Crunch Time in DOJ v. Google: An Ad Tech Market Definition Cluster

A decision is expected soon in the landmark DOJ v. Google ad tech case in the Eastern District of Virginia-affectionately known as "The Rocket Docket"-where Judge Leonie Brinkema heard closing arguments just a few weeks ago. A threshold issue for DOJ is whether it can prove Google has monopoly power in at least one of the three ad tech markets it has defined. The first is an ad exchange market that includes Google's AdX product, which intermediates transactions between advertisers and publishers who respectively buy and sell advertising. Second, there is an ad server market and Google's Double Click for Publishers (DFP) offering, which allows publishers to place and manage bids. Third is an ad network market that encompasses Google Ads, which advertisers use to help them buy ad space.

Needless to say, Google has a different view on how the relevant market should be defined in this case. In addition to critiquing DOJ's attempt to limit the relevant market to "open web display advertising"-which, as ITIF has already noted, is itself highly problematic-Google rejects the idea that the ad tech space in which it competes should be trifurcated into distinct ad exchange, ad server, and ad network markets. Instead, Google argues that the relevant market should be viewed as a broader "two-sided transaction platform that connects publishers and advertisers to facilitate ad transactions" and which subsumes all three ad tech products-AdX, DFP, and Google Ads. In other words, for Google, "[a]dvertisers and publishers may use different tools to access the network but the ultimate product supplied is a display advertising match between a publisher and advertiser."

The controlling precedent is the Supreme Court's decision in Ohio v. American Express, where the Court laid out three core criteria for determining when a relevant market can take the form of a two-sided transaction platform. First, such a market is "two-sided" by virtue of "offer[ing] different products or services to two different groups who both depend on the platform to intermediate between them." Second, the platform must involve the facilitation of a "simultaneous transaction between participants," such that it "cannot make a sale to one side of the platform without simultaneously making a sale to the other." Third, there must be mutual indirect network effects across the platform whereby "the value of the platform to one group depends upon how many members of another group participate."

Of course, the debate over whether Amex applies is not academic. If Google's view prevails and the relevant ad tech market is one two-sided transaction platform, it is highly likely that DOJ will lose the case. Not only would it be hard to show Google had a monopoly, but even if it did, DOJ would have to prove that its conduct was anticompetitive when taking into account its overall effects upon users on both sides of the platform-a showing that, on the facts of Amex, was belied by output expanding dramatically on Amex's platform. And, similarly here, there seems to be little debate that quantity has increased. As Google argues, digital display advertising output increased from $7.6 billion in 2008 to $136.7 billion in 2022. Indeed, for all of its alleged anticompetitive behaviors, Google can certainly identify some procompetitive benefits that accrued to at least one side of the platform.

Who has it right? While both DOJ and Google agree that AdX is a two-sided transaction platform, the extent to which Google is correct that DFP and Google Ads should be incorporated with AdX into a broader two-sided transaction platform market depends upon whether the three criteria in Amex noted above are satisfied. With respect to network effects, Google seems right in arguing that "a tool for advertisers that is connected to more publishers (and, by extension, more users) becomes more valuable to advertisers" and conversely, "a tool for publishers that is connected to more advertisers becomes more valuable to publishers." However, this alone does not resolve the issue: These sorts of positive externalities would still exist if, as DOJ argues, AdX, DFP, and Google Ads were merely separate complementary goods and thus, for that reason, unable to be lumped together in a relevant market.

As to the issue of whether there is a simultaneous transaction, it appears clear that DFP and Google Ads are designed in a way that neither publishers nor advertisers respectively can, in the words of the Court in Amex, "make a sale to one side of the platform without simultaneously making a sale to the other." To this, DOJ responds that Google Ads and DFP, respectively, give advertisers and publishers a variety of specific services they can use without engaging in any simultaneous transaction. While that is no doubt true, DOJ seems to misconstrue the simultaneity requirement of Amex, which involves only the inability "to make a sale to one side of the platform without simultaneously making a sale to the other," with a broader requirement of not being able to use the platform (i.e., when a sale is not immediately being made)-without simultaneous engagement with the other side. Amex does not go that far.

That being said, DOJ's claim that DFP and Google Ads involve extra functionality goes squarely to the matter of whether they should be seen as respectively constituting the "sell-side" and "buy-side" of a two-sided transaction platform united by AdX. And, that AdX is highly integrated with DFP and Google Ads is not enough to show two-sidedness, as it is these sorts of integrations DOJ objects to as anticompetitive, and which thus risks a quasi-Cellophanefallacy. However, that DFP and Google Ads offer functionality that is respectively specific to publishers and advertisers, which does not involve simply making a transaction, should not be disqualifying for defining a two-sided market: Credit cards were within the two-sided transaction platform in Amex despite also giving cardholders the benefit of ancillary services such as payment and account management tools, as well as fraud protection.

At bottom, there are several reasons for thinking DOJ is on shaky ground in attempting to define three separate ad exchange, ad server, and ad network markets given the Supreme Court's decision in Amex. In fact, should Judge Brinkema not be comfortable with Google's approach, she may find that the truth is in the middle. That is, even if AdX, DFP, and Google Ads may formally be separate complementary products, the market realities might dictate that they are part of a broader cluster market akin to what was defined in Philadelphia National Bank in light of, as Professor Hovenkamp has noted, not just demand side complementarity, but supply side economies of scope and the fact that "entering into competition with the cluster is difficult." Those seem to be three propositions DOJ could agree with.