Prebid wants Google bidding in the auction, not running it.
Who Gets the Last Word?
Google escaped a court-ordered breakup of its advertising technology empire. Now Prebid has proposed a way to weaken Google’s dominance without forcing it to sell anything.
At the Prebid Summit in New York on Oct. 13, the industry group floated a proposal that could fundamentally change how online advertising is bought and sold: reverse the order in which advertising auctions happen.
How Google Gets the Last Word Today
Today, publishers can use Prebid, an open-source auction technology, to invite multiple advertising exchanges to compete for an available ad impression. Prebid collects their offers and forwards the highest eligible bid to the publisher’s ad server, usually Google Ad Manager (GAM).
GAM then decides which ad runs, comparing the Prebid offer against other demand, including Google’s own AdX exchange and direct advertising commitments.
Here’s the problem: Google gets to see the winning Prebid price before its own exchange competes. Imagine an auction where everyone submits sealed bids, except one bidder gets to see the highest competing offer before making its own. Fair this is not.
Prebid proposes reversing the arrangement. GAM and other ad servers would submit their eligible advertising offers into Prebid alongside competing exchanges. Prebid would conduct the final comparison and select the winner, while respecting guaranteed advertising commitments.
That would address a central competitive problem created by bundling AdX with GAM. Google would lose its privileged position as the operator of the final auction and become one participant among equals, competing for impressions rather than controlling the process that awards them.
Breaking Up Google Without Breaking Up Google
In September, federal Judge Leonie Brinkema declined to force Google to divest AdX or GAM, despite finding that Google had unlawfully monopolized the publisher ad server and exchange markets.
Instead, the court ordered Google to integrate its technology independently with Prebid, give competing ad servers access to AdX demand, and to share more auction data.
Prebid’s proposal aims to turn those requirements into a new auction architecture. Moving the final auction from GAM to Prebid could level the playing field while allowing publishers to access Google’s enormous advertiser demand through competing ad servers.
Could that be more effective than Judge Brinkema’s behavioral remedies alone? You bet.
What Publishers Get
The biggest potential benefit is freedom of choice. Publishers could select their preferred ad server without sacrificing access to Google’s advertising demand.
They could also gain greater visibility into winning and losing bids, enabling better pricing decisions and faster optimization. There’s even a potential speed advantage. Combining sequential auctions could eliminate a step and improve page-loading performance.
But I would be cautious about promising higher revenues or faster websites. The architecture introduces new integration requirements. Publishers must preserve guaranteed campaign delivery, manage pricing rules, and ensure competing systems exchange information correctly. Server-side auctions have their own costs and delays.
Publishers will ultimately judge the proposal by its impact on revenue, operating costs, and performance.
Will Google Cooperate?
The proposal offers one possible path for Google to comply with Judge Brinkema’s remedies. But that doesn’t necessarily mean Google must embrace Prebid’s preferred architecture.
The proposal favors Prebid Server, which conducts auctions on a separate server rather than inside the consumer’s web browser. Google could instead potentially satisfy some of its integration obligations through Prebid.js, the browser-based technology already widely used by publishers.
If Google supports only the browser-based approach, Prebid’s preferred server-side framework could struggle to gain traction. Other ad servers would also need to participate for the proposed system to become a genuine alternative.
I suspect Google will comply with its legal obligations while protecting as much of its existing advantage as possible. Nobody should expect enthusiastic cooperation from a company being asked to surrender control of a lucrative marketplace.
At the very least, the proposal highlights how much control Google retains over the auction process, despite the antitrust verdict against it.
When Could This Matter?
The proposal remains in public consultation, with formalization targeted for Q1 2027. Implementation would follow in a subsequent Prebid release.
Limited publisher pilots seem plausible during 2027, assuming technical agreement, meaningful Google participation and successful testing. Broad adoption could extend into 2028 or beyond.
Several hurdles remain. Google would have to provide workable integrations. Publishers would have to trust the replacement auction. Guaranteed campaigns would have to continue running properly. And independent testing would have to demonstrate financial benefits.
Large, technically sophisticated publishers will probably experiment first. Smaller publishers will wait for proven implementations and compelling results.
Google’s dominance won’t disappear because somebody reverses an auction. But publishers gaining access to Google’s demand without depending on Google’s ad server would represent a significant shift in the competitive dynamics of programmatic advertising.

Leave a Reply