Photographer: Christopher Leaman/Bloomberg Business

Meet the 21st-Century Political Alchemist Who’s Been Data-Mining for Hillary for the Past Two Years

A deep dive into the modern data-driven campaign with Obama veteran and Ready for Hillary senior strategist Mitch Stewart.

When Mitch Stewart first became a field organizer, in 2002, it looked like the most anachronistic job a young man could seek out in a 21st-century campaign: managing and training volunteers for the work of phone banks and door knocks. At that point, the glamour job in politics was making ads. But now, with big data at the center of modern campaigns, the game is different. Thanks to the availability of information about individuals at a remarkably granular level, along with statistical tools for finding patterns in it, what used to be a matter largely of shoe leather now cranks along with guidance from the most sophisticated analytics anywhere. The 39-year-old Stewart has risen to the highest rank of Democratic operatives by thinking holistically about how those pieces of a modern campaign fit together, and how strategies and budgets need to change accordingly.

A South Dakotan who served as field director for Senator Tom Daschle’s 2004 reelection, in which the minority leader was deposed by John Thune, Stewart became a linchpin of the Obama political apparatus: field director for the 2008 Iowa caucuses; battleground states director for the 2012 reelection. In between, Stewart served as founding director of Organizing for America, the Obama campaign-in-exile at the Democratic National Committee. In 2013, along with fellow Obama field hands Jeremy Bird and Marlon Marshall, Stewart launched 270 Strategies. Arguably the first of the truly new-wave political firms, 270 markets itself to clients inside and outside of electoral politics as a general consultancy equally versed in collecting petition signatures and targeting digital ads.

One of its highest-profile clients has been Ready for Hillary, a super PAC founded in early 2013 as a vessel to nudge Clinton into another presidential race. The group was initially viewed suspiciously by Clinton allies, uncertain what purpose a grassroots freelance effort could serve three-and-a-half years before election day in 2016. But the hiring later that year of Stewart’s 270 Strategies as the super PAC’s chief consultant was a purposeful signal to donors that the group’s aims and methods were serious, and that it was worth an investment. (It ultimately raised $15 million, according to a spokesman.) Stewart’s job assignment was simple: Help build a list of supporters that Clinton’s staff could inherit for their use whenever she became a candidate. 

Also on Bloomberg Politics: Inside the First Super-PAC Dedicated to Collecting Data All About You

When her campaign began to take shape earlier this year, Ready for Hillary started to wind down its operation and prepared to hand off—through a legally delicate exchange between two institutions forbidden from coordinating their efforts—the fruits of its labor. Last weekend, Ready for Hillary quantified them: 2.4 million people who signed a pledge to help Clinton’s candidacy, along with four million voters, or prospective ones, identified as supporters the Clinton campaign should attempt to mobilize.

As that process of unwinding the group was getting underway, Mitch Stewart met Bloomberg Politics’s Sasha Issenberg for an exit interview to reflect on this experiment in non-campaign organizing—what Ready for Hillary did, what it accomplished, and what it leaves behind now that Hillary is officially in.  (The conversations were compiled, edited, and condensed for clarity and readability.)

Photographer: Christopher Leaman/Bloomberg Business
Mitch Stewart in his office in Washington, D.C.

Sasha Issenberg: So what was Ready for Hillary?

Mitch Stewart: It was basically a gigantic net. You saw all this enthusiasm across the country shortly after the 2012 re-elect, where people were excited about a Hillary Clinton presidential run. From my perspective, not taking advantage of that enthusiasm when you had it would have been a huge wasted opportunity. One of the things that we learned in 2012 and in 2008 is that the larger the foundation you have, the bigger your impacts will be.

SI: When you say “foundation,” what do you mean?

MS: Well, for me that’s about data: building as big a list as humanly possible. You can do that through e-mail acquisition or getting people to sign up.

SI: When Ready for Hillary started in early 2013, my impression initially was that it was more of a vehicle for activists to stoke enthusiasm for her rather than corral it in a way that would feed directly into an eventual campaign.

MS: I think something Adam[1] and the team did that was really smart was they went out and engaged the Clinton allies from previous efforts that have a lot of credibility in that world. Seeking “permission” might not be the right word, but certainly their blessing. And they got that. I don’t know if you went to any of their conferences, but you saw people like James Carville and Paul Begala—people pretty high up in that world—participating. For me personally, coming as an outsider, when I saw some of those faces and heard those names, I realized that really smart people, both in the Clinton world and in the Obama world, recognized that there is a benefit to having a foundation set up to support a national effort.

SI: How much did you know when you got involved about about what your goals and exit strategy would be?

MS: The whole premise for the effort was literally in the title of the organization, which is to set the framework or the foundation for an eventual national effort. The tricky part was just to figure out how we can do this in a way that was in accordance with all of the election laws. The whole premise was to build as big a list as possible to then provide the eventual campaign a head start.

SI: It seems as if Ready for Hillary has taken on a specific type of early campaign-season work—engaging core supporters and the like—that usually gets forced by a competitive primary.

MS: Take this all way back to 2007. The very first event we had for Obama in Iowa was in Cedar Rapids. He announced in Springfield that morning, and then his second stop was Cedar Rapids—it was at Kennedy High School, we had about 2,200 people there. You had to fill out a ticket to get in that asked for all your personal information. That’s how you build your caucus operation. I and the five staff that I had for the first probably four weeks would literally spend all day doing data entry. That was like the offline version of what we’re doing at Ready for Hillary: having the stickiest web possible. How we looked at our job is: we just want to capture as much data as we can and then try to manage it at the back-end. That’s sort of what we did online, as opposed to offline.

SI: How is the data you get today from an individual supporter who signs up at Ready for Hillary different from what you would have had from someone requesting a ticket to a rally in 2007?

MS: Actually, the barriers for getting that information have completely flipped. Now if all we get is an e-mail address that still provides value.

SI: Because you can match it back to a name?

MS: Even matching an e-mail address is brutal. How we deal is we try get them to self-select: “Get a free bumper sticker, but to do that you have to give us your address so that you we can mail it to you.” You can figure out different ways to move that single e-mail address into something that’s more identifying, so that we can try to match you back to the voter file. One of the things that we did in Iowa is we had a “Caucus Look-Up Tool,” which is a way we were able capture a ton of personal information, because you had to put in your address. But in 2007, all I gave a shit about really was their personal information, their voting address, and their telephone number. The e-mail address was seen as sort of a nice thing if you can get it, but not necessary. Now the e-mail address is absolutely crucial for building your list.

SI: How valuable is an e-mail address if you can’t link it to an identity in the real world?

MS: That’s going to be a voter. You can still see if they’re opening your e-mail. You can still see if they’re taking action. You know there’s a person behind it. And in a perfect world you’d be able to match that back to the voter file to know exactly how to treat them. Are they a voter-registration target? Are they a persuasion target? Are they a get out the vote (GOTV) target? You can’t quantify the impact, necessarily, that you’re having with an individual, but you know that you’re having an impact.

SI: If I gave you a bunch of e-mail addresses and nothing else for people who are politically active, how many could you match to real-world identities?

MS: Maybe you get twenty percent? Then every time you get a field, that match rate goes up. Beyond e-mail address, hopefully you can get first, last, and zip code—then you’re able to do probably forty or fifty percent. I think you can get up to seventy or eighty percent if you get, like, full information.

SI: But whenever you set up another screen that someone’s supposed to go through to give more of that information—

MS: You get a huge drop-off.

SI: How do you weigh that, between trying to get more data or more people?

MS: So let me ask you that question. If you’re running a campaign or you’re running an organization, what’s more important to you: one hundred e-mail addresses, or ten fully-laid out profiles, with details about the individual including the e-mail address? What would you rather have?

SI: I’d respond, in the rabbinic spirit: It depends what type of campaign you’re running.

MS: Sure. But ultimately that’s the Sophie’s choice you have.

SI: Thanks for sticking with the Jewish stuff. Know your audience.

MS: I say it’s better to start with a bigger foundation than a more refined one unless you have a really specific reason, like if you’re doing an academic exercise.

SI: How did you define your objectives here? You weren’t really working towards vote goals the way you would on a campaign.

MS:  One of the things we did at Ready for Hillary, and I think did it well, was being very strategic—about the impact that we were having immediately, but also being mindful of what sort of payoff will have on a potential run down the road. We didn’t want this just to be about Ready for Hillary, this was to supposed to—and did, I think—lay groundwork for a national campaign down the road.

SI: What did you learn in 2011 and the first part of 2012 about trying to recruit and motivate volunteers in the absence of a real opponent?

MS: It’s not easy. If you remember the story lines going into 2011, it was that the enthusiasm’s way down and there’s no way we will be able to replicate the same sort of machine that we had in 2008. There were a lot of reasons to be pessimistic. So we start reaching out to volunteers—the first program we ran was Im In, basically trying to engage. We go to everybody who was active in 2008 and get them to take the pledge: Im in. We know that matters. And so we spend basically the first six months running that program, and it wasn’t very easy. As soon as the Republicans started to have debates, though, and it became crystal clear what the contrast was—even if it wasn’t Romney at that point—it became easier to recruit volunteers.

SI: The early days of the Obama reelection looked to those of us on the outside as though you were running a pyramid scheme: You were recruiting volunteers so they could recruit other volunteers.

MS:  There were phases in the campaign, and so it’s all about capacity-building at that point. In 2012 there ended up being 150 million door knocks and phone calls. If we had waited to start building that infrastructure, say in 2012 as opposed to 2011, maybe we would have hit 90 million, or 120.

SI:  Is it ever too early?

MS:  No, I don’t think so. I mean, at least not right now. I don’t believe that “people are sick of campaigning.” There are ways you can run a campaign that are not in someone’s face. There are millions and millions of people right now who want to support Secretary Clinton, and the earlier you can get those folks involved, trained, and then out by themselves engaging people, the bigger your operation will be.

SI: So Ready for Hillary is collecting e-mail addresses. What happens then?

MS: People are signing up saying Im ready for Hillary, Im in. People start by giving low-dollar donations, then we start walking them up that ladder of engagement—from an online perspective, because we don’t have an offline perspective. Now we’ve done that, let’s start taking you down different paths digitally to see how you want to be involved. 

SI: Can you give me a rough sense of the fractions we’re talking about? What share of the e-mail addresses that you collected have already given or volunteered?

MS: This is not particular to Ready for Hillary, just a test case for another organization that we’re dealing with. What you want for an issue advocacy or non-profit entity, the average is about a 13 percent open rate. What you hope for is then about a 3 percent click rate for the entire list, and then there’s a smaller subset that’ll take action. So it’s almost like you know there’s divisions of 10 there. I’m not saying that applies for Ready for Hillary, but that gives you a sense of the scope.

SI: What are your assumptions about the people who aren’t responding? You’ve contacted them over and over again, we’re getting closer to the 2016 campaign, we’re starting to see who an opponent might be. If they haven’t taken an action now, can you conclude anything about their value as a potential volunteer or donor to a presidential campaign in 2016?

MS: Well, I think the short answer is no. But I would not give up on these folks now. In fact I’d probably run a really aggressive campaign, as soon as the campaign launches, to try to get them to do something. Once there’s news, that’s an organizing opportunity. As far as dissecting or segmenting the list, how I would do it is: first you have your foundation, which is basically your list of supporters, primarily e-mail. Then from there you’re building up. So you know you probably have your list of donors from that and then you have your list of people who volunteer, and people who do both—and so the bigger this foundation is, the bigger everything else will be on top of it

SI: Did Ready for Hillary have field offices?

MS:  No.

SI: Why not?

MS:  It’s a pretty big investment. If you’re talking about having, you know, roughly 20 people on payroll, maybe a little bit more than that, you’re going to have to do some analysis, like where does your dollar stretch the furthest. Rather than field offices, I think the smarter dollar usage would be trying to have strategic impacts within states and within constituencies that we know will be impactful in 2016. So, trying to identify and recruit Latinos, college people, women, African-Americans, the LGBT community.

SI: So what did those twenty-odd people do?

MS: Well, it’s a combination. We hired people specifically for the act of building relationships, and organizations in swing states. Eventually, as 2014 came closer and closer, we distributed a lot of them out to be helpful to people running. 

SI: As a volunteer base for a presidential campaign?

MS: That is correct. You’ve given that future entity hopefully a bit of a head start on how they can start organizing crucial constituencies within those states.

SI: And why those groups in particular?

MS: Because we know if you look at the orchestra that made up the winning majority in a number of battleground states. You want to start with those constituencies, build relationships that then could transfer to a potential campaign.

SI: How do you anticipate what the campaign’s needs will be in each of these areas?

MS: I guess I’d take it back to 2008, 2012. So in 2008 we had 1.2 million volunteers. By all accounts, that broke every record in the book. It would have been impossible for us to try to duplicate. Except in 2012, there were 2.2 million volunteers! From my perspective, that’s additive, and so I suspect that when 2016 rolls around, because the foundation is bigger now than it was in 2008—or even in 2012—that more people will be ready to take action. The volunteer army will be bigger in 2016 than it was in even in 2012.

SI: Why do you think that there’ll be an increase?

MS: It’s both about identifying the low-hanging fruit—by that I mean people who are just interested in supporting a presidential campaign and being involved—and about the skills that those volunteers learn as they move up the ladder of engagement.

SI: At some point do you reach a ceiling? Like there are just so many people who have lefty views and free time and a willingness knock on stranger’s doors and you get all of them. The LDS church may have converted everyone who is susceptible to becoming a Mormon, right?

MS: Fantastic organizing, although I don’t support the religion! What we found to be the limiting factor in this is not the raw number of volunteer prospects, it’s the number of neighborhood team leaders, and by that, the number of field organizers. And so there’s a direct relation between a number of field organizers you have and the number of neighborhood team leaders those field organizers can recruit and then the impact that each of those teams have.

SI: And then the amount of time that they’re given, right?

MS: Right. So you know I can see a scenario where there’s maybe even less enthusiasm in 2016 than there was in 2012, but—because we’re starting at such a more elevated place than we did in 2011 or 2007—the volunteer number will be bigger in 2016 than it was in 2012. You’ll have better-trained volunteers, they know what they’re doing and can create this replicable, scalable volunteer organizations and structures. 

SI: I assume that even if there are as many volunteers, they will have a very different demographic profile than Obama’s.

MS: We looked at that “average” or “typical” Obama volunteer, and it was an older white woman.

SI: How will the next Democratic nominee use volunteers differently than Obama did in 2012?

MS: Now the Analyst Institute[2] is saying it’s not about the quantity, it’s the quality of those conversations. What I always say is if you can’t get your message through in 90 seconds with a voter, move on. Now they’re saying no, have a 10 to 15 minute conversation with that voter—it will be more impactful. Am I going to reject that out of hand based on of my experience? Or am I open to saying: Hey, you know what? Instead of, like, having a really short conversation that’s not very deep or meaningful, sit down and really try to engage them, spend 15 minutes, because we know at the end of the day that one conversation is worth five two-minute conversations. That’s the math that you have to do. So I don’t know.

SI: And so what are your feelings right now about the ability of volunteers to deliver the campaign’s message?

MS: I think it’s the best way to persuade.

SI: How do you assess risks involved in that? If you have two million Democratic volunteers next year, do you want all of them knocking on doors and talking about funding Planned Parenthood?

MS: This is something else that we tested[3]. We had 120,000 people, and did a field experiment: How did they react afterwards? You learn a lot of interesting things. With some groups of people you can have 10 percent impact—so you know if you do 100 persuasion phone calls or door knocks, ten more people will now support them than before. We also found there’s a negative corollary. We knew that if we talk to 100 people, we lost two or three votes. 

SI: Do you feel good about the ability to use modeling to preemptively identify people likely to be alienated by the interaction?

MS: We do now! The saying on the campaign is “don’t poke the badger.” But ultimately you can know who the badger is now, so you don’t engage them.

SI: Do you look back on campaigns that you’ve been on and recognize things you did wrong?

MS: 2004 still makes me sick to my stomach. I had the most sophisticated data-coding system. I had 150 staffers in South Dakota. We spent $20 million dollars, John Thune spent $20 million dollars. I think about that race every day.

SI: You wish you would have allocated resources differently?

MS: Yes. I would have invested way more in volunteer recruitment and retention. I didn’t do that as much—and a lot of people didn’t in 2004. I would have been much more selective in the messages that we used. I would have had a much larger conversation about our direct mail program. Our GOTV operation would have been radically different. I don’t know if we could have overcome four or five thousand votes in a pretty small state, but I would have been a much better advocate for Tom Daschle if I would have known then what I know now.

SI: If you could play God, how would you divvy up the resources and responsibilities between all the different entities trying to elect a Democratic president in 2016?

MS: The national committee, the DNC, has to remain the center point for all data collection and manipulation of that data. So the voter file needs to continue to live at the DNC, it needs to be the hub of all the different spokes that run Democratic campaigns and then feed back into the DNC. Some of the technology that enhances that data, and allows this stuff to flow freer—whether it’s between a volunteer data set, a donor data set, a voter data set—I think that needs to live in the DNC. That’s a huge, critical piece. With the actual candidate campaign committee, that has to be both distinguishing your brand and your opponent’s brand. And I think the field component has to live within that candidate campaign.

SI: Why is that?

MS: I hope that with two successful presidential efforts we’ve broken this, but the idea that volunteers are going to show up because the party asked them to is an outdated model. Volunteers are going to show up because a candidate asked them to, because a candidate builds a specific relationship with those volunteers—motivates them, keeps them engaged. Volunteers showing up in 2015 and 2016 need to know that they’re volunteering for Hillary Clinton. Hillary Clinton needs to know that those volunteers are her people, and that that line is very, very straight between two dots—that you’re not having to go through two or three other subsidiary legal entities to show up for your preferred presidential candidate. 

To me, the relationship that President Obama and the volunteers had, it can be replicated—it can be done for people other than Barack Obama—but doing the John Kerry model in 2004, with ACT[4] being the “super PAC” field organization, is just a waste of money. I would never support something like that.

SI: So where do super PACs fit into this ecosystem?

MS: What Priorities[5] did very well in 2012 was really labeled the opponent in ways that didn’t make sense for the campaign to—went a step further than the campaign felt comfortable branding Mitt Romney as someone who outsourced jobs. So the negative campaigning is where super PACs will continue to spend most of their resources.

But you could see a world where a super PAC is built solely on providing a presidential campaign with technology. You could make a pretty strong case it makes sense to have a continual environment where smart developers can build apps that help campaigns make smarter decisions about how they spend their resources, that can make door-to-door apps, all of those things—but it doesn’t live with the committee, it lives with the super PAC that a campaign or committee would have access to. You can see a world where that exists.

SI: So let’s talk about something I assume super PACs will never do on their own: determine electoral-college strategy for a nominee. You managed the map for Obama in 2012. How early can you start identifying your battleground states?

MS: It’s kind of tough initially when you don’t have the candidates. When we started in 2011, there were six or seven, but you couldn’t really say, based on an Obama vs. Romney poll eighteen months beforehand, what’s a battleground state and what isn’t.

SI: What did that process look like?

MS: Basically the formula we went through was some baseline research, some baseline polling. The president’s positives and negatives, fave/unfave would help inform it, but you’d also look at previous election results, both at the presidential level, and incorporating 2010. Based on that, you’d assign a value to different cohorts—there’s a little art to this.

So, say, African-American women—we expect the president to perform at 94 percent, because if you look at 2008 he got 97, if you look at 2010, the Democratic candidate got 89 percent. Then you just look at the changing demographics: instead of 100,000 African-American women voters in 2008, now there are 112,000. Basically, you go through that line by line, and then you predict how your candidate would perform off a generic Republican candidate. We looked at some places initially, and Florida was just brutal. Colorado was brutal.

SI: “Brutal” means you’re calculating the resources you would have to put in?

MS: That’s exactly right. There are three ways to win an election, or build your list. You can do that through voter registration, you can do that through persuading undecided voters, or you can do it through increasing turnout. With the analytics team, we would go through the math exercise I just talked about and we would say, “right now, if the election were held, we would be at 48.6 percent”—you could be that granular. We say “okay, we have a goal of 51 percent, how do we get there?,” with those three mechanisms I mentioned. We had bar graphs, and most of the states were pretty reasonable to get to 51 percent, but Florida looked like the Sears Tower. The resources we would have to spend when we initially looked at Florida—astronomical amounts of money to win that state. That’s, frankly, why it became the 322nd electoral vote[6] and not the 270th. It was never where Ohio was, for example.

SI: But you always had to compete in Florida, right? Was there a price where you would have said Florida will not be a battleground state?

MS: It’s a two-headed question. You have to ask about the resources you have to spend to win Florida. But another way to look at it would be: What are the resources that the Romney campaign would have to spend to win Florida? The hill you have to climb to get to 270—Romney’s doing the inverse, coming up the backside to get to 270. So before they get to Ohio they have to win Florida. If that means we spend $40 million in Florida, but they have to spend $120 million, and we still lose it—it means that’s $120 million they can’t spend in Ohio, Colorado, Virginia, some of these other states that will be more determinative of who wins and who loses.

So it really is a chess game on resources. North Carolina is a great example. Everyone was like, “oh, you lost North Carolina!” It was never going to be our 270th electoral vote. But the Romney campaign, or the Republicans, have to spend a ton of money to win North Carolina? That means there’s less money they can spend in Florida, there’s less money they can spend in Virginia, all these other states that are more likely to be the 270th electoral vote. That’s really a battle of resources at that point, and attrition—that’s the math we went through.

There’s also some political optics. Not playing in Florida would have had a negative impact in states outside of Florida.

SI: Granted that we don’t know now who the Republican nominee will be, where will the map of the 2016 Democratic nominee start differs from that 2012 map? 

MS: I think she’ll have a very similar map to what we had in 2012. None of the states that he was competitive in, at least in my estimation, would come off. I think the states that were hard for the president would be hard for her. I think she starts in a very strong position, because the fact that we were competing in North Carolina, and the fact that Florida was never really in the picture of being the 270th electoral vote, means that we’re playing on their side of the 50-yard line, and they’re not playing on ours.

In 2000 and 2004, we were playing in Washington and Oregon …

SI: Minnesota!

MS: That’s a great example. In those elections, we were playing on our side of the 50-yard line. That’s changed. There has been a shift in the electoral college in our favor, and it’s just about maintaining that. I don’t know if there’ll be a huge expansion from what the president did, but I don’t think she’d need one. We’re talking about a pretty friendly map for Democrats. Now, it’s not destiny. A lot of things have to happen, I think it’s still going to be a close race, but the map favors Democrats.

SI: Do you expect that the identity of the Republican nominee would change that? Is your instinct that any of their candidates could expand the map or find new places to put the Democrat on defense?

MS: I guess the way to answer that is: If either Jeb or Rubio is the nominee, does that mean Florida behaves differently in a presidential than it would otherwise? I’m not convinced it would. I may be the outlier on that, but I don’t think Jeb Bush being the nominee takes Florida off the table.

SI: You could, though, end up in a situation where the Democrat ends up having to spend more there just to keep it competitive.

MS: For sure. But I don’t think it completely throws the map up in the air.

SI: How do you assess the advantages and disadvantages for a non-incumbent who wins the nomination without a difficult primary contest?

MS: I can see both the pros and cons. Coming out of 2008, I thought that primary process—especially by Indiana, it was a long, long, long fight—did two things. It allowed states to shift: You look at the voter registrations in Pennsylvania, in North Carolina, in Indiana after that primary election, and they’re radically different than they were. So that’s a benefit. The second benefit is potentially testing the candidate through ways that can only happen through a contested fight—seeing how they deal with adversity, attention, drama and all those things.

But you’ll remember there are a whole bunch of party stalwarts at that point saying how detrimental this primary is, and how these relationships will never be restored. So you had a lot of people saying that long primary fight was not helpful. 

SI: That never really seemed to be borne out, right? In retrospect, there seems to be little debate that it was better for Obama as the nominee.

MS: I kind of go back and forth on it. If it’s an easier one, maybe you don’t have a lot of the dissension that folks thought we would have in 2008. What I do know is the Clinton campaign isn’t taking anything for granted—they’ll be fighting for primary votes and caucus attendees block by block, neighborhood by neighborhood, just as they would if it were 2007, only this time with different results.

SI: Could the Ready for Hillary structure have been built and executed for anyone other than Hillary Clinton?

MS: No. The folks who support Elizabeth Warren are trying to do something similar, and I haven’t seen their numbers or how successful they’ve been at building a list. But with some of the other candidates, there isn’t the enthusiasm to support it like there is with Hillary Clinton. To build a list of over a million names, I don’t know who else could have done that. That’s organic support, that’s not buying a million names from other progressive organizations—that’s folks self-selecting It’s a pretty unique phenomenon. As future candidates look at the formula for doing something like this, I think you’ll see their activists copy this. Not all of them, but some of them will.


[1] Adam Parkhomenko is a former Clinton aide who started Ready for Clinton and has served as its executive director. He is expected to take a job on the presidential campaign.

[2] The Analyst Institute is a secret society of lefty campaign operatives and social scientists founded in 2007 to administer randomized-control trials around electioneering tactics. Their empirically tested “best practices” have challenged much of the conventional wisdom about how to most effectively use volunteers.

[3] In 2012, Obama for America contracted with the Analyst Institute to embed experiments in various aspects of the campaign’s operations to measure their real-world impact.

[4] America Coming Together was an independent group founded in 2004 to help the Kerry-Edwards ticket, and funded primarily by labor unions and large lefty donors including George Soros. ACT was structured under section 527 of the federal tax code, which allowed it to register and mobilize voters in battleground states on Kerry’s behalf while the campaign focussed on persuasion. Since the 2010 Citizens United decision loosening campaign-finance laws, 527s have been largely superseded by super PACs, which are permitted to engage in direct advocacy on behalf of candidates.

[5] Through a well-funded media campaign, the pro-Obama super PAC Priorities USA effectively introduced the issue of Bain Capital into national politics during the summer of 2012. In one particularly harsh ad, the group directly blamed Mitt Romney’s work with the private-equity firm for a former employee’s illness and death. Since 2012, Priorities has been repurposed as a pro-Clinton super PAC and is expected to deliver much of the outside advertising on Hillary’s behalf.

[6] Obama won 322 electoral votes to Romney’s 206. Florida was the closest state in 2012, as it was famously in 2000—when a Democratic victory would have tipped Al Gore over 270 votes instead of merely running up the score, as it did for Obama.