Looking back at where Democrats went wrong in 2024, Danielle Butterfield thinks a lot about that June debate. Sure, President Joe Biden’s performance was disastrous. But what compounded that catastrophe, to Butterfield’s mind, was what Democrats did, or rather didn’t do, next.
In early July, even as the right’s muscular media apparatus filled the internet with Biden’s most excruciating on-stage moments, Democrats’ persuasion ad spending online dropped dramatically. They were still fundraising plenty, but ads attempting to sway voters to their side — or at least away from Donald Trump’s — mostly stopped.
There was a reason for that. Future Forward, the dominant Democratic outside funder which raised some $950 million between its super PAC and other entities, had gained prominence for its aggressive approach to ad testing. Run by a small inner circle of number crunchers, the firm commissioned more than 1,500 ads throughout 2024 and, with the methodological rigor of a drug company testing a new vaccine, ran each one through randomized-control trials, surveying millions of voters to determine which ads would be the most persuasive to the most people.
Future Forward’s tests spelled out a pretty consistent theory of the case: Elevating and contrasting Biden with Trump was more persuasive to voters than attacking Trump outright. But the debate scrambled this strategy. How to promote a candidate who had all but self-immolated on stage? And was it worth spending money now when it was anyone’s guess what the next few months would hold? Without something nice to say about Biden, for a few critical weeks, Future Forward didn’t say much at all.
To Butterfield, that was a wakeup call. Democrats, she believed, were allowing data about what supposedly sways people to stand in the way of intuition. For all their parsing and fine tuning, she felt her party was failing to simply read the room. “Trump was attacking us from every angle, and we were not doing anything,” Butterfield said. “We were really letting data drive a decision that should be pretty crystal clear.”
Butterfield is hardly an uninterested party. Since 2021 she has served as executive director of Priorities USA. Once the go-to Democratic super PAC behind Barack Obama’s 2012 and Hillary Clinton’s 2016 presidential campaigns, Priorities was effectively sidelined in 2024. Future Forward received the Biden campaign’s blessing, and Silicon Valley megadonors, drawn in by the group’s data-driven approach, funneled their fortunes into Future Forward’s giant pot. In the end, Future Forward’s PAC spent more than $500 million on ads — nearly 70 percent of all presidential super PAC spending from Democrats according to one count. Priorities, by contrast, spent about $2 million.
In the aftermath of the election, there has been plenty of scapegoating and fingerpointing directed Future Forward’s way. Grassroots groups say it should have spent money earlier. The Harris campaign has grumbled that it focused too much on economic issues and not enough on casting Trump as a threat to democracy. After Tuesday’s elections, the Democratic National Committee is expected to put out its own autopsy, arguing that the party didn’t pay close enough attention to the issues voters cared about. Future Forward, meanwhile, has pushed back on the idea that concentrating on democracy over the economy would have won the day. Future Forward declined to comment for this story.
But Butterfield argues there’s a more fundamental flaw permeating the party — not just Future Forward — and is pitching donors on a different way of doing things as she tries to mount a comeback of sorts for Priorities. It boils down to this: Over the last decade or so, Democrats have leaned too far into the science of campaigning. Now, it’s time to get back to the art.
For too long, Butterfield said, Democrats have taken ad tests as gospel. These tests, which take place online, split large groups of voters into treatment and control groups. The treatment groups are shown an ad; control groups aren’t. Afterward, both groups answer surveys about who they plan to vote for. The difference between the two groups signals how persuasive an ad is. While Future Forward relied on a range of different types of surveys and tests, the majority of them worked this way. These tests give donors and strategists some confidence that their messages are actually moving people. But “a lot of that confidence is faulty,” Butterfield argues.
Getting people to watch and give their opinions about an ad in a vacuum bears little resemblance to the way people actually consume content online. These tests don’t measure whether people will grasp a candidate’s message when they’re half-listening to a podcast at work. They don’t tell you if people get the ick when a slick, highly-produced political spot pops up in the middle of a video of a guy ranking fast food french fries. They don’t tell you if an ad is really “breaking through,” Butterfield said, a hand-wavy term she uses both liberally and intentionally.
“We try to treat advertising like this, like black and white, measurable thing, but in the world we’re dealing with, it’s not,” she said.
Now, Butterfield is attempting to transform Priorities into a different kind of super PAC, focused less on plowing millions of dollars into ads every four years, and more on funding around-the-clock experiments in digital advertising. This year, it plans to invest $8 million in those experiments. That includes in Pennsylvania, where Priorities has been working on digital ads leading up to Tuesday’s contentious Supreme Court race. It’s a trial balloon of sorts for Priorities’ evolving approach in a state that will be crucial in 2028.
As part of its broader shift, Priorities has launched a new for-profit company called Flock, which sells deep dive media consumption research to other groups within the party. It’s conducted one-on-one interviews to get inside the minds of young men on YouTube and is marketing a tech tool called Ad Hawk, which shows Democrats what their opponents’ digital ads are saying and whom they’re targeting. Priorities is running its own tests, too, focused on less traditional categories of political ad targeting like geography and demographics. Instead they’re testing, say, pushing ads toward people who watch a lot of right-wing podcasts.
It’s not that Butterfield is anti-data. One great irony of what Priorities is proposing is that it relies heavily on data — just more and different types of it. Democrats, Butterfield argues, need to be asking a far broader variety of questions, like how effective is an ad at driving people to search? Does it matter whether they watch it or just listen to it in passing? What are the most popular podcasters saying about the issues on their shows? How are people reacting in their comments? And, at a time of shifting demographics are Democrats better off targeting people based on their media habits than their zip codes?
If this sounds like a hodgepodge of unproven ideas pulled straight from the throw-shit-at-the-wall school of politics, that’s because it kind of is. Butterfield’s view is that Democrats should be taking a lot more risks, even if that means stomaching — and paying for — a lot more failed experiments.
This story is based on interviews with eleven Democratic strategists familiar with Future Forward’s and Priorities’ work, some of whom were granted anonymity to speak about confidential matters.
The issues Butterfield is raising cut to the heart of some long-simmering debates in Democratic politics. It’s a debate over the difference between online engagement and actual persuasion, a debate over going broad and going deep. And it’s a debate over how much advertising can really achieve at a time when the Democrats’ fortunes are as bleak as they’ve ever been.
Butterfield got her start in politics on Barack Obama’s 2012 campaign, when a single line on her resume noting her experience with Google ads landed her on the digital advertising team. She soon fell in love with the left-brain-right-brain part of political advertising, which deals as much in dollars and cents as it does in emotion.
It was during that election that Democrats’ data obsession was arguably born inside a windowless room in Obama’s Chicago headquarters known as “the cave,” where a crew of quants — quantitative analysts — used reams of data to predict how persuadable any given voter was. Columnist Peggy Noonan once referred to their approach as “politics as done by Martians” in The Wall Street Journal. She did not exactly mean it as a compliment, but after Obama won, even Noonan had to admit the “Martians” were “reinventing how national campaigns are done.”
Future Forward launched in 2018, two years after the Clinton campaign, which Butterfield worked on, found itself outmatched by Trump’s Facebook-fueled operation. It was founded by two staffers from Obama’s 2012 campaign, as well as the group’s current leader, Chauncey McLean, who worked as director of media tracking for the Democratic National Committee in 2012, before doing a stint in the corporate world.
From the start, Future Forward’s commitment to ad testing was perfectly matched to the way the party’s wealthiest benefactors, who increasingly came from the tech industry, liked to spend their money. Facebook co-founder Dustin Moskovitz and his wife Cari Tuna, some of the party’s biggest donors, are leaders of the “effective altruism” movement, a philanthropic philosophy that involves scrupulously maximizing impact per dollar. In 2020, Moskovitz donated at least $20 million to Future Forward.
Throughout 2024, Future Forward worked closely with the Democratic data firm Blue Rose Research, which tested thousands of ads for both parties, including ads Future Forward commissioned and ads made by other groups. Those tests found that the ads with clear economic messages about lowering the cost of living consistently proved the most persuasive to the most people. Future Forward cut its own ads from an array of ad makers in eight languages and produced different versions to appeal to different demographics. It spent heavily online, outspending Republicans in battleground states two-to-one on digital platforms. But for the most part, Future Forward’s ads stayed relentlessly on message, with 71 percent of them focused on the economy, according to a Future Forward adviser who was granted anonymity to speak with POLITICO Magazine because they didn’t have permission to go on the record.
“This was not a slice and dice election,” McLean, who rarely gives media interviews, said on a podcast in March. “We saw consistently that cost of living, lowering costs, [and] inflation were the driving concerns across every single sub group.” McLean declined to comment for this story.
Future Forward published its ad test results in a newsletter called Flight Radar. It was an effort to lend some transparency to the process, but pitting ads from different partners against each other and funding the top performers rubbed some in the party the wrong way, drawing comparisons to “The Hunger Games” in The New York Times.
It also gave people the impression — the wrong impression, McLean has since said — that the ad tests were the only thing that mattered. “That’s not how the process worked at all,” he said on the podcast. “If I have 1,000 ads, I’m going to look at the top 100 as all usable.” Future Forward would then compare those results to other data points like social media engagement, surveys and focus group results to determine a course of action.
“It didn’t always produce the happiest partners, but it did, we think, produce the best results,” McLean said.
There’s a strong case to be made, of course, that this was a sound strategy. Elections are a game of numbers. Why not try to harness every penny in service of things the most possible voters are explicitly telling you they care about?
But others viewed things differently. Just because an ad persuades the broadest number of voters doesn’t mean it resonates deeply with anyone in particular. “You’re assuming whatever works for the broadest group is going to work for that smaller group you need to move. That’s just not the case,” said Anna Scholl, president at the progressive advocacy organization ProgressNow.
In some cases, this approach can achieve just the opposite effect, producing messages that Butterfield refers to as “low-fat vanilla” — broadly palatable, but “they don’t actually stick with anyone.”
Priorities saw this first-hand with its own ads, Butterfield said. In one example, the group created an ad designed to look like an influencer-style “getting ready with me” video focused on abortion rights. In a world where younger voters have hypersensitive BS-detectors, Priorities wanted to see whether ads that fit seamlessly into people’s feeds would be more convincing than traditional ads that can look alien online.
Priorities targeted YouTube users who were interested in beauty with both the influencer ad and a traditional one. Using YouTube’s own built-in survey tool, they found that the getting ready with me video was significantly more persuasive. And yet, in Future Forward’s tests, which surveyed a broader cross section of voters, the same ad flopped. To Butterfield, that illustrated what ad tests miss. The goal wasn’t to move a big group of people a little; it was to move a small group of people a lot.
Another point of contention was Future Forward’s decision to spend so much of its money so late in the cycle. That decision, too, was supported by the evidence that suggests the persuasive effect of ads dwindles over time. There was also no telling what might happen between spring and fall to change the stakes of the race. Even in an election that cost billions of dollars, every dime Future Forward spent in May would be one it couldn’t spend in October, which is when Future Forward made its biggest advertising push.
But that decision came at a cost as grassroots groups trying to build and secure the Democratic coalition struggled for funding. “It set everybody up for failure,” said Tatenda Musapatike, founder of the non-partisan group Voter Formation Project, which was focused on registering and mobilizing low-propensity voters of color through digital ads. “Let’s say that programs are most efficient in the last two weeks before the election. How is an organization supposed to prepare to do that most efficiently if they don’t have the money in hand, sooner?”
In May of 2024, according to an email exchange viewed by POLITICO Magazine, Future Forward turned down Voter Formation Project’s pitch to fund an early, digital voter registration drive focused on Black and Latino voters in battleground states. “Bringing nonvoters of color into the electorate requires more than pushing an ask a few weeks before Election Day,” Voter Formation Project’s pitch read. But Future Forward deemed the group’s estimated cost-per-vote — $64.38 — to be too high and said the existing research didn’t support making investments early, according to the email exchange. Instead, Future Forward said it would consider supporting an ad campaign focused on the final three days before registration deadlines.
Voter Formation Project did receive $250,000 from Future Forward in mid-September, and another $1 million in early October. With few other pots of money to turn to, Musapatike downsized her plans and wound up reaching a fraction of the voters she had originally set out to target. Early this year, Voter Formation Project shut down altogether. But in its annual report it showed that, among the small group of voters it was able to reach, those who saw its ads early turned out more than those who saw them late.
The Future Forward adviser said the organization distributed money generously to other groups and partners — more than $200 million over the course of the race — but said Future Forward was equally rigorous about asking people to prove their programs would move the needle. “They didn’t want to put a dollar toward doing something that wasn’t going to achieve that goal. Otherwise you end up in a rabbit hole of just giving to people who sound interesting,” the adviser said.
While all of this was happening, Republicans were taking a markedly different approach. If Democrats were being caricatured as a bunch of ivory tower eggheads message-testing themselves into marble-mouthed oblivion, team Trump seemed to have tapped straight into the main artery of internet culture.
It’s not that Trump’s campaign and surrounding super PACs weren’t monitoring the data. An early 2024 survey reportedly convinced the Trump camp to try to peel off young Black and Hispanic men and convinced Trump’s super PAC to try to reach them on streaming TV. But while they were attuned to the data, Butterfield said they weren’t as strictly bound by it either.
One example of this she frequently points to is the infamous “They/Them” attack ad, which Priorities tracked. It featured Harris talking in 2019 about paying for gender-affirming surgery for trans prisoners. The ad began life as a bit of opposition research, posted to the Trump War Room account on X. The post gained steam online until, in September of 2024, the Trump campaign ran it as an ad. The more pickup the ad got — in the news and on podcasts like The Breakfast Club — the more money Trump’s camp put behind it.
“That ad didn’t come from a poll telling them to talk about that clip,” Butterfield said. “They used the internet to signal to them that it was effective rather than a poll or a focus group.”
Democrats never really responded to the claims in the ads, though. Future Forward could see that the ad was effective. But according to the Future Forward adviser, the group decided that it was up to the Harris campaign itself, not an outside super PAC, to decide what kind of stand to take on the issue. According to The New York Times, the Harris campaign did craft a rebuttal ad, but it fell flat in ad tests and was shelved. And so, in another moment that Butterfield likens to the post-debate period, Democrats left one of Trump’s stickiest digs effectively unanswered.
After Trump won, one post-election poll found that, of all of the issues, from immigration to Israel, voters thought Harris had focused excessively on transgender topics. In reality, she hadn’t said much about those issues at all. It was Trump’s team doing the talking.
In April of this year, as Democrats were still crawling their way out of November’s wreckage, the party got another chance to prove its value to voters — this time, in the Wisconsin Supreme Court race. It was seen as an early litmus test of voters’ views on Trump’s chaotic first few months. It was also an early test of Priorities post-2024 plan.
In the months after Harris’ loss, much was made over the demographic shifts in the party. Young voters, Hispanic voters and Black voters — all central to the Democratic coalition — had moved substantially to the right. But Priorities’ post-election polling revealed an equally staggering swing: Between 2020 and 2024, people who consumed at least three hours of YouTube a day had moved 10 points toward Trump — a bigger swing than Black voters, Black men, young voters and a slew of other categories that were dominating headlines.
“The driver of behavior in ‘24 was so much more about your media consumption,” Butterfield said.
After the election, Butterfield and her team launched a new media research project called Warbler, housed under a new LLC. Warbler’s focus is on better understanding how, where and why people are spending their time online and using it to inform Democrats’ strategy.
In Wisconsin, concerns about voters simply disengaging as they did in the presidential race were top of mind, said Chris Walloch, executive director of the Democratic state PAC, A Better Wisconsin Together, which partnered with Priorities on its digital strategy. “We needed to make sure that we were communicating with all audiences and that we’re not leaving folks off the table,” he said.
The Warbler team went digging into its media consumption data and found that the Wisconsin voters who said they hadn’t heard anything about the election yet were spending a lot more of their time on audio streaming platforms like Spotify and Pandora. That finding drove Walloch to shift some of the PAC’s digital outreach to audio streamers in hopes of reaching potential voters who would have been otherwise overlooked. “They have been an indispensable partner,” Walloch said of the Priorities team, though he acknowledges it’s difficult to know how big of a difference those ads made to Democrat Susan Crawford’s ultimate victory.
More recently, in Pennsylvania’s Supreme Court race, Priorities worked with the statewide communications hub, Commonwealth Communications, to come up with an ad campaign to define the opposition. Unlike in Wisconsin, Pennsylvania voters will be asked to vote “yes” or “no” to retain the current slate of Democratic judges. A no vote would require new appointments to go through a confirmation process that could leave the court in a partisan deadlock until new elections in 2027. In the absence of an actual opposition candidate, said J.J. Abbott, executive director of the state group, “one of the hardest tasks was to try to define for people who ‘no’ is.”
A traditional approach to that assignment might focus on issues like abortion or voting rights, which are all at stake in the race. But throughout the summer and early fall, Priorities’ data consistently found that the Epstein Files were the one political topic that continued to break through online. That led Priorities to create an ad about the shadowy powers of the billionaire class, drawing a straight line between Trump, Epstein and the Pennsylvania megadonor Jeffrey Yass who’s backing the opposition. The ad doesn’t make any claims about Yass, but it pushes the idea that money in politics corrupts. “With money and power, who knows what you can get away with,” the voiceover intones.
Like the “They/Them” ad before it, it’s a bit of a conceptual stretch, and it didn’t perform particularly well on ad tests. But amid a crowded field of ads featuring stale B-roll of judges in robes, Abbott said it’s worth leaning into a message “that’s catching people’s attention.” In the first few days the ad ran, Priorities said, it performed better on YouTube surveys than other more traditional ads about the race, and users were twice as likely to actually click through. Whether that makes any difference on Tuesday still remains to be seen, but it’s one of many tests to come of Priorities’ new strategy.
Beyond their work with individual campaigns, Warbler is also producing a steady stream of research to help guide players in the party. In one recent briefing, the Warbler team walked other Democratic groups through the actual YouTube watch history of 1,000 survey participants. In another, they presented anonymized video interviews with a cross section of voters — including Trump voters — about why they like the podcasters they do and what their biggest turnoffs are on YouTube. Overly polished and high-production videos? Out. Explainer videos with catchy thumbnail images? In. The Priorities team has asked much the same questions in focus groups of creators, trying to get to the bottom of what excites and enrages their audiences.
The goal in all of this, Butterfield said, is not to yield one definitive strategy, but to arm political operatives with a lot more information. “No singular data source is ever going to tell the whole story,” she said.
Priorities is aggressively testing the value of targeting ads based on people’s online behavior, not their demographics or geography. This work is based in part on a test Priorities conducted in partnership with the Latino research firm Equis Research in 2022. In that study, they targeted an ad for Arizona senator Mark Kelly at a few different groups of users, dividing them based on the types of videos they watch — from mainstream news consumers to fans of far-right podcasters like Candace Owens. They named this latter group “right hive,” and in a surprise finding, discovered that, of all of the groups they studied, people in the “right hive” group were substantially more persuaded by the Kelly ad than any other cohort.
“Traditional thinking is that’s not an efficient use of money, because those people aren’t with us,” Butterfield said. “But we’re actually seeing an effect.” Priorities now hopes to run more experiments like this in the future.
Of course, all of that depends on other people — most notably donors — buying what Butterfield is selling. “It’s part of why we’re trying to tell this story,” she acknowledges openly. “We need to act as that incubation lab for new ideas, and we need donors to subsidize that.”
But convincing donors to do that may be a challenge, particularly at a time when the left’s big money donors, feeling burned from 2024, are withholding funding across the board. As one major Democratic fundraiser, who was granted anonymity to discuss private conversations, put it, “We do like being able to look at the numbers. We do like being able to see an ROI.”
Just because Harris lost doesn’t mean there isn’t a good deal of post-election evidence on Future Forward’s side. For one thing, she lost by less in the battleground states where her campaign and Future Forward were fighting. As Brookings put it in its analysis, that could be evidence that “the more voters saw of the campaign, the less they were inclined to move away from the Democrats.” No one’s going to slap that on a bumper sticker, but it is at least some evidence of impact.
At a wildly precarious time for the party, when the Democratic brand is badly damaged and the Trump administration is dismantling the federal government and arguably stretching the rule of law, it’s a risky thing to place bets on unproven ideas that may wind up failing.
There’s an argument to be made that what Democrats need now is not tinkering around the edges with new ad targeting categories on YouTube or new insights from creators, but a wholesale rebrand. Ads, after all, can only accomplish so much when millions of people see the underlying product as fundamentally defective. There are those who are understandably dubious that any number of minor tactical changes would have altered the outcome.
Butterfield doesn’t disagree that a broader rethink is in order. But, she argues, now in the midst of that rethink, is precisely the right time for testing new ideas so that two years from now and two years after that, when billions of dollars are on the line, Democrats won’t have to learn those lessons on the fly or repeat the mistakes of the past. Even so, she knows it’s a lot to ask. “We’re basically going out there saying, ‘Hey, what we’re doing is immeasurable. Give us money anyway. We can’t tell you if what you’re giving to us is going to be successful or not. Give us money anyway,” she said. “Believe in us.”
