USA v. Google Weeks 9 & 10
Google passed through the looking glass in Week 9. It is naturally the goal of their defence to challenge the key tenets of the DOJ’s case — chief among these being the anticompetitive nature of Google’s myriad payments to secure default status. But this week, Adrienne McCallister, VP of Global Partnerships, launched an extraordinary argument that these payments are by nature pro-competitive. Unpacking McCallister’s justification for this point is spelled out below.
Differentiation and Competition in End User Devices
McCallister has worked for Google for twelve years. As is consistent with other long-term Googlers that the defence has called as witnesses, much of her testimony portrayed Google as a force for good. According to this viewpoint, Google’s RSAs improve competition in the smartphone market. These agreements accomplish this feat, as McCallister suggested, because they help to differentiate Android phones from Apple phones. One might have thought that for consumers the features, functions and prices of Google and Apple mobile devices would achieve differentiation as they do in all electronics goods markets without the need for the extra payments. Indeed, the payments appear to be designed to restrict differentiation by distributors. The DOJ produced an email exchange from 2020 in which one of McCallister’s colleagues flagged to her that AT&T, one of Google’s largest RSA partners, ‘want to differentiate the experiences [on Android] so not all Android devices look the same.’[1] The frustration AT&T feel is with the restrictive parameters of Google’s deals that enforce homogeneity to drive search revenue.[2]
Which is right? Google’s view that payments help to keep Android phones distinct from Apple handsets, or distributors such as AT&T who are looking to sell mor handsets that users want to buy?
The DOJ also produced evidence that Google’s RSAs limit players such as Samsung from building a competing apps store and Apple from building a competing search capability. The DOJ’s series of example certainly appear to build a picture of restrictions arising from the agreements rather than the opposite. This point is significant because it administers a healthy dose of suspicion to the Court’s reception of Google’s arguments on this front. Such suspicion is vital for an unimpaired analysis of the wider arguments Google made about the RSAs this week. Their expert witness, Professor Kevin Murphy, made the bold claim that Google’s default agreements are themselves the result of a competitive process; companies compete to secure these valuable market arrangements and the company with the best bid — historically, Google — emerges victorious.
In response, the DOJ’s expert witness, Professor Michael D. Whinston, returned to the stand armed with the simple rebuttal that, ‘when bidding for an exclusive contract, a dominant firm can use the monopoly profits it protects to make sure it wins.’[3] Whinston’s point identifies the source of Google’s success in these ‘auctions’ — to borrow the defence’s term — as a direct consequence of its pre-existing market share. The ‘competition’ for RSAs is thus implicated in the economies of scale and network effects that have subsumed digital markets — eliminating competition.
Is this a Library? Is this a Library?
The DOJ also referenced Elizabeth Reid’s comment from Week 8 that Google’s goal was ‘organizing the world’s information … like a library’ for a scathing purpose in this week’s testimony — to emphasise the gap between the knowledge-work accomplished by Google’s search engine and its revenue-generating function. How far Google functions as a ‘library’ and as a modern technology for information storage and review or is it just an ad vehicle? To some, Google might be only the latest iteration of the Enlightenment pursuit of an ‘encyclopaedia’, or a total knowledge system. However, as with Diderot’s original ‘encylopedie’, Google’s data-system structures its knowledge along certain ideological fault lines.[4] It’s not just that any attempt at organising the world’s information will have in built flaws, it is that the Google system has in built bias because its funding drives it to categorise, limit and distort results of its searches for its commercial ends.
What are commercial queries have become known during the trail; searches for a house or a car or insurance are top commercial questions. Each give Google the opportunity to present results where it gets a benefit from doing so- and commands a slice of the action, whether through deals with certain websites over others or through advertising. Non-commercial or less obviously commercial queries are also used by Google to inform its view of which cohort or marketing category to place the user in for advertising purposes. Tempting a person who has a search history of Google ancient temples in Greece with a Greek holiday is a different commercial proposition to someone with a search history of searching for Magaluf or Aya Napa.
This point is worth emphasising as we reach the closing stages of the trial because it illuminates the remedy that is feasible at the end of these proceedings. If the Court finds against Google, it will have an opportunity to restructure how our modernity organises and accesses knowledge. Defaults could be banned but what good would that do? Apple will still have $19 billion reasons not to build its own search engine. Would it be possible to leave Google with a search infrastructure and relevance engine and enable competitors, including Apple and others, to hook up their own search results pages to Google’s relevance engine to create a search competitor?
The DOJ is asking for Google to be broken up – so how it can be restructured to remediate the competition that it has eliminated along the way is key. A separate “pure search” business solely focused on retrieving results that are relevant to queries is attractive, and could be used by many, but will it be possible when the evidence has revealed that all functions, including the browser and the way the results are displayed, are optimised for advertising?
[1] Trial Exhibit – UPX0997: U.S. and Plaintiff States v. Google LLC (justice.gov), p. 1.
[2] There is other evidence of RSA partners being dissatisfied with Android UX in https://www.justice.gov/d9/2023-11/417997.pdf. See also Trial Exhibit – UPX0129: Google presentation: MADA & RSA: Android Commercial Agreements (justice.gov) for evidence of the link between Android UX and Google’s revenue generation.
[3] Trial Exhibit-UPXD106: U.S. and Plaintiff States v. Google LLC. (justice.gov), p. 26.
[4] See Paul K. Saint-Amour, Tense Future: Modernism, Total War and Encyclopedic Form (Oxford, 2015).
Header Image courtesy of anonymous James on Unsplash (Licensed for free via Unsplash Licence).