Ad tech insiders met the allegations that Google secretly fed publisher ad server data into its buying systems to gain market share and keep prices low with resignation. These claims were revealed in a mistakenly unredacted court filing obtained by the Wall Street Journal.
“This shouldn’t be surprising – walled gardens have access and full control over vast amounts of ecosystem data,” said Ana Milicevic, co-founder and principal of Sparrow Advisers. “The surprising part would be if they didn’t use it in this way to their commercial advantage.”
But publishers still didn’t know the specifics around Google’s advantaged position – for instance, how the tech giant netted $230 million in 2013 thanks to Project Bernanke.
“We have not heard of this practice, but it is the type of news that we typically point to when we talk about lack of auction transparency creating erosion of trust,” said Nicole Lesko, SVP of data, ad platforms and monetization at Meredith.
Who was harmed?
The unredacted filing doesn’t fully explain how Google actually used the ad server data however. And in ad tech, the devil is in the details.
“There are a lot of missing details and a big conclusion,” cautioned Scott Messer, SVP of media at Leaf Group.
If Google used ad server log data to advantage itself, the question is who ended up losing the stacked game: publishers, SSPs, or someone else.
For instance, publishers would be harmed the most if Google used the data to tell ad buyers like DSPs to bid less, causing publishers to receive lower prices for their ads.
“When the dust settles, it will turn out that digital programmatic is not actually THAT bad a business for local publishers aside from an opaque duopoly that allowed value to be siphoned off at every link in the chain,” tweeted Damon Kiesow, Knight Chair in Journalism for the University of Missouri and a former head of product for McClatchy.
On the flip side, independent SSPs would have been most harmed if Google used ad server data to change the floors on some inventory, helping its own advertiser demand win. Advertisers who used Google’s DSP would benefit, but marketers who used third-party DSPs would lose access to inventory.
“The control of the publisher ad server has been critical in Google’s domination of digital advertising,” said Romain Job, Chief Strategy Officer of Smart, an ad server and SSP competitor to Google. “While it may never have been a cash cow, ad server control allows Google to block other players from competing for share.”
So depending on if Google used the ad server data to do bid shading (which would hurt publishers) or change price floors (which would hurt competitive SSPs and advertisers using non-Google DSPs), the practice would hurt different parts of the ad ecosystem.
The lack of transparency makes it harder for publishers and advertisers to figure out who benefitted from Google trying to route more ad spend through its pipes.
“The mechanics…would determine whether or not it ultimately helps or hurts publishers,” Lesko said.
Does Project Bernanke still exist?
Project Bernanke happened before header bidding changed the auction environment, and it’s unclear how its data sharing program evolved after header bidding took off.
But the same principles that defined Project Bernanke could be applied through Google’s open bidding feature. Publishers see evidence they already are.
“We see increased CPM and win rates from the same exchange through [open bidding] vs. [header bidding],” even when accounting for timeouts, Lesko said.
Exchanges use open bidding to access more data, which improves their auction logic so they look better to DSPs.
“Google could be trying to stunt header bidding in favor of open bidding where it can exploit bid-level data to its benefit,” Job said of Smart.
So if Google is still quietly sharing ad server data with other Google ad products, while withholding that same data from competitors, Project Bernanke lives on.