by Sean Spicer
We had to upshift the downshifters.
We also followed up with a strong ground game. The RNC had added more than 172,000 voters to the rolls than the Democrats had in the battleground states. To make the most of this advantage, our field staff had become a large and capable army. In Colorado in 2012, we had forty-six field staff members deployed. In 2016, it was 933. In Michigan, we had thirty-three people in 2012. In 2016, it was 778. Our volunteers had knocked on the doors of 14.5 million likely Trump voters nationwide against 11.5 million for the entire 2012 cycle.
We made almost 26 million phone calls to voters based on big data.
Trump himself had never wavered in his belief that Pennsylvania was gettable. The RNC political staff saw hidden opportunities in Wisconsin and Michigan. What we saw in the aftermath of the second debate and the RNC follow-up was a strong, upward trend. Michigan provided a clear illustration of the national trend.
In September, the RNC model depicted Trump with 260 electoral votes (Wisconsin was tied and Trump was even closing in on blue Minnesota and Colorado). After the release of the Access Hollywood video in early October, our model showed us dropping nearly 100 electoral votes as potential Trump voters “downshifted” into possibly sitting out the election in response to the saturation of negative media coverage.
The “upshifting” of this group occurred after the RNC reached these discouraged voters via social media (Brad Parscale and his team of digital ninjas) combined with old-fashioned, door-to-door politicking (under the direction of Chris Carr and his talented staff in the field). We targeted these voters one by one. We reminded them how important the election was, with big issues at stake, including the future direction of the Supreme Court, the economy, and issues of particular interest to certain voters. Then we confronted them with a binary choice between a blemished candidate (Donald Trump) and an unacceptable candidate (Hillary Clinton). That started moving the trend lines back in our direction in the battleground states. We deployed Donald Trump, Mike Pence, and our top surrogate speakers wherever the data told us we had the greatest chance of shifting voters to win a state.
In Florida, for example, our model showed us that 6 percent of the state’s voters—consisting of 680,000 possible votes—were undecided. Exit polls later confirmed that this undecided 6 percent remained uncommitted until just before election day. Our data showed that most of these undecideds were likely Trump voters, and they were gettable—but we had to reach them.
Ohio, Iowa, and other battleground states were all showing the same upward trend in our direction. We felt that if one major battleground state fell our way, the trend was such that the others would as well because the demographics of the “downshifted” voters in the battleground states were similar—and they were all moving our way.
State after state, we reached our vote goal, and the Democrats did not. The media had completely bought into the idea that no one could match the Clinton machine’s ground game. But, in fact, the Democrats underperformed in many places with early voting and absentee ballots.
Our model suggested a Trump win was possible, but the path to victory was narrow.
Electing a president and a congressional majority was our overriding goal, but we also had another institutional stake in the outcome. Reince had bet the farm that if we could build a superior data operation, our candidates up and down the ballot would have an advantage over their Democratic opponents. Donors had bought into the plan. We had to prove that the RNC had provided the right technology and necessary resources and that our data had been accurate and valuable. So, we created a digital presentation that showed not only the investment the RNC had made but also what the data revealed heading into the final week of the campaign, state by state. The presentation laid out the vote goals Trump needed to meet in order to win and our predictions (assembled by Bill Skelly and his team) of what each state’s overall vote totals would be.
Katie Walsh, Chris Carr, and I briefed key political players, pundits, and the media. The data showed where we stood in every battleground state. More importantly, it showed how many votes we needed and where they were. Win or lose, no one could say that the RNC had not done its work, and we were increasingly confident that the trend was moving our way and that Donald Trump had a shot at becoming the forty-fifth president of the United States—even if, at the time, no one in the media seemed to believe us.
In retrospect, one of the most interesting aspects of this race was Donald Trump’s rallies. Until then, political professionals considered rallies to be in the same vein as yard signs. Having a lot of them doesn’t mean you’re going to win. It just means you’ve dedicated your resources to cultivating your committed supporters.
But there was something different about the Trump rallies. A good example is the last rally Trump held before the general election. It was at the DeVos Place in Grand Rapids, Michigan. He easily filled that giant venue with a line of supporters stretching across the Gillett Bridge to the Gerald R. Ford Presidential Museum on the other side of the Grand River. To get in at 8:00 p.m., some people took their place in line at 9:30 that morning.
“If we win Michigan, we will win this historic election,” Trump announced to wild applause.5
We had never seen anything like it in terms of energy, enthusiasm, and dedicated support (enough to sustain hours of line standing). Those intangibles count. They cannot be manufactured. They cannot be bought. And there was one more factor: there is an old saying in campaign politics that the candidate who’s having the most fun wins.
Donald Trump was always having fun—especially at his rallies where he reveled in the excitement and adoration of his supporters and shared their enthusiasm. At his rallies, he was a rock star. The rally field staff would work the lines to sign up volunteers and register voters. In the past, the campaign would have to spend money to advertise a rally and drive people to attend. In Trump world, the cost was a mere tweet from the candidate.
On the morning of election day, I checked my phone and saw that Trump, at age seventy, was still an Energizer Bunny. He had campaigned into the late hours and was up again in the early morning.
At that point, the campaign had one remaining mission: to get on television and explain why we were going to win. Cable news outlets need to fill election day air time before the polls close, so it’s another way to play get-out-the-vote. Confidence on election day is critical. The early misreporting about Bush’s loss in Florida in 2000 almost sank him in the rest of the country as discouraged supporters avoided going to the polls. We couldn’t let that happen again.
We were close to winning in ten states, not where we wanted to be, but well within the margin of error. We were convinced the polls underreported Trump’s strength among voters who were shy about announcing their support for a candidate excoriated by the media. The determining factors were whether our positive trend lines would continue in these final, critical hours and whether undecided but potentially pro-Trump voters turned out to vote.
I did a quick interview on MSNBC with Kate Snow, explaining some of what we saw in our internal polling data. While at the studio I tracked down Mika Brzezinski and Joe Scarborough in a small green room down the hall from their set. My assistant, Vanessa Morrone, and I walked them through our state-by-state polling data on an iPad that showed how we could hit the magic number of 270 electoral votes. They remained skeptical, but after the election, Joe Scarborough noted on the air multiple times that I had, in fact, shown him how Trump could win. That was notable not only because most of the media refused to believe our data but also because some outlets actually reported the opposite. Politico, for one, declared, “RNC Model Showed Trump Losing,” which was absolutely not the case.
Next, I went to the Fox News studio where I did a short interview on Fox Business. When the interview was done, Bill Shine, then co-president of Fox News, entered the studio and asked if I had a couple of minutes.
We rode the elevator up to Bill’s office on the executive floor.
> Waiting for us in office chairs were Rupert Murdoch; Fox executive Suzanne Scott (now president of programming at Fox News); Jay Wallace, another Fox executive; Brian Jones, head of Fox Business; and several other executives. I had known Bill and some of the others, but this was my first conversation with Rupert Murdoch beyond an exchange of pleasantries at charity functions.
“So, Sean, can you give us the lay of the land?” Murdoch asked in his deep, gravelly, Australian accent.
I pulled out the iPad and explained what our internal data showed and how we were poised to exploit the trend. Murdoch leaned over the screen, asking informed questions about each swing state and why we thought we would win it. As I spoke about swing counties and key demographics, I could see the wheels spinning behind Rupert Murdoch’s eyes. He was doing the math himself and arriving at the same number we did—and it would likely be one larger than 270.
I kept running the math through my head, playing with electoral maps, and by my guess, we could get to 270, 271, or 272, including Maine’s second congressional district.
But not everyone in the media was interested in what the data said. Just about every news host, reporter, pundit, and political analyst thought we were going to lose.
At 7:30 p.m. on election day, a bunch of us—including Steve Bannon, Kellyanne Conway, Dan Scavino, Hope Hicks, Bill Stepien, Mike Pence, Reince Priebus, Jared Kushner, and Ivanka Trump—gathered on the fifth floor of Trump Tower. We were in a large utility room with a projector feeding us RNC and Associated Press election data on county-by-county vote counts. From the outset, Donald Trump had a laser focus on winning the electoral college, and we all knew that winning swing states was largely a matter of winning swing counties, which is why those county vote totals were so important—they were the best data we had on voter trends within a state.
The exit polls coming in were not good; but if exit polls selected presidents, Al Gore would have been in the White House. We didn’t let the exit polls rattle us. More important were voter percentages within counties. If there was big turnout in Trump-leaning counties, that was an indication that the undecided voters we needed to win were breaking our way.
How was Trump faring in the Florida Panhandle compared to Romney and Bush? How were we doing in Broward County, Florida? How was the turnout for Hillary in Fairfax County, Virginia? Where were we outperforming? Where were we underperforming?
While we fielded one nervous question after another, Ivanka talked on the phone with her father. “You’ve got to come down to five,” she said. “We’re all here.”
“No, have everyone come up to fourteen.”
Bill Stepien, political director of the Trump campaign, packed up the projector and took it upstairs. Bill’s computer was plugged into the projector, and the data was shown against a wall just off the stairway that connected the fifth floor to the fourteenth. Around 10:00 p.m., the outlines of a trend were starting to become clear. We did, indeed, have the wind at our backs. One by one, key states fell into the Republican-red column. Indiana, then Kentucky, West Virginia, South Carolina, Alabama, and a host of red states from Tennessee to Texas—that was all well and good but not enough. No surprises yet, but at least we had reason to be cautiously optimistic.
Just after 10:00 p.m., I sent Rebecca a text to tell her I was about to go live on ABC with George Stephanopoulos. She had just left an election return party at the Washington Post headquarters and walked across the street to have a drink with some colleagues. When she received my text, she immediately asked the bartender to change the TVs to ABC. She knew I would only be going on the air at that hour if we were seeing some encouraging returns.
Then, at 10:21 p.m., Ohio and its eighteen electoral votes were called for Trump. Better than expected returns in south Florida confirmed that we were in the running there, too. Then all of Florida came in, along with North Carolina, awarding us a combined total of forty-four electoral votes. Both states were must wins for us—and we had gotten them. We were also outperforming in states the media had predicted we would lose.
Our little crowd, which grew bigger by the minute, went nuts with cheers and high fives every time a state was called. Dana Bash on CNN said that fingers must be bleeding in Brooklyn “because there’s no more nail to bite.”6 Our candidate, standing right behind me, was hardly biting his nails. Donald Trump was encouraged, but he was not about to tempt the gods by declaring a premature victory, just like when I called him “president-elect,” and he replied, “not yet.”
The Associated Press called Wisconsin in our favor but not Pennsylvania. Fox called Pennsylvania for us but not Wisconsin.
If we won any of the remaining battleground states, we would win the race. It felt like we had scratched off five winning numbers in the lottery—and were now starting to scratch off the sixth.
Just after 11:15 p.m., Trump went up to his residence to gather his thoughts. By 1:00 a.m., we led in the electoral college 244 to 215. Around 2:30 a.m., Trump headed over to the New York Midtown Hilton, just a few blocks away, as it looked likely that we would win Wisconsin and Pennsylvania, which would put us over the top. Julie Pace of the Associated Press texted Jason Miller that the AP was calling the race for Trump. Jason immediately shared the news with our candidate. Within seconds, Clinton aide Huma Abedin called Kellyanne.
The two aides connected their principals.7 Hillary conceded.
The ballroom was electric. The crowd, warmed up by Vice President-elect Mike Pence, shouted, “USA. USA,” with an enthusiasm I have never seen.
At 2:47 a.m., President-elect Donald Trump, his family, and top aides emerged from a balcony and descended to the stage.
At 2:49 a.m., he said, “Sorry to keep you waiting. Complicated business.”
Melania stood behind him, beaming. Donald Trump made some kind remarks about Hillary Clinton and called on Republicans and Democrats to “heal the wounds that divide us.”
The Trump I saw that night was elated but serious. He clearly appreciated the magnitude of what had just happened.
Shortly after 3:00 a.m., Wisconsin and Pennsylvania were called for Trump, their combined thirty electoral votes putting us over the 270 threshold. Michigan came in later that morning with another sixteen electoral votes, final confirmation that the Trump strategy of reaching out to blue-collar conservatives—people ignored by many Republicans, defined as deplorables by Hillary Clinton, and alienated by the elites—was the right one.
At 3:29 a.m., Trump’s motorcade pulled up to the residence side of Trump Tower. As I stood there welcoming him back, I was somewhat surprised at the light police presence at the home and office of the president-elect of the United States. It wasn’t until just after dawn that barricades started to go up.
Donald Trump won a historic victory that night. But the press immediately tried to deprive him of it by pointing out that he lost the popular vote and that his electoral college margin of victory was smaller than Obama’s in 2012 and 2008. It was historic not only because the media had all but declared Hillary the victor before the event but also because virtually no one believed that Trump could win 306 electoral votes. When you win big after nearly being counted out, it is a victory like no other.
The media had invested their credibility in Hillary winning, and now they were proved utterly wrong. The New York Times had, after all, given Donald Trump at best a 15 percent chance of winning.
Many in the media rarely take any blame for their mistakes. Instead, they blame others. In this case, they blamed the polls, Nate Silver, Russian meddling, anything and everything but their own reporting and analysis. In their race to keep up with the twenty-four-hour news cycle, many reporters now spend less and less time trying to understand what’s happening and more and more time pontificating—ill-informed pontificating at that. We had the data showing that we could win, and it had been dismissed.
In the morning, I stopped at a newsstand to buy a Powerball ticket.
“Let’s see how long this lasts,” I said to the perpl
exed man behind the counter.
It felt like luck, but our victory the night before had been anything but luck. A lot of work, strategy, and research was behind this election.
Later the next day, I took the train back to Washington and the RNC where we all pored over the results. In almost every state, we nailed our total vote goal—what we needed “to finish the house”—and Hillary underperformed.
In Ohio, RNC data had estimated Donald Trump needed 2.8 million votes to win the Buckeye state, and he got 2.84 million votes.
Our model showed Donald Trump winning Michigan by 0.2 percent. He won by 0.3 percent.
Our model also nailed the vote totals.
In Florida, our model projected that 9,409,777 people would vote. Actual voter turnout in that state was 9,420,039. In other words, the RNC had projected turnout to within 0.1 percent.
Ironically, the two battleground states that did not fully embrace RNC data and field operation were Nevada and New Hampshire, where both incumbent Republican U.S. senators lost, and we lost the presidential vote as well.
The RNC projected a national total vote of 135,824,656. The actual number was 136,401,627. We had predicted the total turnout to within four-tenths of 1 percent.
But more than anything else, we had won. We accomplished what we had set out to do—win the White House and maintain control of both the House and Senate.
The big questions now were what the new administration could achieve and what my role, if any, might be.
CHAPTER SEVEN
BAPTISM BY FIRE
Over Thanksgiving, our family gathered at my mom and dad’s house in Barrington, Rhode Island. While we never openly talked about it (welcome to an Irish Catholic family), we were all thinking the same thing: these were my father’s final days. The four young grandchildren weren’t fully aware of what was happening, but they knew it was a special, somber time when my dad said goodbye to each one of them. It was the most difficult holiday of my life. My father was unable to eat solid food. While we sat as a family in the dining room eating takeout turkey, my dad was upstairs in his bed fading in and out, taking nourishment through a tube twice a day and wanting what he couldn’t have—to sit down to Thanksgiving dinner with his family. It was painful to watch him fade. We had stayed close over the years, and I had often called him to share stories and debate the news. That was all coming to an end.