I hadn’t heard from Rick Bruner in years. A great guy, if I remember correctly. Then suddenly—boom—there he is in AdExchanger, dropping a long sermon about how randomized geo experiments are the One True Way to measure advertising. It read less like an op-ed and more like a religious revival, complete with the kind of certainty that makes CMOs weak in the knees and consultants dust off their retainers. If you didn’t know better, you’d think he’d returned from the desert carrying stone tablets with “Thou Shalt Randomize” carved on them.

This is the kind of article adtech loves because it offers a messiah in a business that thrives on confusion. You’ve got endless dashboards, endless metrics, and endless self-congratulatory “insights” that rarely match the sales reports. Then along comes Bruner with a promise of salvation: forget the noise, forget the lipstick-on-a-pig models, just run randomized controlled trials and your answers will be clean, honest, and scientifically unimpeachable. It’s bold, it’s seductive, and—like most things in this industry—it’s only partly true.

Let’s be clear: advertising measurement is broken. Always has been, always will be. Attribution, the backbone of digital for two decades, is the world’s worst magic trick. You dump billions into Google and Facebook and then let them tell you, in their infinite generosity, that—surprise!—your ads worked.

They’re essentially selling you the gun, the bullets, and the crime scene cleanup, then handing you a report saying you’re the hero.

Multi-touch attribution? It’s the same nonsense, just with fancier charts and longer acronyms. It’s correlation pretending to be causation. It doesn’t measure incrementality—it just spreads credit like butter on Wonder Bread.

And then there’s MMM—Marketing Mix Modeling—the great old dinosaur lumbering across CMO PowerPoints. Sure, it gives you a sense of the forest, but the trees? Forget it. By the time an MMM spits out results, you’re on a new campaign, the budget cycle has shifted, and you’ve already fired the intern who collected the data. It’s glacially slow, hopelessly coarse, and about as useful for tactical decisions as a weather report written six months after the storm.

Bruner is right to call this a crisis. Billions are being wasted every year because the industry refuses to admit its tools are fundamentally flawed. We are operating on junk math, and everyone knows it.

This is where Bruner’s gospel comes in. Randomized Controlled Trials (RCTs) are the “gold standard.” Everyone knows it. If you want to prove causality, you randomize. You split your audience into test and control, you expose one group to the treatment, and you measure the difference. It’s basic science. It’s how drugs are tested, how policies are evaluated, how every serious researcher avoids being laughed out of the room.

And in a world where privacy laws have ripped user-level tracking out of marketers’ hands, RCTs start to look like the holy grail. Enter the geo experiment. Instead of tagging individuals and risking another trip to the GDPR penalty box, you just carve up the country into Nielsen’s 210 DMAs, randomly assign them into test and control, and measure the sales lift. No cookies. No IDs. No clean rooms. Just math.

The pitch is intoxicating:

For CMOs on the edge of being fired if they can’t prove ROI, this sounds like deliverance. No more trusting vendors. No more shrugging at the CFO. Just hard numbers that let you say, with a straight face, “Yes, that $50 million we poured into connected TV wasn’t lighting cigars for Roku execs—it actually moved product.”

This year more than ever, measurement is existential. AI is the new god in marketing, and like any god, it’s only as powerful as the sacrifices you put on the altar. If your training data is garbage, your AI becomes a garbage factory—churning out optimizations that look brilliant in dashboards but translate into nothing at the register. Garbage in, garbage multiplied.

On top of that, privacy walls are everywhere. Regulators in the U.S. are sniffing around, Europe is still swinging the GDPR hammer, and California thinks of privacy legislation the way Vegas thinks of buffets—never enough. User-level tracking isn’t just dead; it’s radioactive. Anyone still pushing pixel-based attribution at scale is either delusional or running their business like it’s 2012.

So when Bruner shows up saying, “You don’t need any of that—just randomize across DMAs and you’re free,” he’s striking a very raw nerve. CMOs are desperate for independence, for methods they can defend in boardrooms, and for results that don’t look like they were drawn up by a drunk magician.

And let’s give him credit: he’s torching the right villains. Attribution is a fraud. MMM is too blunt. Platforms have been running measurement theater for years, trotting out “lift studies” that make snake oil look like Nobel-winning science. Advertisers are too quick to swallow the lies because inflated numbers make everyone look good—until the CFO asks why revenue isn’t budging.

RCTs and geo experiments, for all their flaws, do offer something radical in this industry: independence. They let advertisers test without bowing to the platforms’ black-box metrics. They give CMOs a fighting chance to say, “No, actually, we ran the experiment ourselves—and this channel really did move sales.” In an industry hooked on vanity metrics, that feels almost revolutionary.

So yes, Bruner’s gospel resonates. He’s right to call out the junk math, right to say attribution is lipstick on a pig, right to argue that MMM is too slow, and right to remind us that causality beats correlation every damn time.

But let’s not pretend this is a silver bullet. Randomization might be the gold standard, but gold standards in advertising are like unicorns—they’re mythical, everyone claims to have seen one, and if you’re lucky enough to catch it, you’d better check it for fleas.

RCTs and geo experiments are tools, not salvation. They can clean up a lot of the mess. But in an industry addicted to easy answers, don’t be surprised if half the people who claim to be running them are just selling a shinier version of the same old snake oil.

A highly relevant reminder came in June 2025 from Richard Oppy, Marketing VP at Carlton & United Breweries (a division of AB InBev), who posted on LinkedIn:

“50% of my marketing money is wasted. Which half?”

Oppy’s remark, borrowing Wanamaker’s century-old line, is more than just a wry quip—it’s an indictment. Even inside one of the most sophisticated, data-heavy marketing machines on the planet, uncertainty still reigns. Advanced attribution models, geo RCTs, and dashboards stacked ten layers deep haven’t solved the essential mystery: what spend actually drives incremental sales, and what spend just props up vanity metrics.

The resonance of this quote is exactly why it stings. It underscores that no matter how many millions are poured into analytics, the “meta-problem” persists: marketers don’t know which half of their budgets are pulling weight. The industry may be awash in dashboards and regression models, but the humility baked into Oppy’s admission is what’s missing from most boardrooms.

For CMOs, operations leaders, and data scientists alike, the message is blunt: we’re still playing a guessing game with billion-dollar stakes. And until the culture shifts to face uncomfortable truths, measurement theater will keep selling tickets.

If you’ve read this far and think you’ve got the whole story—you don’t.

What you’ve seen so far is the surface. The sermon, the altar call, the seduction of randomization. But the rest of this issue goes places most of the industry doesn’t want to.

Inside, we break down how geo experiments can rescue billion-dollar channels one week and deliver false negatives the next.

We show you the operational messiness—the scale problem, the underpowered tests, the external shocks—that turn the supposed “gold standard” into a very fragile one. We pull quotes from Nielsen’s 2025 report and Richard Oppy at AB InBev that will make you rethink what’s actually driving ROI.

And then we go deeper. We expose the real problem: the addiction to measurement theater. The way CMOs buy whatever model flatters them most. The way AI amplifies garbage data into billion-dollar mistakes. The way certainty trumps science because nobody wants to be the person who admits half the budget is wasted.

This is the part where you realize the fight isn’t technical—it’s cultural. No methodology can fix cowardice. No geo split can solve vanity metrics. And no CMO wants to tell the CFO that beloved channels aren’t actually working.

Or, as a senior exec at Haus told me off the record: “The problem isn’t whether we can measure incrementality—we can. The problem is nobody actually wants to know the answer.”

If you’re not reading the full issue, you’re missing the sting. You’re missing the parts that will never make it onto a conference stage, the truths buried under keynote slides and “case studies.” You’re missing the receipts.

And if you’re fine with that? Well, then you’re not the audience.

But if you want the buzz—this is what ADOTAT+ delivers.