Last Wednesday, on 16 December 2020, Texas and nine other US States filed an antitrust lawsuit before the District Court for the Eastern District of Texas with respect to Google’s conduct in ad tech (reminder: almost two months earlier the Department of Justice sued Google with respect to its practices in general search; click here for Damien’s analysis). The lawsuit takes issue with various Google practices in the arcane world of ad tech, where publishers monetize their ad inventory and advertisers market their products with the help of ad tech vendors such as publisher ad servers, ad exchanges and Demand-Side Platforms (DSPs). This a strong lawsuit showing how Google has used each and every opportunity to distort competition and act against the interests of its customers, be it through unilateral conduct or even unlawful agreements with rivals.
As a background, Damien and I have written extensively on these issues, exposing Google’s anti-competitive practices in a series of papers starting in late 2018 (see here, here and here). Our papers explain how over the years Google has used its position in the ad tech value chain to engage in a variety of discriminatory and prima facie anti-competitive practices while also charging fees that are not transparent to publishers and marketers. We also explained how, instead of removing our concerns, Google’s recent switch to a unified first-price auction continues to tilt the balance to its favor while injecting additional opacity to its (already obscure) auction dynamics.
In the meantime, we have also written extensively on Google’s decision to phase out support for third-party cookies on Chrome, explaining how this may further undermine rivals and benefit itself (note: the CMA is currently looking into this issue following a complaint filed against Google by an association).
It is thus only natural that US States scrutinized Google’s conduct in ad tech and eventually filed an antitrust lawsuit. But the evidence uncovered by Texas shows that things are actually quite worse than what we had originally thought. I hereafter summarize the complaint’s key points:
(1) The lawsuit largely confirms our analysis and is in line with the findings of the CMA in its report on online platforms and digital advertising. According to the lawsuit, Google wields monopoly power across all relevant markets: publisher ad servers, display ad exchanges, display ad networks and display ad buying tools for small marketers. The CMA similarly found that Google has the strongest position across the ad tech value chain, boasting market shares of at least 40% (and in some cases up to 90%). The lawsuit then takes issue with a series of Google practices we have also criticized, such as (1) using its publisher ad server DoubleClick for Publishers (now Google Ad Manager) to grant preferential treatment to its ad exchange AdX (now part of Google Ad Manager) through features such as Dynamic Allocation and Enhanced Dynamic Allocation, (2) forcing publishers wishing to access demand from its display ad buying solutions to use its ad exchange and its publisher ad server; (3) restricting information rivals could access to advantage itself; (4) trying to quash an industry innovation that promotes exchange competition called Header Bidding through a series of measures (including the design of the Accelerated Mobile Pages format); (5) Google’s switch to a unified first-price auction; and (6) forcing marketers to use Google’s ad buying solutions to purchase YouTube inventory. A common theme throughout the lawsuit is Google’s severe conflicts of interests (acting as the buy-side, the sell-side, and the exchange in the middle), which would be unimaginable in financial exchanges. Google’s practices are said to have had anti-competitive effects across the ad tech value chain to the detriment of marketers, publishers, ad tech vendors and ultimately consumers.
(2) The lawsuit relies on a wealth of evidence based on Google’s internal documents. Unfortunately, the juiciest parts are redacted from the complaint but still the evidence seems quite damning. For instance, the lawsuit cites documents where Google employees agreed that in the future they should not directly lie to publishers but instead find ways to convince them to act against their interests. Another example is how Google would tell its customers (publishers) that its Open Bidding feature enabled competition and improved revenue while it had secretly designed it to favor AdX and foreclose competition at the expense of publishers.
(3) What is truly new in the Texas lawsuit is the allegation that Google colluded with Facebook to undermine Header Bidding. The complaint explains that when faced with the threat of Facebook supporting Header Bidding (which could erode Google’s monopoly in ad tech), Google tried to bring Facebook to the negotiating table. The outcome was a formal agreement between the two tech giants that Google internally gave it a code name based on a Star Wars character. The deal was the following: Facebook would significantly curtail its Header Bidding initiatives and in exchange Google would grant Facebook a series of advantages in its auctions – including a portion of auction wins. These are very serious allegations, and seem to be based on written evidence (such as the agreement between Google and Facebook, which apparently was known to top executives). If proven, they could very well establish the existence of a cartel between Google and Facebook (an infringement of Section 1 of the Sherman Act), and lead individuals behind the bars.
(4) The lawsuit explains how Google has used privacy concerns to advantage itself, but violates privacy when doing so is convenient. The complaint refers to an agreement with Facebook according to which Google was granted access to millions of encrypted WhatsApp messages pursuant to a 2015 agreement. (Note: Facebook claims access to encrypted messages is technically impossible). The complaint also refers to a 2019 closed-door meeting between Google, Facebook, Amazon, Apple, and Microsoft, whereby Google was allegedly interested in coordinating to lobby the government to delay or shelve privacy legislation. These are very serious allegations. If proven, they could change profoundly how regulators (and lawmakers) approach Google’s tactic of justifying various controversial practices on privacy considerations (think of the deprecation of third-party cookies in Chrome and the Privacy Sandbox). Google will simply lose credibility with regulators if behind closed-doors it lobbies against privacy legislation. (By the way, Google has been recently fined by the CNIL in France for dropping cookies without having first obtained user consent in breach of the ePrivacy Directive). In any event, as Damien and I have explained, Chrome’s policy change serves to distract attention from the fact that the most pervasive user surveillance takes place on Google and Facebook’s own properties. This is why its privacy benefits are likely to be limited in practice.
(5) The lawsuit asks among others for structural relief, i.e. a breakup of Google’s ad tech business. Considering how Google’s ad tech suite is rife with conflicts of interests, this is likely to be the only way effective competition can ever be restored in ad tech (in fact Damien and I have urged the CMA to consider structural solutions, but arguably US regulators are best placed to pursue such remedies).
All in all, this is a very strong lawsuit, and Google seems to be in a tight corner. Meanwhile, its ad tech practices continue to attract regulatory attention across the world (e.g., UK, Europe and Australia). We wouldn’t be surprised if the collusion allegations in the Texas lawsuit led to fresh antitrust proceedings in other countries and follow-on damages actions.
P.S. In other news, last week Google was hit with another antitrust lawsuit, this time from 38 US States, led by Colorado, with a focus on specialized search.