The “Performance Web” Thesis, Revisited

Performance marketers have long hailed search and social as the “performance channels” of digital advertising. They capture intent, drive conversions, and deliver outcomes. Meanwhile, the open web has had a less coherent relationship to performance. 

But increasingly, a new narrative about the open web is taking root: the performance web, powered by AI and differentiated by transparency and control, may be coming for performance ad dollars. 

Taboola says nearly 75% of performance marketers have seen diminishing returns on social and that more than half were expanding into additional digital channels beyond it. As a market signal, that tells you what performance advertisers are anxious about and, just as importantly, what adtech companies think they want to hear.

It’s a narrative that has the potential to lift spirits across the open web: if performance is down on social, perhaps the pendulum could swing toward the programmatic ecosystem.

Maybe. I’m not here to litigate attribution models or diagnose Meta’s performance. What interests me is how this shift functions as a go-to-market signal and what it tells us about the GTM stories adtech companies are adopting.

A familiar playbook 

The performance TV thesis, the idea that TV can deliver measurable, optimized outcomes, helped push performance marketers into CTV. That same thesis (touted by companies like tvScientific, MNTN, and Vibe), is now being retooled for the broader open web and applied to display and OLV.

For instance, Taboola’s Realize is framed as an “industry-first platform” for “performance outcomes at scale” that explicitly extends beyond its old native placements into display and video. It pitches proprietary publisher signals, AI-driven audience identification, and creative optimization for ROI across campaigns.

Viant pushes this even further. Its “Outcomes” product is framed as the “first fully autonomous open internet ad product,” powered by Lattice Brain, where advertisers simply define the result and let the machine do the rest. Or just look at Teads’ homepage: “The Global Platform for Elevated Outcomes,” designed to work for the open internet.

The core pitch hasn’t changed much: the walled gardens are crowded and expensive, the open web can now be optimized more intelligently, and AI can finally make these environments behave like true performance media.

It’s a strong narrative, particularly because it aligns with what we already know about the Outcomes Era — namely, that platforms win by proving they drive performance. Even when the language changes, the structure of the argument is remarkably consistent.

From reach to yield

The challenge is that growing open web traffic isn’t on the table. The structural shift away from the open web toward apps, closed platforms, and AI-native interfaces is already established. You’re not going to reverse that trend through better storytelling or smarter sales.

What is in play is how much value you can generate from the traffic that remains. This is where the performance web thesis starts to make sense. Instead of making the funnel bigger, the goal is to increase yield. That means maximizing the value of every open web audience interaction.

That is also why AI features so prominently in this new wave of positioning. In one form or another, these companies are all arguing for the same set of capabilities: creative that adapts faster, targeting built on better signals, autonomous optimization, smarter allocation, and clearer measurement. 

Whether it’s Taboola emphasizing AI-powered audience discovery, Viant promising autonomous decisioning, or Teads tying omnichannel intelligence to outcomes, the task is the same: grow the open internet by using a generational technology to make it perform.

The open web’s performance moment 

That’s the essence of the “performance web” thesis. A year ago, it was coming. Now it has become a recognizable market claim.

We’re seeing more companies claiming they can turn open web media into a precision engine for outcomes. And we’re hearing a lot more about performance in CTV, display, native, and OLV. The evidence is sitting directly in product launches, investor language, and homepages.

The performance web thesis is here. The question is who can actually back it up.