada was a much tougher state for Bernie. Hillary Clinton was ahead by 33 points (61 to 28). As in Iowa, Biden’s entry into the race would have helped Bernie. The result in that case was Clinton, 48; Sanders, 23; and Biden, 20—a deficit of 25.
South Carolina was even worse. Clinton was ahead by 52 points (69 to 17). Adding Biden helped somewhat, with Clinton taking 51 percent, Sanders at 12 percent, and Biden at 25 percent. That lowered Bernie’s deficit against Clinton to 39 points but pushed him into a solid third place. Regardless of whether Biden was in or not, South Carolina was going to be a tough state for the campaign.
Because the polling in South Carolina was not positive, we decided to test another southern state to see if we might find one that was more favorable. We wanted to make a stand somewhere in the South, but we wanted to invest resources where they would most likely produce results. We decided to test Georgia. It was better, but only barely. Clinton was ahead, 69 to 20—still a 49-point gap. With Biden in the race, the outcome was Clinton, 51; Biden, 28; and Sanders, 16. That was a 35-point gap with Clinton. Ouch!
Given that South Carolina and Georgia were very similar, we decided to fight in South Carolina because it was smaller and less expensive. The South Carolina primary was only a couple of days before most of the other southern states voted. If we could improve there and have a respectable showing, it might produce a bump a couple of days later.
Overall, our situation was precarious. In Iowa, we were down almost 20 points. Hillary Clinton erased Barack Obama’s lead and won New Hampshire in 2008, even though he had finished solidly in first place in the Iowa caucus. If we really stumbled in Iowa, our single-digit lead in New Hampshire could very well evaporate. Our weakness in Nevada and South Carolina gave greater weight to the need for momentum coming out of the first two states. In addition, it was clear we had to increase our support among African American and Latino voters.
One issue that weighed on voters’ minds was Bernie’s electability. It is a sad irony, given the outcome in November 2016, that a year earlier we were spending time and resources trying to figure out how to convey to voters that Bernie Sanders was more electable than Hillary Clinton. As early as the summer of 2015, public polling consistently showed that Bernie Sanders was a stronger candidate against Trump (and most of the other Republicans) than Hillary Clinton would be. The media minimized and dismissed these polls as soon as they reported them.
As part of his presentation to get retained by the campaign, Ben Tulchin prepared a memo at the end of July showing that Bernie was the only major contender who had net positive ratings with voters. Ben’s data came from public polls. Bernie was beating all the Republicans tested, even though only 45 percent of voters in a CNN/ORC poll conducted in late July had enough information to give him positive or negative ratings. Hillary Clinton, on the other hand, was underwater with voters (as were all the Republicans tested) and looked to be slipping with voters since her announcement.
Our internal polling did reveal a number of bright points, which Ben Tulchin laid out in an October 30 memo. The first point was that, despite still being behind, Bernie had made considerable gains with voters since his announcement with no paid media whatsoever. He was even winning in New Hampshire. The Clinton campaign had been on TV in New Hampshire and Iowa since August 6. The other important fact was that the two central issues of Bernie’s campaign—wealth and income inequality and a corrupt campaign finance system—were the top issues for Democratic primary voters.
The information from Tulchin’s polling would guide the program of paid television and radio advertising that the campaign began in November. To continue the process of introducing Bernie to voters in Iowa and New Hampshire, we would present a biographical ad that tracked him from his childhood, through his years in college fighting segregation, to being mayor of Burlington, to his time in Congress. We would follow that up with a pair of ads that combined the two most powerful elements of Bernie’s platform—a rigged economy held up by a corrupt system of campaign finance. The third piece would be an effectiveness ad that highlighted his record of accomplishments in Congress and as mayor of Burlington. Finally, we decided to run testimonial ads, including one from a dairy farmer and one from a nurse—both Vermonters. There is no better validator than your neighbor. We wanted people in Iowa and New Hampshire to see that Vermonters sincerely supported Bernie.
This new research also allowed us to flesh out the state-by-state road map to the nomination that Mark Longabaugh had developed and was continually refining. Winning a presidential nomination is not a single contest like a Senate race. It is fought out serially in every state and territory and in the District of Columbia and with Democrats abroad (American Democrats living abroad who vote in a worldwide primary).
Our early focus was the first four states. As an insurgent campaign, we had to win early. We knew Bernie would not get support from most of the establishment without a strong show of strength. Early wins would also demonstrate to the media, our small donors, our volunteers, and voters that this was a campaign that could go the distance. That strong focus on the first four contests meant that the vast majority of Bernie’s time was spent in those states.
The strategy fit well with his idea of how Bernie wanted to run. The first four states rely heavily on the person-to-person grassroots organizing that he prefers. He saw his national rallies as a complement to our work on the ground in the early states. “The people in Iowa watch the national news,” he would remind us, whenever we suggested that he spend more time on the ground in the early states. To him, the large rallies he would hold throughout the fall all over the country were a key element to doing well in the first four.
A heavy focus on the first four was in line with my thinking on creating local superiority against the Clinton juggernaut. In military terms, local superiority is achieved when a smaller force concentrates its numbers on one part of the battlefield, or in one battle, to gain a numerical advantage that allows it to defeat one part of a larger force. This can be followed up by defeating the larger force piecemeal or by inflicting enough damage that the larger force ultimately loses the will to fight. Because we were outmatched organizationally, the only way to prevail, in my view, was to institutionally and financially achieve local superiority.
We could be as big as they were, if we were focused in terms of resource allocation and Bernie’s time. In these smaller states, the Clinton resource advantage would be irrelevant. We could bring as many forces to bear as they could—or more than they could.
Of the first four states, New Hampshire seemed like the most likely place for us to score a victory. It was our “must-win” if we were to have any chance at the nomination. New Hampshire was next door to Vermont. Vermont television bled into western and northern New Hampshire, so the voters there had seen Bernie on TV for decades.
We would also have to make an early play for Iowa. We were behind in Iowa, and we would have to contend with the state’s complicated caucus process. But much of Iowa is small towns and rural communities—places like those in which Bernie did very well in Vermont. And unlike New Hampshire, Iowa had never taken to the Clintons. Hillary Clinton had lost in 2008, as had Bill Clinton in 1992.
Nevada and South Carolina, as our polling revealed, would be tougher. Tad, in particular, was of the view that victory in the first four states would constitute a knockout blow from which the Clinton campaign would not be able to recover. That seemed like a forbiddingly ambitious goal, but we would play for the early knockout. Tad’s resolve strengthened at the end of the third quarter, when we came within $2 million of Clinton’s fund-raising total. Given our lower burn rate, driven in part by our decision to not follow Clinton when she went up on TV, we could potentially achieve an overall resource advantage.
Tad’s thinking went this way: The Clinton campaign was relying primarily on larger donors—many of whom had given the federal maximum. They were under pressure to find new donors. We, on the other hand, were relying on many more small donors who had the capacity to give repeatedly. Our gaining a resource advantage would spook her prospective donors. The Clinton campaign could find itself in a money pinch. Winning the first four states could spell the end of their campaign. Achieving a win in the first four seemed very unlikely, but everything that we had accomplished up to that point had been unlikely—at best.
In the event that didn’t happen (and unfortunately it did not), the path would be a long slog to the convention using Mark’s road map. We would still have to do well in the early states. We would have to win New Hampshire and beat expectations in Iowa. We would have to avoid being “killed” in Nevada and South Carolina. Then we would fight for wins every week thereafter when there was a contest.
Mark’s go-the-distance road map also identified key battleground states. Very early on, we knew Michigan would be one. It was relatively early in the calendar (March 8). It was a diverse state, and one that had been devastated by the very trade deals that, over the years, Bernie had opposed and Clinton had supported. If we were doing well at the time of the Michigan primary, it would confirm Bernie’s momentum. If we were not, it would serve as a reset going into delegate-rich March 15, when Missouri, Illinois, Ohio, North Carolina, and Florida held their contests.
During the fall and early winter, we accelerated our staffing so we could compete in the first four and beyond. At the end of July, we had only twenty-nine staffers nationwide. We started adding state directors in many of the eleven states that held primaries and caucuses on March 1.
By February 1, the date of the Iowa caucus, we reached a peak of some 1,200 staff and paid canvassers. It was a herculean effort to recruit, vet, and deploy that many people in only a few months. In terms of vetting, we checked criminal backgrounds, social media usage, and other sources. All campaign staffers have to go through a vetting process, and the vetting of so many candidates bogged down considerably in the fall.
I made a conscious decision that our campaign would practice what it preached in terms of helping people who had had encounters with the criminal justice system reintegrate into civil society. Domestic violence and crimes against children were automatically disqualifying. No one else was automatically disqualified. If anyone’s vetting revealed any other criminal conviction, I personally reviewed the case. In virtually every case, the otherwise qualified applicant was hired. By magnitudes, far more people were not hired because of the content of their social media activity than because of a criminal conviction. On a campaign of our size, it goes without saying that there were a number of problems that arose. However, I am not aware of a single instance in which the hiring of a staffer with a criminal record was the source of any of them.