Sasha Eisenberg, author of The Victory Lab and The Lie Detectives, joins Shaan to break down the science of political campaign marketing. Topics include the neighbor-shaming voter turnout experiment, how the Biden 2020 team built a “Harm Index” to prioritize which disinformation to fight, Trump’s Facebook-as-fundraising-machine insight, and why Trump’s community approach beat Obama’s list-building approach. Recorded just before the 2024 election — with observations on Harris’s TikTok moment and the Rogan question.
Speakers: Shaan Puri (host), Sasha Eisenberg (guest)
Note on This Episode
Shaan is the sole host. Sam does not appear. Guest is Sasha Eisenberg, political journalist and author of The Victory Lab (2012) and The Lie Detectives (2024).
Introduction: Campaign Operatives and Attribution Problems [00:00:00]
Shaan: I’m fascinated by the marketing machine underneath political campaigns — regardless of which candidate you support. Both sides are spending over a billion dollars trying to persuade people to push a button. The science, tactics, and persuasion techniques underneath it all. I invited on Sasha Eisenberg, who’s studied this for two decades and wrote a book called The Victory Lab.
Sasha: One of the hardest parts of reporting in this area is that when something works, the incentive is to tell the world how genius you are. When it doesn’t work, it’s “the politician had no charisma — nothing we could do.” During a campaign, nothing leaks. The day after the election, the campaign ceases to exist and everyone starts talking.
Sasha: The Obama analytics department in 2012 had 54 people. They called themselves “the Cave.” The campaign manager was literally asking who talked to reporters. Then the day after the election, Eric Schmidt was helping them launch firms and they were giving interviews to everyone. The issue: an election has a binary outcome, and no one thing shaped that outcome. It was a confluence of dozens of factors. Beware of anyone offering a monocausal explanation for why an election turned out the way it did.
The Neighbor-Shaming Voter Turnout Experiment [00:10:00]
Shaan: One of the most interesting things in your book was the neighbor-shaming tactic. Walk me through that.
Sasha: Political scientists ran an experiment where instead of sending voters a typical “please go vote” mailer, they sent a mailer saying: “Here is your voting record. Here is your neighbors’ voting record.” And voter turnout in those precincts went up dramatically. Just making voting public information — or creating the perception that it could become public — was more powerful than any other turnout intervention.
Shaan: It’s like the Elf on the Shelf. You don’t have to explain why they should vote. Just: someone’s watching. Someone’s tracking this. That simplicity is extremely powerful.
Sasha: The Cialdini connection is real. Norm-based messaging — “most people in your neighborhood voted” — consistently outperforms arguments about civic duty or policy outcomes.
The Biden 2020 Harm Index [00:20:00]
Sasha: In 2020, the Biden campaign was flush with cash. The campaign manager asked: if you had an extra $10 million, where would you spend it? The answer was: fight disinformation. But they quickly realized that “don’t let an attack go unanswered” is the wrong playbook when your attacker is a Macedonian teenager farming clicks rather than the opposing campaign.
Sasha: The old model was whack-a-mole: whatever viral story trends today, respond to it. The Biden team reframed it as a demand-side problem. Most viral content won’t change any minds. The question is: which specific narratives actually move persuadable voters?
Sasha: They built what they called the Harm Index. Three questions for every viral narrative: (1) Are you familiar with this story? (2) Do you think it’s true? (3) Would it make you less likely to vote for Biden? The x-axis was reach (how many people had heard it), the y-axis was impact (how likely it was to affect their vote).
Sasha: Hunter Biden corruption: high awareness, low impact — persuadable voters didn’t think Biden was fundamentally driven by personal financial gain. Biden’s age: high awareness, high impact — this was in the upper-right quadrant. But the focus groups revealed something important: voters weren’t worried about his physical fitness. They were worried about his political weakness. They saw him as perpetually the supporting character, never the main one. The fix wasn’t photos of him biking. The most effective ad was 15 seconds of Biden looking directly at camera, talking in a plain voice about what he cared about. Unedited, unglossed. People who were susceptible to the “senile Joe” attack just wanted evidence he could articulate his own worldview.
Trump 2016: Facebook as Revenue Center [00:35:00]
Sasha: Trump in 2016 was outspent by Clinton roughly 2:1, and yet he was spending an unusually high share of his budget on Facebook relative to TV. That’s almost unheard of at that level. The reason it happened was Jared Kushner came to Trump and reframed it: this is not an ad budget, it’s a revenue center. If you target the right people with the right message, every dollar you spend will come back as fundraising. Trump didn’t like spending money, and he didn’t use a computer — someone literally sat next to him and typed out his tweets as he dictated them. But he understood the ROI argument.
Sasha: The other shift was community vs. list. Obama had the biggest email list in politics. His theory was: collect millions of supporters, then ask them for money and volunteers. Very transactional, very one-way. Trump did something different, mostly by instinct: he retweeted his supporters. He amplified their content. That’s antithetical to how political professionals think — you’ve spent hundreds of millions testing your message down to the syllable, and now you’re going to let some random supporter speak for your campaign?
Sasha: But what Trump created was a community of people who felt invested in the campaign. MAGA online became a place where people got reinforcement, recognition, a feeling of participation. Obama never gave his supporters that. He collected them. Trump made them members.
The 2024 Election — Harris vs. Trump [00:50:00]
Shaan: This is being recorded just before the 2024 election. How do you see the two campaigns?
Sasha: Trump for most of the year was running as good a campaign as he could run. More recently I see a mismatch between their stated strategy and their execution. Their strategy depends heavily on turning out non-voters — young men especially, young men of color. That kind of voter activation requires high-quality face-to-face contact from volunteers who come from that community. They’ve outsourced a lot of that to Elon Musk’s America PAC, which uses paid day laborers, not volunteers. That’s a terrible model for turning out infrequent voters. You need socially meaningful interactions — “where’s your polling place, here’s how this affects you” — not a paid stranger at the door.
Shaan: What is Musk doing with the petition thing — paying people to sign petitions?
Sasha: They pay you some bounty to get your friends to sign a petition supporting the First and Second Amendment. The theory is: collect contact info, then target those people with digital ads and door-knocking. It’s a reasonable first step in a multi-step process. The problem is that the second, third, and fourth steps require a very precise, well-targeted operation, and I haven’t seen evidence they’ve built that.
Sasha: Harris is running a far more traditional Democratic campaign. She built a campaign from scratch in four months, which is remarkable. She went viral on TikTok immediately — the “coconut tree” jokes, the JD Vance song, clipping every crazy thing Trump says. The Biden team was allergic to giving Trump oxygen. Harris’s team was unshackled.
Shaan: Was it a mistake for her not to go on Rogan?
Sasha: I think her team was still building the infrastructure for those kinds of conversations. Going to The View and Call Her Daddy in one week was a moment, but it was also a signal of how little she’d done early on. Rogan said come to Austin, three hours unedited. One reaction: what a missed opportunity with 30 million listeners. The other: how dare he have conditions, she’s the Vice President. My read: they were probably afraid the downside of a bad two-hour conversation outweighed the upside of the audience.
Business Models in Political Marketing [01:05:00]
Shaan: What’s the most insane business model you’ve seen inside this industry?
Sasha: Media buyers. They get a commission — historically around 15% — off every ad placed. Buying television doesn’t scale with effort: buying twice as many ads in Wisconsin takes no more work than buying half as many. There’s only so many TV stations in a given market. The commission is the same. It’s been that way for decades. Some of the bigger firms are doing revenue in the hundreds of millions.
Shaan: What’s an underrated opportunity you’d invest in if you were building a political tech business?
Sasha: Out-of-cycle voter communication. The best period for Trump may have been 2021-2022 — he was out of the news, Democrats had stopped attacking him, and nostalgia for the Trump years was building in a vacuum. If you’d kept reminding people what 2020 actually looked like, you might have prevented that. Political money disappears between election cycles. But there’s a real opportunity in party-level branding — not candidate advertising but building the long-term brand equity of the party itself, in off years, when people aren’t being bombarded.
Sasha: Also: democratizing the sophisticated tools that presidential campaigns use. Twelve years ago, only presidential campaigns could afford serious data modeling and voter targeting. Now a county executive candidate could theoretically use those same tools. But the gap between what’s available and what a volunteer campaign manager can actually deploy is still too large. There’s probably a business in packaging those tools for small-scale campaigns.
Shaan: The book is called The Lie Detectives. Where should people find you?
Sasha: Sasha Eisenberg dot com — links to both books. The first one was about the global sushi business, which may also be interesting to your audience.