At Sell Side Summit Austin, publishers explored how AI, first-party data, and diversified ad tech stacks are helping them reclaim control.
At Sell Side Summit Austin, publishers faced the future head-on, and, yes, I even broke it down in song live on the event stage. Here’s my ad tech remix of Beyoncé’s “16 Carriages”:
16 crawlers driving traffic away
Pub rev is down, and GAM is going away
To the Q4 sunset on a cookieless night
On a programmatic road, all the tears pubs cry
It’s a reminder that the challenges publishers face are real, from the chaos of cookieless advertising to workflows being disrupted by AI. But as we sang, the recurring refrain at the summit rang out: evolution is coming, revolution is lurking, and control over the audience and the programmatic supply chain is the prize.
From Arena Group’s loyalty-driven reinvention to The Yield Doctor’s crash course in AI evolution vs. revolution, one thing is clear: Publishers need to step out of the passenger seat and reclaim the wheel.
AI’s Evolution vs. Revolution
James Deaker, aka The Yield Doctor, took the stage in Austin to explore the biggest question of all: how much the world of ad operations is about to change under AI.
To put the scope of AI’s potential impact in perspective, Deaker drew a clear line between evolution and revolution. “Evolution is like getting a new car model,” he said. “Perhaps it has better electronics, or it’s more efficient, but you still know how to drive. Revolution is a self-driving car.”
According to Deaker, AI will have an evolutionary impact on how publishers handle ad serving, audience segmentation, forecasting and reporting.
“Ad servers have had more than a decade of machine learning investment. We’re talking clustering, pacing, and delivery mechanics,” he said. “AI will make it incrementally better, but not fundamentally different.”
Deaker also highlighted how AI is creating incremental efficiencies without eliminating human roles. He predicted that ad operations personnel will become interpreters of AI outputs and that most of what they do day-to-day will remain familiar. However, he advised ad ops teams not to get caught up chasing minor improvements with AI, but to focus their energy on changes that could really shake things up.
As far as where AI becomes truly revolutionary, Deaker pointed to emerging communication protocols for AI agents.
The Model Context Protocol (MCP) and the Advertising Context Protocol (AdCP) are laying the groundwork for machines to communicate efficiently with each other. Deaker referred to the development of AdCP as “the new USB-C for advertising workflows” – similar to how USB-C established a standard connector for devices, making previously widespread legacy and proprietary connectors obsolete.
Now that universal protocols for agentic AI applications in advertising exist, Deaker said, the next wave of AI-based offerings could fundamentally shift how publishers and advertisers execute campaigns, generate creative and make decisions.
Making First-Party Data Work Harder
While there was plenty of forward-looking speculation about the future of AI at Sell Side Summit, publishers also shared how the tech is already transforming their businesses.
Stephanie Mazzamaro, Head of Programmatic, Addressability & Ops at The Arena Group, walked the audience through how her company is leaning into newsletters and first-party data to deal with today’s audience growth and monetization challenges–and how AI plays a role in both areas.
At the heart of The Arena Group’s strategy is engagement, and lots of it.
The company uses its AI-driven intelligence platform, Encore, to break down audience engagement into different buckets, from new users to loyal subscribers and super fans. By understanding behavior at this granular level, Arena Group can craft experiences that feel personalized and, in turn, create monetization opportunities.
For example, new users who are making their first site visits in a year account for about 80% of Arena Group’s audience. For these users, Encore tailors the site experience to show them more content recommendations. The goal is to encourage them to click through to more content, increasing their dwell time and maximizing revenue per session.
Then, once a user visits and Arena Group site four times in a single month, the AI prioritizes funneling them toward subscriptions and newsletters.
“Our single sign-on system drives roughly 50,000 signups a day,” Mazzamaro said, noting that acquisition is just the start. The real challenge, she explained, is in continuing to deliver value for the registered audience.
In that sense, Mazzamaro said, newsletters are less about collecting emails and user data and more about building meaningful experiences – and AI is taking on a bigger role in crafting those experiences. “It’s about creating reasons for users to return to the platform, not just capturing their information,” she said.
Reclaiming Control from GAM
But integrating new AI solutions into their stack isn’t the only way publishers are changing the current ad ops status quo.
Walt Houseknecht, Executive Director of Ad Tech Operations at Politico and Carlos Bracho, VP of Global Ad Tech at Politico’s parent company Axel Springer, illustrated how they’re moving away from a heavy reliance on Google Ad Manager and reclaiming control over their ad operations. Central to that strategy is Axel Springer’s AdLib product, an adtech management and optimization solution the company rolled out last year.
AdLib powers Politico’s ad tech layer, optimizing placements, generating first-party IDs, and enabling real-time bidding across platforms through a flexible mediation system.
Houseknecht was candid about Politico’s starting point: without dedicated ad tech resources and limited flexibility, the company’s previous GAM-dependent setup left them exposed to inefficiencies.
Now Politico’s stack is much more diversified, shifting from a Google-centric approach to a multiplatform stack. The new stack processes requests with Microsoft as the primary ad server. In addition, it has a dynamic mediation layer that can simultaneously query a DMP, Prebid, and other ad servers within 700 milliseconds. This allows real-time competitive bidding, which helps drive up CPMs, said HouseKnecht.
Politico has seen a 26% increase in programmatic eCPM year-over-year and a 41% increase in total available inventory eCPM, which Houseknecht attributed to the changes in its tech stack.
Plus, by reducing its reliance on GAM, Politico also significantly reduced its dependency on Google’s SSP, AdX, which previously accounted for more than 60% of its demand.
Bracho framed AdLib’s approach as providing options, not replacing other platforms. “We just want publishers to have multiple levers and fewer dependencies,” he said.
For Houseknecht, the flexibility of not solely using GAM meant Politico could experiment more freely, introduce new ID signals, apply its internal audience and content taxonomy, and develop more targeted advertising solutions. All of these improvements translated to more efficiency in the company’s ad business, he said.
Politico’s experience demonstrates that publishers don’t have to accept the status quo. By integrating complementary solutions and reducing reliance on a single system, they can increase control, improve monetization, and build a more resilient revenue strategy.