Erroneously unredacted court docs reveal Google ran a secret program called “Project Bernanke” that used past bid data to boost its win rate in ad auctions
Program used past bid data to boost tech giant's win rate in advertising auctions, according to court filing
Context & Ripple Effects
The disclosure emerged from the Texas antitrust case after an unredacted draft of the states’ lawsuit detailed an alleged Google-Facebook advertising deal. The newly visible material adds a separate allegation about how Google’s own ad-buying system operated inside auctions.
The story became a focal point beyond the case: Sen. Warren’s request for a CFTC probe centered on the alleged use of past bids, while later state claims broadened scrutiny to ad-pricing practices affecting publishers and buyers.
First-order effects
- Google must defend the alleged Project Bernanke auction mechanism in the Texas litigation, while state plaintiffs gain a more specific account of how its ad-buying system may have used historical bid information.
- Advertisers and publishers implicated by the auction system face sharper questions about whether auction outcomes and prices reflected a neutral process.
Second-order effects
- The disclosure gives regulators and plaintiffs a concrete theory to examine alongside the alleged Google-Facebook deal, rather than treating the case solely as a partnership dispute.
- Google’s ad-tech customers face greater pressure to seek clarity on auction rules and pricing after the later state allegations of pricing distortions tied publisher sales and buyer costs to Google’s conduct.
Third-order effects
- If similar internal auction practices become central to enforcement, ad-tech competition cases will increasingly turn on the transparency of market-design rules and the operator’s access to bidding data.
- The episode points toward greater scrutiny of vertically integrated ad platforms whose auction operator also participates through its own buying tools.
The trend: Digital-advertising antitrust scrutiny is shifting from platform partnerships toward whether auction operators can use privileged market data to advantage their own services.