It’s bloated. It’s underfunded. It’s increasingly irrelevant to Gen Z. It’s pitching AI banners while begging buyers not to run away to the walled gardens. And in the middle of this storm, three men with wildly different vantage points just tried to talk it out.

Paul Bannister, Chief Strategy Officer of Raptive, who’s spent more time trying to monetize the web than most of us have spent online.

Andrew Eifler, the thoughtful optimist from TripleLift who wants to build AI agents that don’t just assist your team—they replace it.

Andrew Casale, the Canadian transplant running Index Exchange, who casually suggests that most of ad tech’s current value props should be priced like Visa swipe fees, not hedge fund commissions.

These are not your average panel warriors. They’ve built things. Scaled real companies. They’ve survived the banner ad apocalypse, the waterfall wars, the cookiepocalypse, and the rise of a hundred LUMAscape logos now lost to history. And they’re not trying to sell you vaporware. They’re here to warn you the water’s rising—and you better know how to swim.

So let’s unpack what they said, what they didn’t, and what’s making them quietly panic.

Paul Bannister said it first, and he said it for all of us. “I’m so tired of talking about cookies.” You could hear the exhaustion from a decade of panels, promises, and half-baked replacements that never quite worked.

And he’s not wrong: the ad industry has been declaring the death of the cookie longer than we’ve been pretending CTV metrics make sense. And yet here we are—still baking, still tracking, still panicking. Paul doesn’t see salvation in blockchain (don’t even start), and frankly, neither does Eifler, who believes the whole conversation is stuck because no one wants to rip off the Band-Aid.

Eifler’s point is sharper: the only way to truly fix the open web is to automate the stupid out of it. His take? Half the jobs in programmatic are people doing menial work that software should’ve handled five years ago. Enter the “agentic AI” phase: actual task-based bots doing planning, QA, and optimization. Not just LLMs pretending to be your copy intern, but real performance machines embedded into the tech stack.

That doesn’t scare him—it excites him. But it’s not just about replacing jobs. It’s about freeing the web from the weight of inefficiency. The idea? Automate the grunt work so humans can build better paths forward… maybe even rediscover creativity. You know, that thing we used to care about in advertising?

Bannister hits a nerve here: publishers are squeezed on every side. Curators take their margin. DSPs flood their pipes. Buyers chase MFA crap that looks better on a spreadsheet. And now, publishers are expected to be influencer agencies, event companies, podcast studios, and occasionally, maybe, actual journalism operations.

Paul’s frustration is clear: the monetization path is too narrow. You either go big—massive scale, pools of remnant inventory—or you go hyper-niche and serve a loyal, targeted audience with expensive CPMs. But if you’re stuck in the middle? Good luck. That’s the red wedding of ad-supported content.

Casale agrees—and calls BS on a lot of what’s passed as “publishing” in the last five years. You can’t call yourself a publisher if your entire model is AI-slop plus Taboola clickbait plus paid traffic from Taiwan. That’s not survival. That’s arbitrage. And Casale’s proud that Index was one of the first SSPs to aggressively dump MFAs. Why? Because they weren’t built to support arbitrage—they were built to support publishers.

Of course, that raises the uncomfortable question: if the most profitable sites in your exchange aren’t real publishers, what does that say about the buyers?

CTV was supposed to be the golden child. Instead, it’s a hormonal teenager, flipping the dinner table because it doesn’t understand its own pod structure.

Casale doesn’t mince words. The problem? We took the protocol for display and tried to jam it into CTV without actually redesigning it for how TV works. Imagine trying to sell 15-second slots in a commercial pod using the same logic as banner rotations—and wondering why the metrics are on fire.

The other CTV headache? Burst traffic. When a million people watch the same show and hit the same ad break at once, you don’t get smooth auction behavior. You get chaos. Floods of bid requests. DoS-level surges. DSPs choke. Costs balloon. Infrastructure melts. And that’s before we even talk about how opaque the show metadata is. (Good luck knowing if your spot ran next to a prestige drama or a rerun of Duck Dynasty.)

Casale is working with the IAB on standardization, but let’s be real: half the platforms in CTV still won’t even pass you the name of the show because of “rights issues.” So yeah—connected TV is the future. But it’s also a terrifying game of "guess which bundle your ad was in."

This might be the only thing all three agree on. Transparency is real—but it’s also marketing.

Paul jokes that transparency basically means “we’re slightly less shady than the others,” and honestly, he’s not wrong. Casale goes further: Show me your log-level data. Index has offered a full API for buyers to audit every transaction. Guess what? No one used it. Until now.

Why? Because we’re in the efficiency era. Post-2020, with inflation rising and margins shrinking, CFOs are asking where the money’s actually going. Suddenly, buyers want to know: where did that dollar go? How many middlemen took a cut? Why does a $10 CPM only yield $1 for the publisher?

Casale believes that infrastructure fees should be Visa-level, not SaaS-level. We’re talking single digits—maybe even 1.5%. He even lays out the math: if 100 SSPs consolidate to 10, the volume will cover the margin drop. That’s how exchanges work in other industries. Why not here?

Of course, if that happens, half of today’s ad tech “platforms” disappear faster than an ad verification script on an MFA site.

Yes. And also no.

Yes, because the economics are broken. Because publishers are squeezed. Because buyers are tired of fake transparency. Because resellers and bid duplication are still a thing in 2025. Because Gen Z thinks "the web" is Instagram’s in-app browser. Because half of what gets measured doesn’t matter, and what matters can’t be measured.

No, because smart people are finally saying the quiet parts out loud. Because some platforms—like Index, like TripleLift—are making hard choices, sacrificing margin for principles. Because AI, if done right, could actually fix the sludge. Because there’s still a need for an open web. A place where independent content thrives. Where publishers don’t have to worship at the altar of TikTok. Where ad buyers aren’t just tenants in someone else’s walled garden.

But it’s going to take a rebuild. A reckoning. A realignment of what we value—and what we’re willing to fund.

🔍 What You're Missing in ADOTAT+
While everyone else is still busy name-dropping cookies like it’s 2013, we’re pulling back the curtain on the real issues keeping ad tech execs up at night. In ADOTAT+, we decode what happens when Bannister, Eifler, and Casale—three of the industry’s most pragmatic insiders—actually agree on something: that the open web is underperforming, programmatic infrastructure is bloated beyond reason, bid duplication smells like day-old MFA, and transparency is still more tagline than truth. Forget the headlines—this is where the bodies are buried. If you're serious about surviving the next era of ad tech, you can't afford to miss this breakdown. Spoiler: it ends with a prisoner's dilemma and a six-hop supply chain no one can explain. Welcome to the reckoning.