Alexis Hochleutner's wedding rehearsal dinner fell on the day the banks started falling, in October 2008. She married an equity trader the next day, then left finance for ad tech because it seemed safer. She has spent the years since learning otherwise.
Alexis Hochleutner has worked on brand safety, ad fraud and privacy at Google, Meta, Amazon, T-Mobile and HUMAN, back when it was still White Ops. Her longest run at Google was leading viewability, fraud and brand safety for the DoubleClick platform. She describes her job as saving the world one ad at a time. She says it with a straight face and then immediately admits the joke.
I asked how we get advertisers to care about fraud when most of them want cheap impressions and plausible deniability. She pushed back on the premise first: “advertisers” is not one actor, it is a chain of RFPs, agencies and programmatic teams working with imperfect metrics. And metrics, she pointed out, cut both ways. “As easy as it is to create these metrics is as easy as it is to defraud these metrics.”
Then she said the part I have never heard an anti-fraud executive say on tape.
“Any sort of tolerance for fraud is not great, but there is some impossibility. Because to some extent, you don't want bad actors to know exactly which inventory you're calling fraud. So there has to be some noise put into it. So you do kind of have to sometimes buy fraudulent inventory.”
Read that twice. If you block every bad impression the instant you detect it, the fraudster learns exactly what your detection sees and routes around it. So you let some through, deliberately, to keep him guessing. It is the logic of counterintelligence applied to a media budget. And it means some slice of every advertiser's spend that “got past” verification may have been allowed past on purpose.
It is a defensible tactic. It is also not something most clients have ever been told. The next time a vendor quotes you a fraud rate, it is worth asking how much of it was a choice.
Fraudsters thrive in silos
Her real complaint is not the fraud. It is the finger-pointing. “The biggest problem with ad fraud is that everybody's pointing fingers. We shouldn't be pointing fingers at each other. We should be pointing fingers at the bad actors.”
The proof she offered is a case she worked across two companies. Google had detected a fraudster one way. HUMAN had detected the same operator a slightly different way. Put the two together and the telemetry was strong enough to take him down. He later ran a second operation, she says, and that one brought in the FBI and put him behind bars.
“Fraudsters thrive in silos.” The industry has plenty of working groups, she says, but nobody has really solved the problem of being the first to share. Somebody has to go first. Her prescriptions: know-your-customer checks built into the programmatic supply chain, and real information sharing between companies that are, on paper, competitors.
She now works on exactly that with Matt Katz, formerly of FreeWheel, under a mission she describes as uniting stakeholders to make fraud and abuse unprofitable. When I mentioned Jounce's MFA list going public that day, she raised the obvious counterpoint from the same cat-and-mouse logic: a bad actor can now see who is on the list, and start reverse engineering why. Still, she called it a big step forward.

