Programmatic’s Broken Promise Spurs a New Wave of Full-Stack Ad Tech Models

Programmatic advertising was supposed to be the open web’s great equalizer, a transparent marketplace where quality would win on its own merits. Instead, it became tangled in fraud, fees, and fuzzy measurement. The market’s response is already underway, as major players build integrated systems that borrow from the walled gardens' playbook: direct paths, closed-loop results, and a focus on what actually performs.

That's the perspective of Johnathan Barnes, Founder and CEO of Population Science, a firm focused on bringing new efficiency to programmatic media. A longtime ad tech operator who’s seen every promise and pivot, Barnes keeps his message simple: stop debating the semantics and focus on the strategic moves already underway.

"The promise of what programmatic was supposed to be has failed. It was meant to have similarities to how securities are traded: a Bloomberg terminal for media where buyers could sort out the good from the bad. For a lot of reasons, that just didn't work out," says Barnes.

• Pragmatism, not perfection: According to Barnes, the open web’s complexity and lack of clear measurement have created a performance gap. That gap gave walled gardens like Google and Meta their opening, and many advertisers, desperate for reliable results, turned to their closed-loop models. "We can debate how accurate their measurement is, but it's a good enough north star. It's better than anything we have in the open web," he says.

• Get on with it already: Barnes views the endless buy side versus sell side squabble as a tired distraction. The market has already moved on. He points to companies like AppLovin, which built a massive business as a closed ad network, as a clear sign that the market has already voted with its feet. "I'm just tired of this dichotomy discussion. We've been talking about this for five-plus years now, still hemming and hawing about it. It's time for the industry to make its strategic moves and move forward."

The path forward, according to Barnes, is being demonstrated by companies applying the lessons of the walled gardens to the open web. He highlights the strategy of The Trade Desk, saying it uses its massive repository of buyer data to "identify the quality inventory and activate against that without all these intermediaries taking their taxes in between."

• Ad tech squeeze: The result is a pincer movement on the old model, as major players build full-stack systems from opposite ends of the ecosystem to recreate the efficiency of the walled gardens. "Magnite has ClearLine, and they want to get more buyers in there directly. They're thinking, 'We've got all the supply, let's go get the demand.' The Trade Desk is doing the opposite. They're thinking, 'We've got all these buyers, let's use OpenPath to go get all the premium supply,'" Barnes explains.

So what makes this new structure possible now? In a word: AI. Barnes sees agentic workflows as a means to address a long-standing operational bottleneck: the inability to scale direct deals. That process was historically stymied by manual work until frustrated buyers retreated to the open market. Now, standards like the Ad Context Protocol are automating the work that once made scaling impossible.

But what about AI taking everyone's jobs? In the near-term, Barnes isn't worried. "For the rest of this decade at least, media buyers will be the pilots and AI will be the copilot," he says. "It's going to let the humans behind the controls get to the things they haven't been able to, like addressing fraud and poor quality. It's the work they've been wanting to do for a long time but just haven't had the bandwidth for." For Barnes, ad fraud is the industry’s easiest win and biggest embarrassment, and AI finally gives everyone the bandwidth to do something about it.