Inside Our Wildly Expensive, Dubiously Effective Political Ad Machine

It was Friday, July 19, one of the last days before Joe Biden dropped out of the presidential race. Kamala Harris was rolling calls in a pre-lit film studio at the Democratic National Committee headquarters on Capitol Hill, as someone in the room described to me. The morning had been consumed by video conferences with panicked donors and volunteers, screens of distant little faces looking to Harris for some kind of encouragement. The calls were followed by a series of scripted, direct-to-camera spots. In one, Harris, wearing a black blazer that set her off against the beige and white background, asks for “a donation to support Joe and me. This is going to be a tough fight and we are all up for it.”

And that was true, to an extent. Biden was still in the race, but since the world witnessed him flailing during his debate with Donald Trump, his campaign had been running on fumes. His staff flipped into damage control. The donations dried up. Once Thomas Matthew Crooks put a bullet through Trump’s right ear, the gears ground to a halt. Biden’s team pulled their attack ads from the air and hunkered down until they got a read on the national temperament. By that Friday, as Harris made her calls, the campaign’s whole media apparatus—spread between Biden’s campaign committee, the DNC, and an array of allied PACs—had been hanging in a state of suspended animation for nearly a week. If the prospect of Trump getting a second term as president was distressing to you, these days were agony.

And then they were over. On Sunday, Biden announced he was bowing out of the race and endorsed Harris. By evening, Harris’s campaign had filed with the Federal Election Commission. That began the process by which she could assume control of the $96 million in Biden’s campaign committee. By Monday, donations were pouring in at record pace: $81 million in the first 24 hours. Endorsements followed in short order. In a matter of days, Harris had become the presumed nominee—and stupendously loaded. But all those dollars were still just promissory. There was as yet no clear plan for how to convert them into the real grist of a campaign: the ads.

In Athens, they had the agora. In Vienna, the coffee shop. Here in the states, we have a cyclical barrage of multiplatform political advertising. Regulations governing campaign finance have been rolled back considerably over the last several decades, and ever since Citizens United in 2010, when the Supreme Court decided that money had First Amendment rights, there have been virtually no limits on fundraising or spending in our political system, while only the most paltry regulations govern how our data is harvested. The results are awing. Every couple of years, our country is thrown into paroxysms by campaigns that start sooner and set new benchmarks for how data scientists try to extract dollars from likely donors and place ads in front of possible supporters. The bills for this add up. In 2016, just under $3 billion was spent on ads. In 2020, that figure had climbed to $9 billion.

Until Harris swapped in for Biden, this cycle was no exception. Biden announced plans to run for reelection in April 2023 and was making ads by the fall. The independent expenditures supporting him snapped into action soon after. Future Forward, one of the biggest pro-Democrat super PACs for the past several cycles, announced in January that it was buying $250 million of ads for later in the year—what it claims is the single largest buy any unaffiliated group had ever made. Trump, for his part, had never spent as lavishly as other candidates on ads. In 2016, Trump won the presidency while spending almost half of what Hillary Clinton did. But the big money was with the super PACs that backed him—Preserve America, MAGA Inc.—and they were going up with a major buy on TV spots.

It’s estimated that this election cycle will burn through $10.7 billion on ads—a new record for our country and, quite likely, the world. For comparison, that’s on par with Montana’s total state spending for the fiscal year 2023.

When campaigns ask you for money, they typically list a variety of expenses. Harris, in that ad she cut on July 19, mentioned hiring organizers and staff and opening field offices. She didn’t say a word about advertising, but the lion’s share of your money, north of 60 cents on the dollar, will go to making and placing ads, primarily video spots, which are the real big-ticket item within the industry. In a literal sense, then, these videos are the main grist of all that data and cash that drive modern American politics. And now, with only a little more than 100 days to go until the polls would close, Harris’s campaign needed to start churning them out at a clip.

The logistical complexities of that problem were many. A fresh, mediated version of Harris needed to be conjured up, with image and messaging calibrated for the campaign map. This Harris—ad Harris—would then need to be delivered to these potential supporters across a fragmented media landscape. Daunting as that project was, in practice the choice before Harris was simple. As our elections have swelled into multibillion-dollar media spectacles, the private sector has correspondingly spawned an array of for-profit firms to which campaigns outsource this work. Harris’s choice boiled down to who would run her paid media, and what constellation of firms would get the contracts to make and place her ads.

And at that point, political decisions also touched people in their pockets. “There’s just this massive pile of money, and it is the greatest treasure in politics as far as what we do,” a Democratic media consultant told me shortly after Harris launched her campaign. “So you can imagine there’s a lot of feelings about how that pie will be split.”

I had been following this world since early summer, when I began corresponding with a freelancer beset by a nagging sense of doom. (Everyone directly involved in making ads spoke to me on condition of anonymity, because they were not authorized to talk to the press.) The freelancer had been shuttling between battleground states for weeks, living out of a suitcase while he made testimonials that would be used in ads supporting the president. The work brought him into factories and union halls, into people’s homes. In these settings, ordinary Americans, handpicked by data analysts and producers to resonate with key demographics, explained to the camera why the country should reelect Joe Biden. The thing was, so far as the freelancer could tell, none of it was sticking.

The freelancer had made campaign ads for several cycles before this. Four years ago, even while the country was still in the grips of the pandemic, the atmosphere had been hopeful. Biden was running strong on a message of change and sanity, and although much of the film industry had been idled by Covid, the production process had felt light-footed. Teams of underemployed editors, working out of their homes, were shipping video spots at a lightning pace. Four years before that, during Hillary Clinton’s 2016 run, the freelancer had called the election for Trump by the end of summer. The consultants writing Clinton’s ads seemed to have open contempt for their client. This year, in a different way, the vibes were bad again.

At the top, Biden’s media team was pretty much a reshuffled version of the people who had carried it off in 2020. Mike Donilon, a longtime Biden confidant and a seasoned media maker, had come on again as head strategist. Jen O’Malley Dillon and Anita Dunn, both of whom had worked on 2020 campaign, were also back, Dillon as campaign chair, and Dunn as senior adviser to Biden himself. Overseeing the ads was a young firm called Blue Sky Strategy. Blue Sky was founded in the run-up to the 2022 midterms, but it, too, really had its roots in 2020. At its helm were the former head of paid media for Biden’s 2020 campaign, Patrick Bonsignore, and Jon Fromowitz, a media consultant who worked for the super PAC Future Forward.

Four years ago, this team had marshaled an impressive advertising campaign built around testimonials from nonactors and clips of Biden, parading around in a Covid mask or holed up in an office with dark turquoise walls, talking directly to a camera. This year, with the pandemic behind us, the cameras had been able to follow Biden to his campaign rallies, capturing tape of him working the lines and chopping it up with smiling supporters who had come out to see him. Even so, many of his ads still leaned hard on testimonials, using a chorus of ordinary Americans to underline the appropriate message for whatever demographic track to which the ad was targeted: veteran, faith, Latino, rural, etc. The effect was Whitmanesque, framing Biden as the focal point of a broad coalition of voters who supported his administration and feared a second coming of Trump. In this, one could discern the campaign’s belief that the voters who had carried Biden in 2020, and even more notably those who had turned out during the midterms in 2022 to break the supposed red wave—these voters were a reliable base, and so long as Biden could hold them together, he could run through the tape again.

What this all looked like on the ground, with the camera running, was not so reassuring. The freelancer and his team were not finding people who were enthusiastic about the president. Among the score of interviews the freelancer had done this summer, only one person, he figured, was an actual Biden fan. The others certainly wanted Biden to win again, but they seemed lukewarm on the man himself. Other interviewees didn’t want to talk about Trump, as if to do so would have been unsuitably partisan. None of this had been an issue in 2020.

In contrast to the excessive prudence of their interviewees, Trumpian ardor surrounded the film crew. “It’s every shoot, every single shoot, we have to be careful,” the freelancer told me. “Like the guy across the street is pro-Trump, so we don’t want to block the driveways [with the crews’ cars] and get into this confrontation.” Union reps told stories about arguing with their rank and file over which candidate would be better for labor. In a factory in Pennsylvania, the person giving the tour pointed at a guy on the line and said that he would take up arms if he found out the film crew was working for Biden.

Then there was the problem of selling Biden’s accomplishments. Talk to ad makers, and they will tell you that, out of all the various means campaigns use to reach voters—direct mailers, email, texts, static ads in your Instagram feed, viral marketing—video is the medium for persuasion. (The rest are there mainly to drive you to the polls or to tip you over the edge to donate.) The combination of image, sound, and narrative found in video is uniquely capable of evoking emotion, and it is through such amalgamations—where political content is fused with a specific structure of feeling—that a reluctant voter might allow herself to see a candidate or issue in a new light.

The affective alchemy could work when the ad highlighted the threat Trump posed, especially when it concerned abortion access. The 2022 midterms demonstrated how effective that message was in the wake of Dobbs. But selling Biden’s own record proved much harder. “I remember like when Bidenomics started being talked about, it was just like this is never going to work,” the freelancer said. “No one cares about this.” He had spoken to people whose whole towns were rejuvenated by the Inflation Reduction Act, but the ads that came from this weren’t connecting that well. I’d heard similar things from a camera and lighting technician who’d made ads for Biden. Tasked with an ad about the environmental upshots of the IRA, they arrived on location to find nothing but a bunch of storage sheds in the desert. Their interviewee was a single engineer, who could talk shop all day but struggled to connect the dots between the recondite technical problems he was facing and Biden turning the tide against climate change.

In many ways, it seemed to me, that was Biden’s problem tout court. He had changed the Democrats’ position on antitrust and corporate power at large, while he also saw through some of the most ambitious legislative accomplishments of any president since Roosevelt. These were political feats, particularly given how slim the Democratic majority was and how it relied on the intractable Joe Manchin. But for far too many Americans these victories might as well have taken place in the astral plane, so remote were they from our lives. Consultants and ad makers I spoke with complained that Biden wasn’t getting the credit he deserved, but they knew it was a hard sell.

What was obvious in the present was that buying groceries had gotten more punishing, that rising housing costs were forcing working people onto the streets, that credit card debt had shot up. And then, good Lord, there was the sustained atrocity of Israel’s assault on Gaza, all those maimed and dead children, the gory horror of the prison camps, and the president signing off on every new arms shipment put before him. A large number of Americans seemed inured to this suffering caused by our government, but more than half the country had come to view Biden as too weak to effectively run foreign policy, and only 33 percent approved of his handling of Gaza. The utter fecklessness of Biden and Secretary of State Antony Blinken in the face of Benjamin Netanyahu’s open defiance set a bad tone for Biden’s final year in office, one that his advanced age only accentuated. His administration appeared listless over these past months, weighed down by decisions that muddled its accomplishments and tarnished it with blood.

On these points, the freelancer and I did not see eye to eye. He was still a true believer in Biden. What I saw as the president’s noxious bovarism, he saw in a more tragic light. Biden’s was the story of a good man depleted by the ordeals of the office and the excessive contradictions of American politics and the Democratic coalition. And despite being clear-eyed about the challenges Biden faced, he wasn’t sure there was anyone better. “I felt like I could make the argument of why Biden could or should be president again, should win, is the best person for the ticket,” the freelancer said when I spoke with him on July 21, after Biden announced his withdrawal from the campaign. “I felt that way yesterday. I felt that way this morning, you know.” At other times, the divisions in the country seemed as if they ran too deep. “We are Bud Light,” he told me during another conversation. Just as Bud Light couldn’t stanch their loss of market share after conservatives boycotted the brand for hiring a trans spokesperson, so Team Biden struggled to move the needle, no matter how much money they threw at the problem. It didn’t matter how good the ads were. Greater forces seemed to be at play.   

For what it’s worth, that sentiment is largely borne out by academic work on political advertising. Given that our political system is organized around raising money to pay for campaigns, and that advertising is by far the major cost those campaigns incur, it is remarkable that the political scientists who study these matters agree that, on the whole, ads have only marginal effects on the vote share. Evidence of this goes back to some of the earliest academic work on how voters make up their minds, as Adam Sheingate shows in his history of political consulting, Building a Business of Politics. In the run-up to the 1940 election, the sociologist Paul Lazarsfeld tracked a group of people in Erie County, Ohio. Only 8 percent changed their views over the course of the survey, and these shifts were best explained by their socioeconomic standing, religious beliefs, and where they lived. The ads piped through their radios had only a faint presence in the data.

Successive generations of scholars have refined Lazarsfeld’s conclusions, but the fundamental insight—that ad effects are extremely modest—has stood the test of time, even as the media landscape changed from radio to TV and from TV to our current platform multiverse. A landmark study from 2011 found that TV ads could shift voters’ sentiments by up to 5 percentage points, but those feelings vanished within days of the ad going down. A macro study from around the same time disaggregated whom ads were affecting the most, concluding, “political ads influence mostly the political ignorant.” For others, the effects were minimal. There is, of course, plenty of anecdotal evidence suggesting that advertising can turn the tide in a race. Just this summer, House incumbents Jamaal Bowman and Cori Bush, both outspoken critics of Israeli atrocities in Gaza, each lost their primary races after the American Israel Public Affairs Committee, or AIPAC, spent more than $23 million to unseat them. Still, counterexamples are just as easy to come by, such as when, in the 2020 South Carolina Senate race, Jaime Harrison outspent Lindsey Graham by more than $30 million only to lose by double digits.

“As a scholar of political communication, I could never tell you that the narrative doesn’t matter, right?” said Erika Franklin Fowler, one of three political scientists who run the Wesleyan Media Project, which tracks how political ads are used in federal and gubernatorial elections. “I think the narrative matters. But I think to pin it all on political advertising specifically would be far too great of power to ascribe to this particular medium.”

And yet, like clockwork, the ad buy grows with every cycle. This is because winning is the main thing that matters to campaigns and to our political parties, and, for around the past 30 years, ads have been one of the primary tools they rely on to try to shift the balance amid historically tight races. It is at this level that the effect of ads in our political system becomes anything but marginal. The money needed to finance political ads has turned fundraising into the first test of a would-be candidate’s viability. “You can’t just be the scrappy upstart who’s willing to, like, roll up their sleeves and knock on every door. That’s not gonna quite do it. And to be honest, the party’s not gonna go for that,” one consultant explained to me. “They’d rather you be in the phone booth for 16 hours a day, dialing for dollars.” Gregory Martin, a professor of political economy at Stanford, put it this way: The main effect of advertising “is not changing the vote shares of the candidates who run. It’s in changing who runs in the first place.” The emphasis placed on fundraising means that wealthy candidates have a leg up over others, and that the particular interests of the wealthy loom larger in party platforms. Political scientists have demonstrated as much with historical surveys and regression analysis, but a thought experiment makes the same point: Chuck Schumer has been a reliable advocate for crypto—a industry that originated in libertarian fantasies of stateless monetary systems, has been the source of one financial scam after another, and also happens to be a driver of climate disaster. What’s in crypto for a Democrat other than deep pockets? The low rate of taxation on carried interest, which benefits private equity managers and enjoys support from both parties, is another such issue.

That our politicians are beholden to moneyed interest is a truism, of course, and ads have not made them any more so. For that, the most proximate cause is the rollback of campaign finance regulations. Ads, however, foster the incentive structure that thickens the nexus between American politics and money with every cycle. Running in a major race nowadays requires an immense body of specialized knowledge and technical ability, far exceeding the capacities of any campaign. Included in this are film crews, editors, and graphic designers, but that is only the start. There’s also polling and the ability to track and test ads. The heart of the industry is in the ad buys. That’s where the big money is, and it’s also where the task becomes the most complicated. In the 1970s, when political consultants became a fixture in U.S. politics, buying ads was mainly a matter of using past election results to pick your target audience and then buying up airtime in the desired media markets. Today, it means trying to figure out how best to disburse funds across multiple platforms, where each has not only its own demographic skew but its own systems for identifying its viewers and measuring an ads reach. Analytics are everything here. “We’re aiming one set of ads at persuasion targets, a different set at turnout targets, and avoiding those who oppose us altogether,” explained Jim Margolis, a founding partner of GMMB and former adviser to both Barack Obama and Harris. With smart TVs, which can capture data on what their viewers watch, it gets even more granular. Consultants who know you’ve seen an attack ad against their candidate can make sure you see their response on another platform, like your Instagram feed.

The image of American politics that these systems produce is incredible to behold. I was forwarded a few installments of a weekly newsletter Future Forward puts out called Flight Radar. It circulates among Democratic ad makers and offers an ad-focused summary of the previous week in the presidential race (“Nearly all Republican ads deal with immigration and crime, with the former seeming to pack a heftier punch”), with links to videos of relevant ads and the latest data. Clicking through on the data sends you into a spreadsheet of ad buys in all the major swing states over the previous week. These figures are further broken down by the source of spending (whether party or campaign committee, or a super PAC), by the major target track (“White Women 35+” or “Latino,” for example), and by the efficacy of the ads being aired. Functioning as a public audit of political ads for the presidential race, Flight Radar lets everyone in the industry see what the others are making and how well their products and the messaging that underlies them are performing.

In a given week in July in Harrisburg, Pennsylvania, for example, Republicans had outspent Democrats by more than 600 gross rating points, a unit that measures the reach of an ad. All of these were bought by the super PACs MAGA Inc., SAG Inc., and Preserve America. The ads they were running, you could see, were working fairly well, improving support for Trump by around 2 percent in randomized-control-trial ad tests. The following week, the Democratic side had cut that difference in Harrisburg to less than 200 GRPs, with significant ad buys by the campaign and Future Forward. At the same time, the Democrats had opened up a whole new front in North Carolina, where for the time being the Republicans had yet to commit a dime. Looking at all this, I had the sensation of being a general taking in reports on the state of the war, where with enough resources, the right messaging, and powerful creative, victory seemed within reach. You have to pinch yourself to remember that under the hood of the system producing such data is a profound contradiction—whereby our parties try to reach small percentages of persuadable voters through ads, using money that welds the same parties to interests wholly unrepresentative of their voters.

A Democratic strategist put it to me this way: If Ritz wanted to roll out a new line of snacks, the company would want a 12-week media campaign, at minimum, to lay the groundwork for a successful launch. Harris was given just a bit more time than that, during what was historically the fiercest part of a race, when about 70 percent of the spending occurs. Given a longer runway, political ads tend to follow a sequence. Early spots introduce the candidate to the public, boost name recognition, and frame the kind of race the candidate would want to run. Later, the ads begin hammering on the campaign’s messaging while drawing contrasts with the opponent and also trying to respond to attacks. In the last weeks, when you’re not likely to win any more voters to your side, the emphasis shifts to driving out the vote. With Harris, the whole race would be squeezed into an election window more like those in Europe, where regulations prevent the protracted ordeals we allow ourselves. Everything would have to happen almost simultaneously. “There’s kind of a walk and chewing gum at the same time” thing, a consultant said.

Trump’s side had started in almost the instant Harris received an endorsement from Biden. That same day, the super PAC MAGA Inc. went live with an ad accusing her of covering up Biden’s “mental decline.” A slew of other ads followed, many of which focused on immigration. The GOP had already calibrated their ad campaign to attack Biden’s handling of the border. When Harris entered the race, they simply rolled over the playbook. She became the “border czar,” shown dancing alongside other Black folk, with a filter on the footage to distort it. Over this, a male voice insinuated that Harris allowed ISIS to establish a beachhead in the United States. Most of the GOP’s best-testing ads followed in the same vein—tendentious, conspiratorial, clear dog whistles. One exception, bearing the title “Meet San Francisco Liberal Kamala Harris,” was basically a super cut of old interview tape of Harris from when she took more liberal positions than she has this year. In one clip, she says she does not believe in treating undocumented people who cross borders as criminals. In another, she insists that putting more police on the street won’t fix crime. At a time when Trump still appeared confused about how he was going to talk about Harris, the people cutting his ads had set a clear tack.

Still, none of it could compete with the earned media Harris had gotten simply from entering the race. In the dark days after Biden crashed and burned in the debate, it had been cathartic just to imagine what it would be like to have a candidate that was not so encumbered. And then it was a reality, and the elation spilled from social media into Harris’s rallies and then into her ads. Among the ad makers I spoke with, there were some lingering resentments about what had happened to Biden. “They’re just using texts and tweets and other stuff to call this guy all the names in the book,” one consultant said. “Then, the minute he drops out, he’s the greatest patriot that we’ve seen since Thomas Jefferson and George Washington.” But even this person had to admit that the enthusiasm around Harris was impressive. “There’s some energy, which is, you know, you can’t fake energy.” Seeing that and the outpouring of support for her, they could feel that maybe the switch wasn’t a complete disaster.

To her great credit, Harris showed remarkable political instincts in these early days, riding the good vibes while gradually defining her candidacy. Her ads followed suit. An ad internally dubbed “Freedom,” which arrived on July 25, relied on footage of Harris’s first rally as the presumed nominee in Milwaukee only two days earlier, along with a big assist from Beyoncé on the backtrack. It was necessarily a rush job, but the images were gorgeous, warmed by natural light and filled with motion from ample use of a crane and a Steadicam rig. The footage benefited, too, from the uncontrived delight that broke across Harris’s face every time she paused for the cheers. In her first speech in Milwaukee, Harris had opened with fulsome praise for Biden before lightly discarding the emphasis he’d placed on Trump’s threat to democracy. Instead, Harris offered a more forward-looking vision. “Freedom” did something similar stylistically. The footage had been shot and cut by some of the same people making Biden’s ads, but the feeling was far different. Aesthetically, if you overlooked the distorted clips of Trump and JD Vance that were spliced in here and there, “Freedom” lived in the vicinity of a Nike commercial.

As an ad, “Freedom” was mainly doing fan service. At one minute and 19 seconds, it was of irregular length and was not intended for broadcast. Its main home would be Harris’s Instagram grid and other social media platforms. It took another week before the campaign came out with a more standardized spot aimed at persuadable voters. This was “Fearless,” which dropped July 30. Stylistically, “Fearless” was a tamer affair, hewing close to the familiar grammar of a political ad, with an anonymous male narrator relating Harris’s background while text callouts hammered home the salient points (“Prosecutor For 20+ Years,” read one). Contentwise, “Fearless” was packed, stuffing Harris’s biography, agenda, and her catchphrases (“We are not going back!”) all into 60 seconds.

In these ads and her rallies, Harris made it clear she was building her campaign atop lessons learned from Biden’s run. Some of the best-testing messaging for Biden this cycle focused on the threat Trump posed to the Affordable Care Act. In “Fearless,” Harris says that Trump wants to end the ACA. Likewise, the claim that Donald Trump “has no plan to help the middle class,” which was featured in a Harris ad from early August, was a callback to a Biden ad from July, which contained the exact same line. Yet from the start Harris also made her campaign about affordability. It was probably not a coincidence that message testing from earlier in the year showed that rising prices were the most salient source of displeasure with Biden, beating out even his advanced age.

What followed was a stretch of weeks in which a peculiar dynamic set in. Outwardly, Harris’s campaign was spreading joy around the country and watching its polling numbers rise. Her choice of Tim Walz as her running mate was a high point here, onboarding Walz’s commonsense verdict that Trump and JD Vance were “weird” and yielding the best campaign hats I’ve ever seen. But inside the campaign, the transition was still working itself out. Mike Donilon, Biden’s head strategist, was on his way out. I had heard that about a week before it was confirmed in the press, but that was the only information that seemed solid. Rumors abounded about who would come in to replace him and what would happen to the rest of the campaign staff. On August 2, The New York Times reported that David Plouffe, who had run Obama’s campaign in 2008, was joining Harris’s team, although no one I talked to seemed sure how real Plouffe’s role would be. There was more certainty about Jim Margolis from GMMB. Margolis had also appeared in the news as a new addition to the team, but everyone I spoke with said so far his position existed in name only. When I talked with Margolis, he told me that he ended up working with Future Forward and other super PACs instead.

As for the situation a bit further down the food chain, people seemed to be on tenterhooks. One described the uncomfortable challenge of keeping your head down enough not to cause problems while also making sure everyone knew you were still “seen in the room.” “I can’t speak to anyone’s particular contractual deals and whatnot, but I can tell you that the tension that is there is not just about power or responsibility,” another consultant told me. “It’s about money and who gets it.” Fears of being replaced (“Or layered, as they say in the corporate world,” one consultant remarked) mixed with fears about Harris’s campaign strategy: whether the campaign would decide to bring in cheaper and less-experienced people; whether she was going a bit too hard on her digital ad buy over linear TV. These were self-interested concerns, as those expressing them admitted freely. But they also seemed to come from a genuine desire to see Harris win. Campaigning in the United States is a business, after all, and no bright line cordons off one set of concerns from the others. It helped that Harris’s fundraising machine was cranking along so well. “There’s going to be some really dumb, dumb money happening in the last 30 days because they’re going to spend it all out,” one consultant said.

Whatever contretemps were occurring behind the scenes, Harris’s team and the super PACs backing her kept banging out ads. (I counted 32 different spots on air by the end of July.) Her first policy speech in Raleigh, North Carolina, went off well, too. In proposing anti–price gouging regulations for groceries, she picked a fight with the neoclassical economics hive mind, who dutifully went to press insisting that she didn’t understand the principle of efficient markets. But she also allayed many of the fears of those who had worried that Harris, with her close connections to Wall Street and Silicon Valley, might reverse Biden’s successes in regulating big capital. 

I talked to Russ Schriefer, a seasoned Republican political strategist, around this time. He was full of praise for Harris’s campaign. “I’ve gotten up campaigns pretty quickly, you know in 24 [or] 48 hours, and they did it really well,” Schriefer said. About three weeks after Harris had tapped in for Biden, the lead she had opened in the polls was holding, and, Schriefer thought, it was clear “the fundamentals of the race” had changed. The media consultants on Harris’s side were similarly feeling positive. “You want to be optimistic, but also the trajectory is good,” said one. “It feels like the information environment could be good for the next two weeks or a week.” All eyes were on Chicago.

The Democratic National Convention began on August 19, one month to the day after Harris had filmed her last ad for the Biden campaign. A watershed was expected, and Chicago delivered. Save for the handful of uncommitted delegates I saw wandering the halls of the United Center in kaffiyehs, hoping in vain that Harris would allow them to address the crowd with a pre-vetted speech, the Democratic Party showed itself to be united behind Harris. And except for a small breach of the police perimeter on Monday, the protests against sending more military aid to Israel were likewise kept at bay. The order of the day was a tight ship, and in her speech accepting the nomination, Harris hewed closely to the script that had already been laid out in her ads and her rallies. Her speech even followed the standard progression of a media campaign, beginning with her biography, continuing into a list of contrasts with Trump, and finishing with a call to vote. The most notable difference that night was the overarching conservative tone, even as she ran through meticulously hedged lines (“We can create an earned pathway to citizenship—and secure our border”). Harris signaled she would not withhold arms shipments to pressure Israel to negotiate for peace, and she promised to “ensure America always has the strongest, most lethal fighting force in the world.” Notably, she never mentioned the anti–price gouging regulations she had touted in Raleigh a week earlier, and she spent little time talking about the problems of affordability that had been central to her ads. Some commentators read all this as a sign that Harris was addressing herself to undecided voters. If so, she had chosen a sliver of the sliver: the neocons.

In retrospect, the convention marked the end of the beginning of Harris’s campaign. The aura of possibility that had defined her first weeks in the race had evaporated, and Harris’s departures from Biden’s platform also seemed to foreshorten as the distance between his exit from the race lengthened. Typically, a polling bounce is expected. Harris’s lead in the national polls remained tight, and the people making ads for her hunkered down for a grueling few weeks. (At the start of September, more than 125 ads supporting Harris were airing in battleground states. Trump’s side was running at least another 50, some of which had started hitting Harris on economic issues, on top of a still healthy list of spots going after her over immigration.) A Democratic strategist told me about breaking a bad habit of checking polling numbers right before bed. Another was more sanguine. “Depending on how this all turns out, it could really change the way it’s done going forward,” they said. “I don’t know if you can necessarily replicate everything about it, but why would you want to run a two-year campaign when you could run a 200-day campaign?” There had been so much “mission creep” in the past decade. Harris might show this wasn’t needed. Although, if so, they added, “I know it would hurt my own bottom line.”

Around this time, I caught back up with the freelancer. Like many of the people I had met with then, he had weathered the transition from Biden to Harris and was back making ads. He’d done a couple of quick spots that had already gone online. Now, he was moving on to some new testimonials. These were in keeping with the decidedly conservative tone Harris had struck during her speech accepting the nomination. They were focused on swing voters who had decided to go with Harris, with the idea these people might offer a framework for like-minded folk to find their way to Harris as well.

The guy the freelancer had shot recently was a white man, whom they framed as generically as possible. That way, if the testimonial came out well, his ad could play to almost any audience. The goal was to see if he might polish up Harris’s credentials as a moderate. During the shoot, though, it had been hard to direct the interview into these matters. Mainly, the guy just appeared to like Harris’s energy and the joy of her campaign. It seemed so much lighter and brighter than what Trump had on offer. “But, you know, that’s not what the ad wants to be about, at least from our side,” the freelancer told me. So, during the shoot, he prodded his subject about Harris’s economic policies. The man couldn’t identify any. “He said the word ‘vibes’ at some point,” said the freelancer.