A Deep Dive into the Antitrust Remedies that Spared Google's Ad Tech Business

The memorandum explaining the court’s remedies ruling for Google’s illegal monopoly over advertising technologies (ad tech) is now unsealed. About two weeks ago, Judge Leonie Brinkema posted a short summary outlining the gist of it: Don’t expect any divestiture orders. Remedies will be behavioral only, with some modifications ordered by the court.

To refresh our memories, the remedies trial that ended a year ago was mainly focused on deliberating structural remedies. The plaintiffs—the Department of Justice and a coalition of 17 states— argued that without structural separations, there would be a door open for Google to unlawfully leverage its control of the entire ad tech stack.

The plaintiffs sought a divestiture of Google’s ad exchange, AdX, and its publisher server, DFP. As the liability ruling found, the platforms were unlawfully tied, and Google also created dozens of policies to disadvantage competitors and manipulate ad auctions to favor its own business. Such policies, dubbed “First Look,” “Last Look,” and “Unified Pricing Rules,” among others, fortified Google’s monopoly, overcharging publishers for connecting them to advertisers and gaining full control of how they priced their ad spaces.

Why Google keeps its ad tech stack

As egregious as it sounds, however, the Judge ended up giving deference to Google’s stance against a break-up, as well as to the remedies ruling in the Google Search case (which not only rejected to break up Chrome, but struck down key behavioral remedies). She deemed a break up would be too complex and represent too much work for Google, and that behavioral remedies such as data sharing and interoperability remedies would fulfill the same function. In addition to that—“as Google correctly argues,” the ruling reads—divestitures aren’t applied to unlawful tying cases.

The Judge also downplayed the plaintiffs’ evidence that Google had repeatedly come up with new tactics to advantage itself whenever it phased out a certain policy that clients found coercive and unfair, or that Google had also violated remedies ordered by competition authorities in other nations that investigated the same unlawful practices this case did. Under her analysis, structural remedies are not imposed to provide “certainty” that a monopolist would not have incentives to break the law again.

Her focus was in not disrupting the market, albeit the market has remained undisrupted precisely because it’s captured by a monopolist. Somewhat surprisingly, the Judge also criticized that plaintiffs requested divestitures without having secured a buyer, even in the face of testimony of interested parties, who also said—as it is typical when deciding over mergers or acquisitions—that they couldn't commit then to buying AdX without doing due diligence first. To the Judge, this all made it seem like an unviable divestiture.

Somewhat hidden and underreported is that the Judge also considered AI as a potential disruptor that could “threaten the stability and growth” of open-web display ads. This consideration was also a factor in the Judge rejecting the proposed divestitures, even after acknowledging that AI is more nascent in the ad tech market than in online search. This reflects the same line of thinking as Judge Amit Mehta, whose analysis of AI challenging Google’s search monopoly has aged so far to show exactly the opposite is happening.

Behavioral remedies: A good first step

What we’re left with is the plaintiffs’ behavioral baseline remedies to unbundle AdX from DFP, achieved through mandated interoperability and non-discrimination provisions. Moving forward, publishers using DFP will not be obligated to use AdX to access Google’s advertiser demand on AdWords, and vice versa, advertisers wouldn’t have to connect to AdX to access DFP either.

Google’s DFP has to interoperate with other exchanges the same way it does with AdX, which means that they have to share real-time data coming from both platforms with all competitors participating in ad auctions, including with header bidding firm Prebid.

On the buy side, an important change will be the non-discrimination requirement for Google to operate AdWords (its most popular ad buying tool). If implemented properly, Google would not be able to route AdWords bids to Google’s own AdX and DFP, unless instructed by advertisers to meet return on investment goals. AdWords will also be prohibited from creating any direct bidding channels with DFP, but it can do so with other competitors such as Prebid.

As the ad tech firm PubMatic’s CEO explains in a blogpost, this opens the market for both demand and supply for the first time in about 15 years: “For advertisers, the principle is straightforward: budgets should flow toward the inventory and technology that best deliver their objectives. For publishers, more demand competing fairly for each impression could create more control over how inventory is sold.”

These remedies will mark a new beginning for how Google’s ad tech stack operates. At the same time, though, these were all mostly uncontroversial proposals from the plaintiffs and from Google itself. Enforcement will be key, so we have yet to see how the court-appointed monitor and technical committee will work.

A huge loss on this front is that the Judge rejected almost all anti-retaliatory provisions, reduced the compliance period from the plaintiff’s proposed fifteen years to Google’s six, and spared Google’s DV360 platform—used by big advertisers to buy ads across multiple formats, including on publishers—from being subject to any data sharing or interoperability orders.

Without divestitures, though, the duration of the ruling will also be shorter: six years, down from the 15 years proposed by the plaintiffs. Considering what it will take to implement enforcement mechanisms, the expectation would be for Google to lose its monopoly over ad tech in essentially five years.

What about the publishers?

The memorandum also suggests that the market participants who were harmed the most by this monopoly were publishers—among which news publishers were key witnesses of the case. Although the plaintiffs did not bring forward a case for financial restitution, they did include a provision for Google to establish an escrow to be funded by 50% of Google’s revenues from AdX and DFP (since they were tied). This remedy represented disgorgement of undue gains from monopoly rents.

Plaintiffs proposed the escrow to fund the administration of DFP by an open-source entity, as well as to defray costs from publishers switching out of DFP, and for any other uses the court would deem appropriate, such as establishing some sort of compensation mechanism to publishers, as my employer, Open Markets’ Center for Media & Digital Governance, proposed.

But the Judge rejected the entire proposal, essentially for a couple of reasons: because the plaintiffs included DFP, and the Judge didn’t see it appropriate since the liability ruling didn’t find DFP fees problematic, and also because there was no explanation about how the funds would be distributed among publishers — all issues, however, that could have been ordered to be specified by Plaintiffs.

There is one silver lining, though. Google will have to provide publishers with historical and configuration data from its publisher server, DFP, which publishers will be able to export to competing ad tech platforms, potentially enabling them to switch out of DFP as well. Additionally, they will get AdX bid data that will allow them to see, practically for the first time, what bids for ad spaces won or lost an auction through AdX.

What this means for antitrust enforcement in the US

Taken altogether, the ruling delivered somewhat effectively over low-hanging fruit: an injunction, prohibition of the unlawful conduct, and a change of practices and policies that even Google’s personnel questioned during trial. The effects should be global as well. However, read alongside the Google Search remedies ruling, the Google Ad Tech case changes very little for Google’s monopolies.

More significantly, the two cases send the unfortunate message to the market that antitrust violations in the technology sector are practically shielded from real deterrents to break the law. As my colleague Sandeep Vaheesan writes: “Despite being found to have broken federal antitrust law in two cases, Google generally gets to keep its unfair advantages in the market. As a result, in the words of a 1947 Supreme Court decision, in both cases, ‘the Government has won a lawsuit and lost a cause.’”