The Vetting Function Nobody Owns

I don't usually reach for a Senate race to make a product management point. But the last few weeks in Maine handed me a case study so clean I'd have rejected it as too on-the-nose if a business school wrote it.

Quick version, if you haven't been following: a Senate candidate got recruited and packaged as the working-class, authentic alternative to the establishment pick. He won his primary with 72% of the vote. Twenty-nine days later he was out of the race entirely, after a string of disclosures that included a credible sexual assault allegation, and a Nazi tattoo, and a biography that turned out to have some very expensive gaps in it. I'm not going to relitigate the politics of it here, that's not what this newsletter is for. What I want to talk about is the mechanism, because it's a mechanism I've watched play out in product launches more times than I care to admit.

The product had product-market fit. That was never the problem.

Here's the thing people keep getting backwards about this story: it wasn't a failure to find demand. The demand was real. Small-dollar donors loved it. Online activists loved it. The primary electorate loved it, 72% worth of loved it. If you were only looking at the metrics that were easiest to see and easiest to celebrate, you'd have called this a slam-dunk launch.

That's exactly the trap. Those numbers were measuring enthusiasm in a beta cohort, not readiness for general availability. Primary voters, online donors, and the terminally-online commentariat are not the same population as a general election electorate, and treating strong signal from the first group as validation for the second is the oldest mistake in the product manager's playbook: falling in love with a proxy metric because it's the one that's easy to rally around. Engagement isn't retention. A viral fundraising page isn't a satisfied customer. If the metric you're optimizing for isn't the metric that actually determines whether the thing survives contact with its real market, you will ship something gorgeous that dies on arrival, and you'll be genuinely confused about why, because your dashboard told you everything was great.

Nobody ran QA, and the person who was supposed to had a conflict of interest

This is the part that should make every PM in the audience wince. The consultant who recruited the candidate later admitted he'd spent only a few days vetting him. Days. On a U.S. Senate campaign. And it came out afterward that this same consultant had previously been barred from another campaign over multiple complaints of misconduct, a fact that, in hindsight, should have disqualified him from being the person anyone trusted to catch red flags in someone else's background.

Sit with that for a second, because it's not just "they skipped QA." It's that the QA function was staffed by someone who had every incentive not to look too hard, and nobody upstream asked who was checking the checker. I've seen a milder version of this in plenty of companies: the person running user research also owns the roadmap they're supposedly testing, so surprise, the research keeps validating the roadmap. When your vetting function reports to, or is staffed by, someone with a personal stake in the outcome, you don't have vetting. You have theater with a clipboard.

Skip the graduated rollout at your own risk

Every mature product org understands staged rollout instinctively: alpha with the internal team, closed beta with friendly users, open beta, then GA. Each stage isn't just a lower-stakes rehearsal, it's a data-generation mechanism. You learn things at each stage that you cannot learn any other way, because you're watching the thing perform under real, if bounded, conditions.

Politics has the exact same ladder and everyone in this story stepped over three rungs of it. This candidate had never held elected office beyond a planning board appointment before jumping straight to one of a hundred national seats. Compare that to Zohran Mamdani, who won a state assembly seat first and built an actual legislative record, including getting tested publicly and repeatedly by bad-faith attacks, before running for mayor. That's not luck, that's a longer rollout with real telemetry at each stage. You can't tell whether someone performs under pressure until you've put them under pressure and watched, and a viral clip is not pressure, it's a highlight reel.

The buyer wasn't the user

Last thing, and it's the one I'd hammer hardest if I were giving this note in a product review. A huge share of the money and momentum behind this candidacy came from small-dollar donors scattered across the country, giving twenty bucks a pop for the emotional payoff of feeling like they'd struck a blow against the establishment. That's a real transaction, but it's not the same transaction as the one made by the people who actually had to live with the result: the state's own voters, who don't get a refund if the product breaks in production.

That's a buyer/user split, the exact pattern that kills enterprise software rollouts when the person who signs the check and the person who has to use the tool every day are different people with different incentives. The buyer gets the satisfaction of the purchase decision. The user gets stuck with whatever ships. Every PM who's watched a beloved champion's pet project get force-fed to a user base that never asked for it knows exactly how this story ends, and it isn't with the users writing five-star reviews.

None of this required a crystal ball. It required someone with authority over the launch to ask "who's actually validated this, and against what population, and who benefits if the answer is nobody?" That's not a political question. That's Tuesday, in product management. We just don't usually watch it play out with this much at stake, or this fast.

If you're the person who skips the beta because the demo went great in the room, I've been you, and I've got the postmortems to prove it. Tell me where I'm wrong at prodbistro.com. The comments are open and I read them.