Overheard at Prebid Summit: ‘Way too often, there’s a delta between what people do, and what they say they will do’

Below is a compendium of they key on-stage talking points in order to help you take the pulse of the media industry’s sell-side.

• ”I don’t think the auction is inherently a sell-side function… We believe that his is a race between the open internet and the walled gardens, this is about creating efficiency for all… Way too often in our space there is a delta between what people do, and what they say they will do.”

—The Trade Desk CEO Jeff Green — who opened his session by quipping that he thought he wouldn’t be allowed back into the conference after many interpreted his company’s recently announced OpenAds tool as rival sell-side product — maintained his outfit is there to be a team player, and that his latest launch wasn’t a rebuttal to Prebid’s Transaction ID.  

• “Tomorrow [i.e., that’s today Oct. 15] a group of companies are coming together to announce the beginning of a new era of agentic advertising. Today, the agentic advertising era is the equivalent of the day before we took header bidding live [in 2008]… We need a new definition of what it means to be a supply-side platform.”

—One of “the Godfathers of ad tech” and Scope3 CEO, Brian O’Kelley continues his tradition of fraying the nerves of his PR and marketing teams by teasing the launch of an industry-wide effort to usher-in a sustainable ad-funded model to the publishing sector in the AI-era. More will be revealed at 11 am ET.    

• “We [Prebid] will be launching two task forces: one on content monetization, and one on agentic advertising Microsoft and myself will be involved.”

—Prebid director, and Microsoft principal product manager, Dr. Paul Farrow, informed attendees of his efforts to help publishers monetize, in a presentation where he also teased that his company’s recently leaked plans for AI monetization have firm revenue models in place. 

• “The platonic ideal of two separate sides has gone away… the platonic ideal sounded great, but it’s not really working. The [next] natural step is for everyone to collide.”

—Ad tech influencer, and author of Yield: How Google bought, built and bullied its way to advertising dominance, Ari Paparo offers his assessment on the state of the industry, which currently finds itself in the “constant consolidation” era. 

• “We saw this with MFA… and AI-slop at the end of the day is MFA. We as an industry need to come together and have a better definition (as tough as it is) with something that’s almost as subjective as this.”  

—MediaVine co-founder and CEO Eric Hochberger, shared thoughts on the most existential issues in the history of publishing, and drew parallels with one of advertisers’ more recent preoccupations — anyone remember the made-for-advertising furor of the early 2020s — and urged attendees to learn the lessons from the past.

• “We’ve been experimenting with CloudFlare’s bot-management suite… even with the LLMs, you have to understand which ones you want to block and in which scenarios… you have to figure out which ones you might be able to get money from in the future.”  

—Renown, sell-side industry commentator and chief strategy officer at Raptive, Paul Bannister, shared insights on how the potential for “overblocking” crawlers — an operational risk given the de rigeur proposed ‘pay-per-crawl’ model — a mode of engagement that can inadvertently lead publishers to prevent sources of revenue, such as demand-side platforms, from placing bids. 

• “We’ve injected mentions of advertisers into the content [of our publisher partners], so, essentially, whenever an LLM would reply [to a user’s request]… it would at least mention content [theoretically] sponsored by Coca-Cola.” 

—Prebid Summit panelist, and Sigma Software CEO, Katherine Tuluzova shared insights on how the software outfit was experimenting with alternative models for publishers to generate revenue. She later went on to mention how her company was able to get “60%-to-70% mention rates” for brands in such instances, though did not provide exact figures. “It was interesting, because we learned a lot about how the LLMs’ cache actually works.”