Adtech isn’t dead.

But if you listen closely, you can hear it gargling.

And Andrew Eifler, Chief Product Officer at TripleLift, doesn’t just know where the bodies are buried—he helped design the cemetery. In our hour-long conversation, he delivered a masterclass in what’s holding programmatic back: not lack of innovation, but an industry designed to reward chaos, duplication, and scale-for-scale’s-sake idiocy.

The current system, he tells me, looks like this: “The SSPs are standing in a semicircle and every SSP has a fire hose… the DSPs [are] standing in the middle, it's like a single person and they have their mouth open. And the game is how much water can you get in the DSPs mouth.”

Yeah, I said ewwww also.

Oh—and by the way, “the water costs money.”

Because of course it does.

That’s not just a vivid, slightly nasty metaphor.

t’s a brutal indictment of the business model we all pretended was innovation. We built a system that encourages bid duplication, redundant infrastructure, and artificial volume—not to optimize campaigns, but to bloat revenue metrics for platforms playing hot potato with demand. And the result? Even when everyone does their job “right,” no one’s happy. Not the publisher, not the buyer, not even the DSP engineer trying to trace 50 million bid requests back to a single damn user.

Everyone in ad tech loves to say “AI is going to change everything,” but what they mean is “We need a new buzzword to put in the pitch deck.”

Eifler? He breaks it down into actual phases of functionality.

First wave? Chat interfaces. Think: talking to machines, novelty style.

Second wave? “Code assist… pretty much every company in ad tech now uses some amount.” This is GitHub Copilot and its cousins making mediocre devs look competent.

But the third wave? That’s where things get real. “It’s the agentic wave,” he says. “Where ad tech really creates AI agents to help with specific tasks.”

We’re not talking about ChatGPT hallucinating strategy slides. We’re talking about bots that do actual work—pulling reports, managing campaigns, normalizing bid logs, and eliminating half the drudgery that keeps ad ops teams up at night with a bottle of ibuprofen and a half-built dashboard.

“There are so many tasks in programmatic advertising that are technically complex, manually labor intense, and require accessing data that’s difficult to get your hands on,” he points out. “It’s gotta be like 50% of all the work that’s done in our industry.”

And no, he’s not worried about AI taking jobs. The threat isn’t elimination—it’s irrelevance. If we don’t automate the sludge, we’re just going to drown in it. “What will happen is the pie will grow 100X. The jobs will change. The jobs will be higher leverage.”

Ask most DSPs about privacy and you’ll get three things: vague nods, a Chrome timeline update, and a white paper no one read.

Ask Eifler? You’ll get a clean, no-BS answer: “Publishers are making privacy-first advertising work.”

That’s right. Not the platforms. Not the intermediaries. Publishers.

“For years, publishers have been using first-party DMPs… to observe what their users in a privacy-safe way do on their site,” he says. And rather than hoarding the data like dragons, they’re building curation packages advertisers can actually buy.

TripleLift didn’t just nod along. They acquired OnePlusX out of Zurich—“one of the leading DMPs”—and basically plugged it into the entire ecosystem. “We’re basically giving the DMP away to publishers for free and creating curation packages across all publishers that can be bought programmatically.”

In other words, they’re doing the thing the rest of the industry just tweets about.

Retail media has become ad tech’s favorite punchline: the magical new channel that’s supposedly going to save everyone’s margins and create a new category of unicorns. Eifler isn’t drinking that Kool-Aid.

“I think the biggest misconception about retail media is that it’s going to produce lots of benefit for every ad tech company,” he says. “And it’s really not.”

Unlike brand campaigns, where “you don’t have to outrun the bear—you just have to outrun the other campers,” retail media is the bear. If your campaign doesn’t produce positive ROAS, there’s literally no budget for you. Retail doesn’t tolerate fuzzy attribution. It’s outcome or extinction.

And the creative canvas? Comically bad. “Forty percent of all web advertising is 300x250 banners,” he points out. “There’s just no way to rearrange the pixels inside a 300x250 banner to create a compelling retail media experience.”

So what’s the answer?

If the term “native ad” still makes you think of clickbait under a news article, you’ve missed the point entirely.

“All native advertising is, is a vertical integration between supply and creative,” Eifler explains. Most programmatic buys today are modular: supply from one vendor, creative from another, data from a third. Native skips the Frankenstein act and builds ads that actually fit.

“We’re not constrained to fixed banner dimensions,” he says. “We can actually create an ad that looks good on every page, no matter how big the page is. There’s no lowest common denominator box.”

That’s not just better UX. In retail media, where you need dynamic product info, real layout flexibility, and data-driven creative—it’s required.

Native isn’t a format. It’s a survival mechanism.

Eifler’s biggest thesis? The open web isn’t underperforming because it lacks potential. It’s underperforming because the ecosystem is clogged with “sand in the gears.”

“The CPMs we produce for publishers are not good enough for those publishers to run thriving businesses,” he says. “And the returns we produce for advertisers aren’t good enough to put the open web above search or social in their priority stack.”

But he’s adamant: this is fixable.

“If we go systematically through all of the issues that are dragging down the open web… we can actually grow.” And he doesn’t mean 5%. He means 10x, maybe even 100x. “We’re focused on the creative aspect. Why do all our ads have to be trapped in 300x250 banners? It doesn’t make sense.”

Andrew Eifler isn’t interested in cute panel slogans or rehashed talking points. He’s trying to rebuild the machine with cleaner code, better logic, and fewer vendors with ten resellers stacked behind them.

The question isn’t whether ad tech can evolve. It’s whether we’ll stop incentivizing garbage long enough to do the hard work of fixing it.

Because yes—right now, programmatic is a mess. But it’s our mess. And if we don’t fix it, someone else—maybe a walled garden, maybe an AI agent that actually does what it promises—will do it for us.

📌 Next Up (Free): “Bid Duplication, Reseller Vampires, and the Secret Cost of DSP Waterboarding”

💸 Unlock ADOTAT+ to read:

Let’s start with a truth bomb Andrew Eifler drops mid-interview like it’s just another Tuesday: “Every SSP that resells… increases their bid volume by 100 percent.”

Think about that. One reseller?

You double your traffic.

Ten?

Congratulations, you’re flooding the DSP like a kid with a garden hose and zero impulse control.

Now, combine that with what Eifler calls the “availability bias built into the ecosystem”—the idea that more volume equals more value—and you’ve got a system that incentivizes bad behavior at industrial scale.

This isn’t a bug in programmatic. It’s the business model.

Remember when bid duplication meant “accidentally” sending the same impression through five pipes? Cute. Now it’s worse—and smarter. The new bid duplication is reselling: SSPs buying and reselling inventory from each other like digital laundromats. The output? More bid volume, more middlemen, and zero added value.

Eifler puts it plainly: “The biggest vector for duplication right now is actually reselling. So it’s SSPs selling to other SSPs or other entities that then resell.”

Why do they do it? Because it juices the numbers. Each hop adds more bid requests. More requests means more perceived scale. And the buyer? Still chasing real people through a fog of mirrored impressions and invisible margins.

Let’s not pretend this is harmless. Eifler’s clear: “Every intermediary you add actually logically extracts margin from the value chain.” And what’s worse? “Right now that’s still an acceptable way to run a business.”

Why hasn’t this been fixed? Because it’s a classic game theory trap with a twist: 100 players, no rules, and everyone's scared to blink.

“There are a hundred SSPs,” Eifler says, “and somebody’s going to do the worst possible thing.” TripleLift took a stand and banned resellers. The result? “It largely hurt us as a business. But it’s the right thing to do, so we’re doing it.”

That line right there? Burn it into the pitch decks of every company claiming they care about sustainability, transparency, or any other buzzword they slapped onto their LinkedIn bio.

This is what principle looks like: losing revenue because you won’t contribute to fraud-by-design.

Eifler doesn’t let DSPs off the hook either. If they actually enforced rules about bid quality or blocked duplicative supply, the whole reseller circus would collapse by Thursday.

“The people that really have the incentive to solve it are the DSPs and the advertisers,” he explains. But right now? Most are too focused on shaving half a cent off CPMs to care that they’re being served ten copies of the same ad call from ten slightly renamed URLs.

What happens if DSPs push back? Eifler has a bleak but honest take: “The big SSPs would probably say the same thing I’m saying… but they’d also say that someone else is going to do it, so I have to do it.”

Sound familiar? It’s the ad tech version of “I’m only evil because I don’t want to lose market share.”

Let’s talk about something almost no one in this ecosystem is calculating: carbon emissions.

Every extra bid request travels through infrastructure. Every hop, every duplication, every reseller adds compute cycles, server loads, and data center strain. Multiply that by billions of bid requests per day and you’ve got an invisible environmental cost that’s being completely ignored in QBRs.

Eifler nails it: “Reselling in programmatic is probably the biggest factor right now that’s climate-changing emissions and data center costs.” And he’s not wrong. We’re burning server racks to resell the same impression ten times with different wrappers.

It’s not just inefficient—it’s immoral.🧠 Final Thought: This Isn’t “Innovation.” It’s Just Organized Inefficiency.

If you’re an advertiser spending millions in programmatic and think you’re “buying premium inventory,” ask your partners how many hops your impression took before it reached the DSP. Ask how many identical bid requests they sent. Ask how many margins were stacked between you and the publisher.

Chances are, no one can give you a clean answer. And that’s the problem.

Eifler isn’t ranting. He’s warning us. “No advertiser that buys in programmatic is like, I’m getting into programmatic to buy a bunch of reselling,” he says. But they are. They just don’t know it.

And until DSPs enforce transparency, SSPs stop chasing fake scale, and we stop rewarding this mess with revenue, we’ll keep drowning.

One bid request at a time.

So you thought MFA was just some backroom click-farm hustle? Cute.

What we uncovered is far messier: a house of cards built by everyone — not just the scammers. MFA was never a bug in the system. It was the system. And while brands were optimizing for “efficiency,” what they were really doing was laundering performance through a fraudulent supply chain designed to reward failure at scale.

🚨 Inside the ADOTAT+ report:

📊 Real data. Real fraud. Real consequences.
Like CPAs dropping 83% the second MFA gets cut. Or 30% of bidstream traffic vanishing overnight. Yeah, it was never about optimization—it was about who got to cash the check.

And it’s not over. Retail media’s got the same infection. CTV’s already coughing. Verification? Always 18 months late and 18 cents too short.

If you're still buying media like it's 2022, this isn't just a wake-up call. It's the fire alarm.

👉 Join ADOTAT+ and read what the dashboards don’t tell you.