In September, at a resort hotel in the Coachella Valley, the California Republican Party held its fall convention. Brad Parscale—forty-four, six feet eight, balding, prolifically bearded—walked onstage in shirtsleeves and tilted the microphone upward, mumbling a self-deprecating joke about being “awkwardly tall.” Parscale has lived in a red county in California and a blue county in Texas, and he now splits his time between Washington, D.C., and two luxury properties in South Florida, yet he still speaks with the neutral accent of Topeka, Kansas, where he grew up. He was one of the top staffers on Donald Trump’s 2016 campaign. “I was the digital-media director,” he said. “So, yes, all that crazy Facebook stuff was my idea.” Other former Trump-campaign officials fill their calendars with paid speaking gigs, padding their remarks with jingoistic platitudes or rapturous accounts of Trump’s improbable victory. Parscale appears in public less often. When he does, he gets to the point.
“We have turned the R.N.C. into one of the largest data-gathering operations in United States history,” he said. He was referring to the Republican National Committee, which has raised two hundred and sixty-three million dollars for the 2020 elections. (The Democratic National Committee has raised just over a hundred million.) As Parscale explained, the Trump campaign has been operating more or less full time since 2016, continually improving its “technology and data operations.” During this period, the campaign and the R.N.C. have essentially merged, sharing staff, voter data, and other resources. The Democrats do not yet have a nominee for President, and some of their systems for acquiring and sharing data are considered outdated by comparison. “You cannot just build an app, or build out data, in the few months you have from the Convention,” Parscale said. “The Democrats will have that problem this time. As they all interfight, we are building for our future.” Two years ago, Parscale was named the manager of Trump’s 2020 campaign. “I know everybody wants me to do it from my laptop,” he joked to the audience. “Not possible. I’ve already done that once.”
Before Parscale worked for the campaign, he was a digital marketer in San Antonio with no political experience. Referring to his work for Trump in 2016, he has said, “I was thrown into the Super Bowl, never played a game, and won.” But it might be more apt to compare Parscale to the technicians who operated Watson, the I.B.M. supercomputer, while it successfully competed against two humans on “Jeopardy!” Machine learning and social-media algorithms are upending most aspects of contemporary life, including politics. One of Parscale’s advantages was that he recognized this fact and didn’t hesitate to make full use of it.
In previous elections, Presidential campaign managers tried to be strategic about where to hold public events, which slogans to emphasize in which media markets, when to give an interview to Elle or to Esquire. These were forms of targeting. We are now in the era of microtargeting, which began, arguably, in 2012—the year of Facebook’s I.P.O., then the largest in Silicon Valley history—and will continue, inarguably, long past 2020. It’s no longer good enough to run one radio ad in Scranton and another one in Pittsburgh. These days, campaigns can carve the electorate into creepily thin segments: Gold Star moms near military bases, paintball-playing widowers in the Florida Panhandle, recovering addicts in Michigan’s Upper Peninsula. And, for anyone who wants to reach a specific audience with an actionable message, there has never been a platform as potent as Facebook. No matter how many bad press cycles or localized boycotts the company endures, the number of users keeps expanding; on average, those users are growing older, and that presumably redounds to Trump’s advantage. “I understood early that Facebook was how Donald Trump was going to win,” Parscale said, in October, 2017, on “60 Minutes.” “Facebook was the method—it was the highway which his car drove on.”
The instant a Presidential election is over, everyone who worked on the losing campaign is recast as a dunce, and everyone on the winning side is reborn as a genius. In 2016, three weeks after Election Day, Harvard’s Institute of Politics hosted a panel discussion featuring leaders of Hillary Clinton’s campaign and Trump’s campaign—the first public reunion of the now dunces and the now geniuses. It got heated.
“I would rather lose than win the way you guys did,” Jennifer Palmieri, Clinton’s director of communications, said.
“No, you wouldn’t, respectfully,” Kellyanne Conway, one of Trump’s campaign managers, said.
Later in the discussion, Mandy Grunwald, another Clinton adviser, rephrased Palmieri’s rebuke as a backhanded compliment. “I don’t think you guys give yourselves enough credit for the negative campaign you ran,” she said, alluding to “the fake Facebook stuff, or the great dark-arts stuff you were pumping out there.” Turning to Parscale, she went on, “I’m fascinated to hear all about that, because it’s so hard for us to track.”
“I’d agree,” he said. “That’s the beauty of Facebook.”
Another morning-after-Election Day tradition is the postmortem. Every political demise has a hundred etiologies. Still, when it comes to the 2016 election, we can’t seem to help ourselves: Was it the Russians? The letter from James Comey? The weather in Wisconsin? These days, the culprit many people settle on is the Internet. “There’s a tendency to turn it into a catchall explanation,” David Plouffe, a Democratic strategist who was Barack Obama’s campaign manager in 2008, told me recently. “Which is understandable, given how powerful it is, and how hard it is for most people to understand.” Plouffe came to prominence at a time when social media was generally perceived as innocuous, even liberatory. He and his team made extensive use of digital fund-raising, organizing, and advertising; after Obama’s victory, they were hailed as innovators. “At the same time, the digital stuff is not a magic potion,” he continued. “It’s an ever-evolving tool. A tool that the Trump campaign, whatever else you want to say about them, used quite effectively.” (Plouffe is now an adviser to the Chan Zuckerberg Initiative, a philanthropic organization set up by Facebook’s founder and C.E.O., Mark Zuckerberg, and his wife, Priscilla Chan.)
Between June and November of 2016, Parscale’s firm was paid ninety-four million dollars, most of which went toward digital advertising. Some of the ads were standard fare about national security or the debt; others were designed to help Trump’s mendacity and nativism go viral on social media, where lies and fractious memes are disproportionately likely to be amplified. Facebook did not maintain an archive of its political ads until 2018, so some of the 2016 campaign’s dodgier efforts may be lost to history. But we do know that Trump tweeted an image, originally circulated on anti-Semitic message boards, of Hillary Clinton’s face, a Jewish star, and a pile of cash; that one of Parscale’s staffers made an ad featuring audio of Hillary Clinton referring to African-Americans as “superpredators” (the intention was to microtarget the ad to black Facebook users in swing states); and that Defeat Crooked Hillary, a Facebook page funded by a pro-Trump super PAC, disseminated several conspiratorial videos, including one insinuating that Clinton was taking illicit drugs and another alleging that she had undisclosed ties to Vladimir Putin.
The point of all this, of course, was to sway the election in Trump’s favor, and, given the election’s narrow margins, it’s highly possible that it worked. (The Internet Research Agency, a troll farm associated with Putin, purchased thirty-five hundred Facebook ads between 2015 and 2017; Parscale and his team bought millions.) None of this is evidence of election-related malfeasance, however. Most social platforms have rules against harassment, hate speech, and violent threats, but they have usually granted exceptions to those rules when the boundaries are tested by a Presidential campaign, or by the President. Parscale used Facebook to move fast and break things, but it seems that the things he broke were long-standing norms, not laws. However, the Internet has disrupted global politics so rapidly, and regulators have been so slow to adjust to the new reality, that there weren’t many relevant laws to break.
In response to questions from The New Yorker, Parscale issued a written statement, which read, in part, “This isn’t journalism, it’s a transparent attempt to divert attention away from the fact that President Donald Trump is on track to steamroll past the socialists and win reelection this November.” He did not agree to be interviewed for this article, but dozens of people did, including people who worked with him and against him in 2016. Predictably, Parscale’s name elicited praise from most pro-Trump Republicans and scorn from nearly everyone else. “I can tell you with high confidence that Brad Parscale is not a genius,” Tara McGowan, a left-leaning strategist, told me. Nevertheless, “he undoubtedly had a massive impact on the outcome of the 2016 election, and he undoubtedly will again in 2020.” For better or worse, she continued, “you don’t need to be a genius to have a massive impact. You don’t even need to break the rules. An average person, given enough time and money and support, can use Facebook to help a demagogue win a national election.”
Parscale often tweaks his biography, in Trumpian fashion, to suit his immediate needs. Sometimes he portrays himself as having been an aimless rube before a fortuitous encounter with Donald Trump. Other times, he says that his work on the Trump campaign was merely an extension of “what I’ve done for twenty years.” Recently, warming up the crowd at a Trump rally in Florida, Parscale spoke as an avatar of red-state alienation: “I was born in Kansas. And people in Kansas don’t matter. I lived in Texas. They tried to push our voice out.” On Instagram, he looks more like a member of the Intracoastal élite, posting a video from aboard his thirty-two-foot Sea Ray cruiser (“Love boating through Fort Lauderdale”), or a photo of an outdoor fire pit next to an emerald-green patch of lawn (“Good fire night! 🔥”).
In a 2016 interview with Wired, Parscale called himself “a farm boy from Kansas.” His childhood home was not on a farm but on a paved cul-de-sac in Topeka, within walking distance of a Sonic Drive-In and a disk-golf course. His parents, Dwight and Rita, were entrepreneurs whose businesses, according to ProPublica, “included a swimming pool company, a scuba shop, real estate enterprises, restaurants and a Western-themed nightclub featuring a mechanical bull.” “Am I worth over a million bucks?” Dwight said in an interview with ProPublica. “Yes. But that’s not that much today.” Brad attended a public high school, Shawnee Heights, which students at a nearby school sometimes referred to as Scrawny Whites. In Parscale’s case, the aspersion was exactly half accurate: by ninth grade, according to a former coach, he was already a sturdy six feet five.
He attended the University of Texas, San Antonio, on a basketball scholarship, then transferred to Trinity University, a nearby liberal-arts school, to study business. Shortly after graduation, in 1999, he moved to Orange County in California to work for his father, who was then the C.E.O. of an animation-software company called Electric Image. The company filed for bankruptcy in 2002, and Brad Parscale returned to San Antonio. Later, in an interview with the Palm Beach Post, he recalled visiting a bookstore and asking for the best-selling book in the business section, which turned out to be an instructional text about Web development. At first, he said, “being a good procrastinator, I didn’t read it.” He got around to it two weeks later, when he was bedridden with food poisoning. “I finished it and thought, I could do that,” he said.
In 2005, he founded a Web-development company called Parscale Media. “Brad is a power player in the international web marketing community,” his online bio read. Actually, Parscale Media, which was run out of a small office next to a tattoo parlor, mostly produced simple sites on behalf of brick-and-mortar businesses around town—Finck Cigar Company, Texas Hill Country Landscaping. But Parscale was also open to political commissions. In 2010, he designed a campaign site for Karen Crouch, a conservative lawyer running for county judge, with a slogan in bold type: “Respect for the Victim. Tough Justice for the Criminal.” For a far-right movement called My America Again!, an alternative to the Tea Party that billed itself as “a phalanx of self-governing citizens bound by Christ, honor, and patriotism,” he built a password-protected site that purported to “harness the power of the Internet” to facilitate a new kind of political activism.
Parscale was also beginning to experiment with social-media marketing, which allowed him to measure, with ever more empirical specificity, where people were likely to focus their attention. In 2010, in an interview with a small Web-development blog, Parscale was asked about a project he’d recently completed for Dury’s Gun Shop, in San Antonio. At first, he said, the store’s owners weren’t even thinking about selling guns online; they’d only commissioned Parscale to make an online catalogue of the store’s inventory. Parscale showed the owners how, through search-engine optimization and “great category management,” they could find new customers outside South Texas. “Dury’s was mesmerized by the amount of hits they were getting from nearly every city in the U.S.,” he said. “Once Dury’s could visualize the potential business from the new web traffic, they were ready to sell guns.”
Time’s Person of the Year in 2010 was Mark Zuckerberg. The accompanying profile was mostly adulatory. (At one point, Zuckerberg receives a friendly visit from the director of the F.B.I., an otherwise taciturn man named Robert Mueller, who is described as “delighted” to make Zuckerberg’s acquaintance.) The piece also notes that Facebook “knows exactly who you are and what you’re interested in, because you told it. So if Nike wants its ads shown only to people ages 19 to 26 who live in Arizona and like Nickelback, Facebook can make that happen.” The article did not mention that the same sophisticated targeting tools, designed to sort the American population into various micro-demographic segments in order to influence their purchasing decisions, could also be used to influence their other behaviors, including the way they vote.
In 2011, Jill Giles joined Parscale Media as its creative director, and the company became Giles-Parscale, Inc. A respected graphic and interior designer, Giles was “responsible for much of the cool look of San Antonio’s chic buildings and restaurants,” according to the San Antonio Current. Giles and Parscale owned two of the more ambitious Web-design businesses in town, and the merger allowed them to focus on their respective strengths: Giles made everything look good, and Parscale made everything work on the back end.
The company moved to a tonier office across town and took on more lucrative clients—the University of Texas, ExxonMobil, a few real-estate agencies in New York. Giles, like most San Antonians, was a Democrat; Parscale, like most Texans, was a Republican, at least in theory. “I had a mediocre voting history, let’s just put it that way,” he said later. (Public records suggest that he registered to vote for the first time at the age of twenty-seven.) “Brad was a businessman,” Quintin Mason, who was Parscale’s college basketball teammate and who remains his friend, told me. “He was savvy, ambitious, all about getting to that next level of success.” Once or twice, he continued, “I heard him refer to himself as having libertarian tendencies”; other than that, he said, “politics never really came up.”
Around 2012, Parscale was at an IHOP, eating a ham-and-cheese omelette, when he got an e-mail from a woman who worked for the Trump Organization. Trump International Realty needed a new Web site, and she invited Parscale to bid on the project. He did, and, whether by luck or by intuition, he met the main requirement for anyone who wants to win Trump’s business: he bid low. As Parscale later told it, Eric Trump, upon receiving Parscale’s written proposal, called him to say, “We think you’re missing a zero, and we don’t know if you’re just dumb or you don’t know what you’re doing.”
Parscale got the contract, and it led to more: Trump Wineries; the Eric Trump Foundation; Caviar Complexe, Melania Trump’s line of skin-care products. In February, 2015, for fifteen hundred dollars, Parscale built DonaldJTrump.com, a bare-bones home page for Trump’s Presidential exploratory committee. Four months later, when Trump announced his candidacy, Parscale, this time charging ten thousand dollars, updated and expanded it to turn it into a full-fledged campaign site.
By then, Parscale had mastered the second requirement for anyone who wants to do business with Trump: obsequious public displays of loyalty. Parscale gushed to reporters about Trump’s “amazing family,” and called working with him “a great honor.” He reserved his most fawning sobriquets (“genius”; “truly a loving person”) for Jared Kushner, Trump’s son-in-law and the de-facto manager of his campaign. In late 2015, according to ProPublica, two Trump-campaign staffers conferred over e-mail about a plan to “transition away from” Parscale’s services, complaining that his sites often crashed under heavy traffic. The next day, the decision was abruptly reversed. “We’re going to stick with Brad,” one staffer explained to the other. “Brad is considered family.”
As the campaign expanded, Parscale’s approach grew more ambitious. At first, according to an associate of his, “it was just Brad, alone on his laptop, buying Facebook ads”; over time, Parscale, drawing on his close relationships with several Trump family members, especially Kushner, persuaded them to devote more of the campaign budget to online marketing. Trump, who adores television and does not seem to know how to use a computer, was dubious. One day, in Trump Tower, according to the Washington Post, Trump loudly berated Parscale for “wasting millions of dollars on Facebook.” Pointing to a nearby television, Trump said, “That’s how people win elections.”
“If you are going to be the next President, you’re going to win it on Facebook,” Parscale responded. Trump relented, but he didn’t seem convinced.
Parscale was despondent—“I hadn’t even seen him yell at anyone, let alone me,” he told “60 Minutes”—and he walked around midtown for hours, thinking about quitting. Eventually, other Trump family members called to talk him down, and he decided to stay on. After that, Trump either changed his mind or stopped paying attention. Parscale’s digital operation kept growing, and the candidate did not stand in his way.
Parscale became Trump’s digital director in June, 2016. The campaign had its headquarters on the fifth floor of Trump Tower, but Parscale ran his operation from San Antonio, in a makeshift office near the airport. “As far as everyone around town could tell, Brad’s whole motivation was: Trump is a big client, and I work my ass off for my clients,” an acquaintance from San Antonio told me. “Brad’s a competitor. Whatever he’s doing, he likes to win.” Parscale used an array of online gimmicks to promote his candidate—Snapchat filters, live-streaming on YouTube, fund-raising by text—but he devoted most of his attention to Facebook.
Parscale’s operation was unofficially called Project Alamo, a reference to the grisly encounter in a nineteenth-century border war between Texas separatists and the government of Mexico. Project Alamo soon grew to more than a hundred people, including campaign staffers, employees of the Republican National Committee, and venders from various tech companies. “He ran the 2016 digital campaign the way you’d run any other e-commerce operation,” a rival digital strategist told me. “He was selling Trump, but he could have been selling sneakers. He looked at the analytics on Facebook, saw what was popping on a given day, and went, Let’s pump money into that and let the algorithm feed it to our audience.”
In a post-election interview on Fox News, Parscale said, “For the first time in history, the data operation ran everything, from TV buying”—placing local television ads—“to where we were on the ground.” Campaign strategists used real-time analytics when deciding where to send canvassers, where to hold rallies, even what Trump should say at which rally. (“It might be, ‘Sir, our Facebook data from this area suggests that people want to hear you talk about tax cuts,’ ” a person familiar with the campaign operation told me. “Whether he actually took that advice was another question.”) The campaign used software to generate an endless stream of ads, each distinguished by one or more tiny variations: a new typeface, a new color, a new aspect ratio, a photo of Trump taken from a slightly different angle. “Certain people like a green button better than a blue button,” Parscale said on “60 Minutes.” “Some people like the word ‘donate’ over ‘contribute.’ ”
If each variation is counted as a distinct ad, then the Trump campaign, all told, ran 5.9 million Facebook ads. The Clinton campaign ran sixty-six thousand. “The Hillary campaign thought they had it in the bag, so they tried to play it safe, which meant not doing much that was new or unorthodox, especially online,” a progressive digital strategist told me. “Trump’s people knew they didn’t have it in the bag, and they never gave a shit about being safe anyway.” Bernie Sanders, who ran as an outsider with a base of avid support, also campaigned aggressively online, using social media to locate an unprecedented number of small donors. His 2020 campaign has found even more small donors, again largely through fund-raising appeals, which have become so widespread that a video clip of Sanders wearing a formidable pair of mittens and intoning “I am once again asking for your financial support” recently became a meme.
Eric Wilson, Marco Rubio’s digital director during his Presidential run, in 2016, told me, “The best online marketers are agnostic, as opposed to prescriptive. Anyone with a lot of money can buy a lot of ads, but what really matters is measurement, because without that you have no idea which ads are having any effect.” This sort of measurement is the province of “ad-tech” firms. Clients decide which metrics they want maximized—often some quantitative measure of success on Google and Facebook, which together control about half of the online ad market—and the ad-tech firms optimize for that outcome. In the summer of 2016, Parscale hired two leading ad-tech firms—Sprinklr, based in New York, and Kenshoo, based in Tel Aviv—to send subcontractors to work for him in San Antonio. Sprinklr also assigned remote employees, stationed in various time zones, to crunch numbers at all hours. In addition to data provided by the R.N.C. and traditional voter files, the Trump campaign had access to a repository of information provided by the Data Trust, a private company that Karl Rove and other conservative bigwigs had established in 2011. There are restrictions that prevent certain kinds of data sharing among nonprofit political entities, but those don’t apply to for-profit companies.
According to a source familiar with the campaign, Parscale pitted Sprinklr and Kenshoo against each other, hoping to inspire a “trading-floor mentality.” The idea came from Jared Kushner, who got it from his friend Gabriel Leydon, a mobile-gaming entrepreneur from Palo Alto. (Leydon founded Machine Zone, now called M.Z., a multibillion-dollar company known for the popular games Mobile Strike and Game of War.) Parscale, hoping to turn his operation into “a meritocracy,” told the ad-tech firms that they would have to compete, and that the winner would earn the campaign’s business. Over time, Kenshoo’s performance started to lag, and the Sprinklr employees were kept on. Sprinklr, which was valued at $1.8 billion in 2016, now lists several of its prominent past clients on its Web site—Nike, NASA, Nasdaq—but makes no mention of the Trump campaign.
One Project Alamo staffer took several pages of notes during the campaign, recounting the operation in remarkable detail. (The notes have never been published, but they have been shared privately with U.S. government officials.) I reviewed the notes and spoke at length with the person who wrote them, who asked to remain anonymous. “I would draw your attention, first of all, to what’s not in them,” the staffer said. “They’re not, for instance, about how I sat next to some guy named Vlad who had a direct line to the Kremlin.” If the Trump campaign was accepting foreign intervention, the staffer was implying, then it wasn’t evident in the San Antonio office.
Parscale relied on Facebook to help him accomplish several campaign objectives, including persuasion, fund-raising, and G.O.T.V., or “get out the vote.” Finding and motivating likely voters through traditional means, such as TV ads or door-to-door canvassing, is expensive and time-consuming compared with social media. The notes refer to a study conducted on Facebook in which likely Republican voters in early-voting swing states were split into two groups: an experimental group, which was given information about early voting, and a control group, which was not. The experimental group was more likely to be aware that early voting was an option—significantly more likely, for example, in Florida, a state that Trump won by a single percentage point.
Some of Parscale’s subcontractors in San Antonio, including a couple of his most trusted advisers, were employees of Cambridge Analytica, the firm best known for acquiring the data of eighty-seven million Facebook users in 2014. Asked about the firm’s impact on the 2016 election, the staffer said, “That’s another story line that gets blown out of proportion”: although some of the Cambridge data was acquired under dubious circumstances, “what they actually did with it was pretty standard data science.” When the data breach became international news, in 2018, it incited a wave of public outrage, not least because Cambridge Analytica had a hand in some of the most misleading political campaigns in recent memory: Brexit; Uhuru Kenyatta’s disinformation-heavy campaigns, in Kenya; the despotic propaganda tactics of Rodrigo Duterte, in the Philippines. Still, the company’s central sales pitch—that its proprietary “psychographic modelling” allowed it to predict each user’s deepest fears and desires—is now widely dismissed as snake oil. “There’s never been any public evidence that Cambridge Analytica brought anything to the table beyond what was standard campaign practice,” Daniel Kreiss, a political-communications professor at the University of North Carolina, told me. In 2016, the Trump campaign paid Cambridge Analytica slightly more than six million dollars. Giles-Parscale was awarded fifteen times more, making it one of the most highly paid venders in political history.
The office culture within Project Alamo was one of brash experimentation—not unlike that of a successful but amoral startup. According to the staffer’s notes, all employees, from executives to interns, were encouraged to voice their ideas, no matter how ridiculous. For example, one junior designer, responding to the popularity of Pokémon Go, made a video in which a Hillary-themed Pokémon was being chased—it didn’t make much sense, and it had little to do with politics, but it turned out to be a viral hit. Mark Zuckerberg once wrote a manifesto of sorts, in which he encouraged his employees to follow what he called “the hacker way”: “Instead of debating for days whether a new idea is possible or what the best way to build something is, hackers would rather just prototype something and see what works.” In San Antonio, Parscale seemed to aspire to a similar ethos.
One of Parscale’s favorite Facebook marketing tools was called Lookalike Audiences. “I mean, it’s why the platform’s great,” he said in an interview with “Frontline,” in 2018. The tool works like this: an advertiser uploads a “Custom List,” an Excel spreadsheet of people the advertiser wants to target. Even if the spreadsheet comprises only scraps of information—an e-mail address here, a mobile advertising I.D. there—Facebook, with its unparalleled accretions of consumer data, can usually fill in the gaps. Lookalike Audiences then multiplies the power of Custom Lists, using Facebook’s proprietary software to replicate the target audience. If you have a Custom List of three hundred thousand people, Parscale explained to “Frontline,” you can use Lookalike Audiences to find another three hundred thousand Facebook users with attributes similar to those in the first group. One of the most difficult tasks of a political campaign—distinguishing likely supporters from the undifferentiated mass of the American electorate—can now be accomplished instantly through artificial intelligence. When the “Frontline” interviewer asked how accurate Lookalike Audiences was, Parscale called it “pretty amazing.”
Project Alamo staffers also experimented with what are commonly known as “dark posts”—Facebook ads that can be targeted to specific, often small, groups of people. Dark posts were not illegal, nor did they violate Facebook’s terms of service, but they were controversial, because they skirted conventions of transparency. In the past, the thinking went, a campaign that chose to run a racist ad would at least suffer blowback from the many nonracists who saw it; in the era of microtargeting, when a racist ad could be served only to people whose online behavior demonstrated a proclivity toward racism, that check was gone.
Two Bloomberg Businessweek reporters visited Project Alamo shortly before the 2016 election. Parscale posed for a photo while hunched over his laptop; on the wall behind him were a “Bikers for Trump” poster, a novelty dollar bill with Trump’s face on it, and an inspirational quote falsely attributed to Lincoln. Parscale told the reporters, “I always wonder why people in politics act like this stuff is so mystical. It’s the same shit we use in commercial, just has fancier names.”
In their piece, the reporters quoted a “senior official” within the campaign as saying, “We have three major voter-suppression operations under way.” The targets of those operations were said to be “idealistic white liberals, young women, and African Americans.” In common parlance, “voter suppression” refers to a narrow set of tactics that are openly racist, unconstitutional, or both (see Georgia in 1960—or in 2018, when its secretary of state was elected governor after purging several thousand people, many of them African-American, from the voter rolls). But the term can also apply to traditional negative advertising intended to dampen enthusiasm for an opponent. Trump’s use of such negative campaigning, enhanced by the latest in targeting technology, seems to have helped: if African-American turnout in Pennsylvania, Michigan, and Wisconsin had been as high as it was in 2008, Clinton might have won. (Parscale has repeatedly denied that any such operations took place. “I would actually say we ran the least amount of negative ads I’ve ever seen in a Presidential campaign,” he said.)
Clinton had a budget of more than a billion dollars, a significant chunk of which was spent on TV ads. Trump’s campaign budget was more than thirty per cent smaller, but he invested more money in Facebook ads, and those ads cost him much less, on average, thanks to the platform’s instant-auction system, which rewards viral success. “A canny marketer with really engaging (or outraging) content can goose their effective purchasing power at the ads auction,” Antonio García Martínez, a former Facebook employee, wrote for Wired in 2018. Parscale responded to Martínez’s piece on Twitter: “This is why @realDonaldTrump was a perfect candidate for Facebook.”
Of all the benefits the Trump campaign reaped from social media, surely the most potent came in the form of free human labor. “I asked Facebook, ‘I want to spend a hundred million dollars on your platform. Send me a manual,’ ” Parscale said to “Frontline.” “They say, ‘We don’t have a manual.’ I say, ‘Well, send me a human manual, then.’ ”
In June of 2016, Facebook dispatched what is often called an “embed.” He was a young man from its ad-sales department who had previously worked for several Republican-affiliated causes. He spent most of the next four months in San Antonio, working with the Trump campaign. Other Facebook employees rotated through the office on a semi-regular basis; Google and Twitter also sent sales reps to the campaign.
“On the commercial side, all big accounts get reps like this,” Tatenda Musapatike, a former Facebook sales rep, told me. “It’s standard. Coca-Cola gets a Facebook rep, working on commission, whose job is to advocate for Coca-Cola within Facebook, and vice versa.” Sales reps were taught that the more useful they were to clients, the more money those clients were apt to spend. “Managers would always talk about ‘earning the badge,’ ” she continued. “As in, you’re so tightly aligned with your client that they think of you as part of their team, and they give you a security badge to get in and out of the building.” In a 2017 paper in the journal Political Communication, Daniel Kreiss and a fellow communications scholar, Shannon McGregor, wrote that embeds “go beyond promoting their services and facilitating digital advertising buys, actively shaping campaign communication through their close collaboration with political staffers.” (Facebook still offers extensive support to political campaigns, but it claims that this support no longer includes embeds.)
The notes taken by the Project Alamo staffer describe a tense office-wide meeting, early in the campaign, during which Parscale made it clear that he distrusted the reps from Facebook and Google, whose bosses presumably wanted Trump to lose. Shortly thereafter, the Facebook embed demonstrated his value: he designed a Custom List of everyone who had interacted with one of Trump’s Facebook pages during the primaries, then sent those people targeted ads asking for donations. The ads cost three hundred and twenty-eight thousand dollars; they raised $1.32 million, a net gain of a million dollars in a single day. After that, Parscale started taking the Facebook embed’s advice.
During the election, the embed did his best to keep a low public profile. The day after Trump’s victory, Gary Coby, the campaign’s digital-advertising director, tagged him in a tweet, calling him “an MVP” of the campaign. The embed was twenty-eight-year-old James Barnes, from Tennessee. He responded to his newfound notoriety by deleting his Twitter account.
Barnes recently told me that, although he grew up in an evangelical family and had long considered himself a Republican, “I despised Donald Trump from the moment I knew anything about him.” On November 8, 2016, after spending months working overtime to help Trump win, he and a few Facebook colleagues went to the polls in Washington, D.C., and he cast a ballot for Hillary Clinton. “My attitude during the entire campaign was, I’m a professional, I’m here to do a job, my personal preferences are irrelevant,” he said. Last year, “after reflecting on a lot of things, including my personal sense of duty,” he quit Facebook. He now works at Acronym, a left-wing nonprofit that is using social-media marketing to try to defeat Trump in 2020. (Acronym is also the main investor in Shadow, the company behind the app that broke down during this year’s Iowa caucuses.)
In December, I spent an afternoon with Tara McGowan, Acronym’s founder. At one point, she met with Barnes and other staffers, most of them former Facebook employees, in a conference room. They were interpreting the results of a large survey they’d just conducted on Facebook—fifty thousand voters across five swing states—sorting the voters according to dozens of metrics: race, gender, media diet, knowledgeability (measured by whether they knew which party controls the House). This was part of a “persuasion-analytics project” that Barnes is calling Barometer—his attempt to reproduce the power of Facebook’s political-marketing tools, now from outside the company. The goal is to gather data on which kind of anti-Trump ad—which subject matter, slogan, tone, and so on—will be most persuasive to each type of potential voter. “If we get even a small percentage of these people motivated to move in our direction, we win,” Harry Hantman, an Acronym employee who left Facebook in October, said. I asked McGowan what made her hopeful that Parscale’s tools could be turned against him. One of her answers was “James’s brain.”
In 2016, while Trump was accepting help from Facebook, Google, and Twitter, Hillary Clinton was offered equivalent services, but her campaign turned them down. “In my experience, the reps don’t add all that much,” a Democratic digital strategist told me. “They may be lovely people, but their job is to sell ads on their platform, and it’s sometimes too many cooks in the kitchen.” Mike Shields, a Republican consultant and a former chief of staff of the Republican National Committee, told me, “Hillary’s people were constantly reading articles about how fucking smart they were, and they let it get to their head. They must have just thought, We’ve got this, we don’t need anyone else. It was hubris.”
Parscale spent Election Night in Trump Tower, poring over returns, before finally heading to the campaign’s victory party at a Hilton around the corner. At 3:45 A.M., as the party was winding down, he tweeted the word “Digital,” followed by “#WINNING.” A few days later, he tweeted, “My goal was to show that digital is the future of campaigns. Check.”
If Aaron Sorkin ever writes a sequel to “The Social Network,” he might set its first scene two days after the 2016 election, when Zuckerberg attended a tech conference at a Ritz-Carlton near Silicon Valley. “The idea that fake news on Facebook, of which it’s a very small amount of the content, influenced the election in any way, I think, is a pretty crazy idea,” Zuckerberg said during an onstage interview. “Voters make decisions based on their lived experience.” But Facebook’s business model is premised on the assumption that there is no solid boundary between social media and “lived experience”—that what people see online affects what they buy, what they believe, and how they behave. “For a decade, our pitch to everyone, especially advertisers, was ‘We can target the exact people you want and make them behave in the exact ways you want,’ ” a Facebook employee who is concerned with the trajectory of the company told me. “Then Trump happens, and it’s, ‘Who, us? We don’t have any power, we’re just a place to share pictures of your dog.’ It was bullshit, but they tried to have it both ways.”
After the election, Parscale attempted to promote himself without upstaging his boss or making the voting public feel that it had been manipulated. “What happened here?” Lesley Stahl asked him on “60 Minutes.” “You’re, like, the secret sauce? The magic-wand person? You’re the Wizard of Oz behind the curtain?”
Parscale made a few modest noises, more demurrals than outright denials. “Nah, that’s a big statement,” he said. “But I—Do I think I played a big role? Yeah.”
Shortly before the segment aired, Parscale bought Google ads that would direct people searching for his name to a new Web site, TheParscaleEffect.com. It featured three gruff-looking photos of Parscale, a few bumper-sticker slogans (“Data drives strategy”), and a contact form (“Find out how the Parscale strategy can advance your business’s success”). If he was hoping to drum up new contracts without getting into trouble, it seems to have worked. In 2017, he appeared before the House Intelligence Committee, behind closed doors, to answer questions about Russian collusion. He didn’t reveal much. (“We got nothing,” Mike Quigley, a Democratic congressman on the committee, told Politico.) The Mueller report mentions Parscale only once, citing a retweet of an account called @Ten_GOP, now known to have been the creation of a Russian troll farm. After the report came out Parscale falsely claimed, “President Trump has been completely and fully vindicated by Special Counsel Robert Mueller.” Later, when asked whether he’d read the report, Parscale said, “I’ve read some of it.”
Ever since Trump began his long-shot candidacy, in 2015, his campaign rallies have given him momentum, press coverage, and an excuse to get out of the house. Since 2016, one of Parscale’s shrewdest innovations has been to turn the continuing rallies into data-mining opportunities. Tickets are free, but they can only be claimed by a person with a valid cell-phone number. The campaign now has a huge database of mobile numbers belonging to people who are motivated enough to attend a Trump rally, many of whom might not have shown up on a voter-registration roll or any other official data file.
“We have almost two hundred and fifteen million hard-I.D. voter records in our database now,” Parscale claimed last year, although his definition of “hard I.D.” is not clear. Even if Trump were banned from every social network, his campaign would be able to reach supporters by text. According to Parscale, the campaign is on track to send “almost a billion texts, the most in history”—and texts are far more likely to be opened than e-mails, social-media posts, or news articles. “We’ve been working on this around the clock for three years,” a senior official who works on the 2020 digital campaign told me. He acknowledged that the campaign doesn’t have the same scrappy, subversive energy as in 2016—“It’s hard to feel like a total underdog when you have the White House”—but, he added, “we’re not slowing down. We’re ramping up.”
In October, 2019, Thomas B. Edsall wrote a long Times column called “Trump Is Winning the Online War,” listing several of the “technological advances that have allowed Trump and the Republican Party to leave Democrats in the dust.” If money were no object, some of these deficits could be overcome quickly; others might not be surmountable by November. “The Trumpies have been really good at persuasion work—being relentless in hitting their target audience with their messaging,” Colin Delany, a digital consultant, said. “That’s most effective when you can repeat it over a long period of time.”
Last year, the Trump campaign spent far more on Facebook ads than any of the Democratic campaigns. Since January, the trend has been reversed, mostly due to two cash-rich and charisma-poor ringers, Tom Steyer and Mike Bloomberg. Bloomberg is currently building a digital operation that could come to rival Parscale’s. His unusually large and well-compensated campaign staff includes Gary Briggs, formerly Facebook’s chief marketing officer, and Jeff Glueck, the former C.E.O. of Foursquare. Sabrina Singh, a spokesperson for the campaign, said, “In comparison to Trump’s operation, Mike Bloomberg is the only Democrat positioned to compete with him on every single digital platform.”
Bloomberg has spent nearly fifty million dollars on Facebook this year, and has given his digital staff an unusual amount of freedom. He wasn’t onstage during the Democratic debate in Iowa in January, but his campaign’s official Twitter account posted incessantly and absurdly (“Mike can telepathically communicate with dolphins”; “WHAT IS THE BEST PART OF THE BODY TO GET A BLOOMBERG 2020 TATTOO?”). Recently, the campaign paid more than a dozen Instagram influencers, including TankSinatra, FuckJerry, and MoistBuddha, to run pro-Bloomberg sponsored content. Viral stunts like this come at a cost, both in dollars and in personal dignity; and it isn’t clear whether the Instagram ads, which winkingly portray the candidate as a stiff plutocrat interested in buying an election, will appeal to the target demographic. But the Bloomberg campaign is an interesting test case: if enough well-placed memes can turn a mediocre hair product or a boring pop song into a hit, then why not a Presidential candidate?
For years, there was no Democrat-affiliated counterpart to the Republican-affiliated Data Trust. In early 2019, the D.N.C. announced that it would partly address this asymmetry, launching an information-sharing operation, the Democratic Data Exchange, to be run by Jennifer O’Malley Dillon, an alumna of the Obama and Clinton campaigns. Before the operation could get off the ground, Dillon left to lead Beto O’Rourke’s ill-fated Presidential bid. Since then, the Democratic Data Exchange, which is now run by a Democratic operative named Lindsey Schuh Cortes, has gone unmentioned in the press. An official familiar with the exchange told me, “We are staying small and quiet for now, by design. We’re not playing in the primaries, but the goal is to be up and running in time for the general. We hope that all the Democrats who are no longer in the race will hand over their data at that time, but participation will be voluntary.” If Bloomberg, a multibillionaire, loses the nomination to Bernie Sanders, who intends to sharply raise taxes on billionaires, it’s possible that Bloomberg would transfer the data his campaign acquired to the Democratic Data Exchange, in the common interest of defeating Donald Trump. It’s also possible that he would refuse.
At Facebook.com/Business, there are dozens of “success stories”: case studies showing how Facebook ads helped a menstrual-underwear company by “broadening brand awareness,” or how an artisanal-jewelry company sold bracelets on Instagram. The case studies use internal Facebook data to demonstrate an ad campaign’s success through quantitative metrics. One details how the 2014 reëlection campaign of Rick Scott, the Republican governor of Florida, used Spanish-language ads on Facebook to target Latino soccer fans (“Buena suerte, Team USA!”). Andrew Abdel-Malik, the R.N.C.’s state director of digital strategy, is quoted in the case study: “Facebook Ads provided us with unique targeting capabilities . . . to reach different sub-groups of Hispanic voters in ways that were simply not feasible on TV and radio.” Scott was reëlected by a single percentage point. Four years later, he was elected to the U.S. Senate in a race so close that it triggered a recount.
The case studies can be filtered by industry, using an alphabetical drop-down menu. In 2018, the journalist Sam Biddle, writing for the Intercept, noticed that the Rick Scott case study had been buried, and that the “Government and Politics” category had been quietly removed from the menu. (“Gaming” is now followed by “Health and Pharmaceuticals.”) This did not mean that Facebook had stopped selling ads to political campaigns, just that the downside of drawing attention to the fact had started to outweigh the upside. (Facebook has since launched a site devoted to government and politics, with no success stories.) A former Facebook employee told me that, after the 2016 election, there was some internal chatter about drafting a case study that would demonstrate, in great detail, how Facebook had been a decisive factor in Trump’s victory. “It would have been one of the most extensive and convincing ones on the whole site,” the person told me. “The evidence was overwhelming. But, given the mood at the time, there was no way they were going to put that out there.”
In May, 2018, hoping to address concerns about dark posts and other controversial practices, Facebook built the Ad Library, which started to archive all political and issue-oriented ads that ran on the platform from that point on. “In meetings, if you bring up problems like misinformation, you’ll hear, ‘Well, we have the Ad Library now,’ ” the concerned Facebook employee told me. “The argument is, ‘If we put all the information out there, then people will find it and become better informed’—even though it’s clear that that’s not actually happening.”
In March, 2018, three M.I.T. computer scientists published a paper in Science comparing the dissemination of false rumors on Twitter to the dissemination of actual news articles. They found that the fake stories spread faster, in part because they were more likely to provoke an immediate emotional response in users. The same phenomenon appears to hold true for other social-media platforms and to apply to misinformation as well as fearmongering, rage bait, and racist propaganda, all of which go viral more readily than calm, patient deliberation. “Stuff that has a more alarmist and hyperbolic tone, or that makes people afraid or upset, is just going to travel better,” the Facebook employee told me. “That fits with human nature, and it’s how the platform is designed.” Without fundamentally altering Facebook’s News Feed algorithm, or the company’s underlying business model, this is unlikely to change.
This past fall, the Trump campaign ran a Facebook ad premised on the incendiary but false notion that the villain of the Ukraine corruption scandal was not Trump but Joe Biden. (Parscale repeated such claims several times on Twitter, adding, “The swamp! They’re playing us and the media is their lap dog!”) The ad, predictably, went viral. Biden’s campaign wrote a letter to Facebook, asking the company to take it down. Facebook’s head of global-elections policy, a former Rudolph Giuliani campaign official named Katie Harbath, explained that the ad would stay up because the platform’s rules do not prohibit lying, at least not when politicians do it.
In October, Zuckerberg appeared before the House Financial Services Committee. Representative Alexandria Ocasio-Cortez questioned him. “I just want to know how far I can push this,” she said. “Could I pay to target predominantly black Zip Codes and advertise them the incorrect election date?” In the ensuing back-and-forth, Zuckerberg clarified that this particular lie is prohibited on Facebook, but that most other lies are not.
Around the time of this testimony, hundreds of Zuckerberg’s employees signed an open letter. “We strongly object to this policy as it stands,” the letter read. “It doesn’t protect voices, but instead allows politicians to weaponize our platform.” The employees suggested six policy changes, all relatively narrow and easy to implement, including “Stronger visual design treatment for political ads,” “Restrict targeting for political ads,” “Spend caps for individual politicians.” Facebook took none of these suggestions. Instead, the company announced that it would “expand transparency,” including by adding more search features to the Ad Library.
Soon after, Elizabeth Warren’s Presidential campaign ran a Facebook ad. “Breaking news: Mark Zuckerberg and Facebook just endorsed Donald Trump for re-election,” the ad claimed. It was a deliberate provocation—a bit of fake news meant to protest fake news. The ad went on to clarify that the eye-grabbing claim was false, then continued: “It’s time to hold Mark Zuckerberg accountable—add your name if you agree.” The stunt garnered some good P.R. for the Warren campaign; it also enabled her to collect the e-mail addresses of many new supporters.
“There are people on every campaign, in both parties, who know how to use all these tricks,” Colin Delany, the digital consultant, told me. “When campaigns decide not to do something—whether that something is microtargeting, or so-called dark posts, or whether it’s outright lies or racism—they’re making a strategic choice.”
“It’s lovely that Democratic campaigns are so principled,” Tara McGowan, of Acronym, said. “I mean that sincerely. And yet it also scares the shit out of me, because the other side isn’t playing by the same rules, and our principles might make it all but impossible for us to regain power.”
A few of the most obvious loopholes were closed after 2016—it’s no longer possible, for example, to purchase a Facebook ad about an American political candidate using rubles—but many of the bigger ones remain. Yaël Eisenstat, formerly Facebook’s head of elections-integrity operations for political advertising, is now a visiting fellow at Cornell. “There’s a lot they could do to protect the integrity of the platform,” she told me. “They could label paid content as paid, even after people start to share it, which they don’t consistently do. They could put a label on every political ad—‘This ad has not been fact-checked’—which might encourage some skepticism.” Still, political ads make up only a tiny percentage of Facebook’s content and less than a per cent of its revenue. It would be much more difficult to fact-check everything that gets posted by every Facebook user, from high schoolers to the President. Eisenstat added, “If the larger goal is to have these platforms contribute to a healthier public square, to leave democracy healthier than they found it, then this is just the low-hanging fruit.” In December, 2016, an internal Facebook initiative called Project P—for “propaganda”—found dozens of right-wing pages peddling fake news. According to a recent Washington Post investigation, Joel Kaplan, an executive at the company who previously worked in the George W. Bush White House, objected to removing all the propaganda, “because it will disproportionately affect conservatives.”
On December 30, 2019, Andrew Bosworth, a top executive at Facebook and a longtime friend of Zuckerberg’s, posted a twenty-five-hundred-word “essay” on a private social network for Facebook employees. The text—by turns contrite and defiant, laden with carefully selected statistics and dubious allusions to J. R. R. Tolkien and John Rawls—was later leaked to the Times. Its central premise was that social media may be poisonous, but ingesting poison is a matter of personal choice. This was a long way from the idealistic posture of Facebook’s official mission statement, “To give people the power to build community and bring the world closer together.” “If I want to eat sugar and die an early death that is a valid position,” Bosworth wrote. “My grandfather took such a stance towards bacon and I admired him for it. And social media is likely much less fatal than bacon.” Bosworth also asked whether Facebook was “responsible for Donald Trump getting elected.” He concluded, “Yes, but not for the reasons anyone thinks. He didn’t get elected because of Russia or misinformation or Cambridge Analytica. He got elected because he ran the single best digital ad campaign I’ve ever seen.” Of course, what’s best for a political campaign, or for a company’s bottom line, is not always what’s best for the health of a nation.
“No one ever complained about Facebook for a single day until Donald Trump was President,” Brad Parscale has said. When the Obama campaign used Facebook in new and innovative ways, the media “called them geniuses.” When Parscale did the same, he continued, he was treated as “the evil of earth.” Despite the bombast and the false equivalence, this is basically true. Some of the public anxiety over Facebook is a response to how easily it can be abused, but much of that anxiety is about the outcomes the platform yields when it’s working as designed. Even leaving aside the Cambridge Analytica data breach and the allegations of foreign interference—even if nobody had ever violated any platform’s terms of service—many of the fundamental problems of social media still remain. Creepy surveillance, dissolution of civic norms, widening unease, infectious rage, a tilt toward autocracy in several formerly placid liberal democracies—these are starting to seem like inherent features, not bugs. The real scandal is not that the system can be breached; the real scandal is the system itself. In a sense, it’s almost comforting to imagine that the only bad actors on social media are Russian state assets, clickbait profiteers, and rogue political consultants who violate the law. If that were the extent of the problem, the problem could surely be contained. ♦
