Seven Takeaways from the Justice Department’s Antitrust Case Against Google’s Search Deals
The relative roles of scale, data, defaults, and quality in Google’s success is as clear as mud
With Vidushi Dyall
We’re now halfway through the Justice Department’s 10-week trial in its antitrust case against Google’s search default deals, and DOJ has rested its case. The state attorneys general are now trying their case — focused on Google’s SA360 product — and then Google will begin calling witnesses in its own defense.
As trial watcher and former DuckDuckGo exec Megan Gray noted, this moment is the “absolute high point of the plaintiffs’ case”:
So, how did DOJ do?
The plaintiffs have the burden of proof in the case. But a few dynamics have made this challenging to assess how well they made their argument:
- First, because witnesses have appeared based on their availability, the plaintiffs haven’t presented their case in any consistent thematic order. There’s an economist witness covering search engine scale one moment…then an advertiser discussing SA360…then a search competitor testifying. The “market power” witnesses have been interspersed with the “monopoly maintenance” witnesses;
- Second, beyond the daily legal newswires, mainstream press coverage of the case has tended to focus on hot documents, news nuggets, and the most high-profile witnesses — and less on how well the plaintiffs have done in establishing the core of their case around Google’s search defaults; and
- Third, a lot of the press coverage of the trial has focused on how transparent the trial has been, not on the facts and theories being presented.
To help address this, my colleague Vidushi Dyall — who has been in court every day since Day 9 — created a Witness Tracker, looking at why each witness was called (what part of the plaintiffs’ case they were meant to bolster), the key wins that the government secured from their direct examination, and the key admissions Google secured from their cross-examination:
Based on Vidushi’s analysis of each witness’ overall testimony (direct plus cross), and where each witness fits in DOJ’s overall case, I identified seven key takeaways from the government’s case:
1. DOJ’s Search Market Definition Rests Largely On One Witness’ Shoulders
In order for DOJ to succeed at the core of its case, it must first establish that Google has market power in the two markets it defined, the first being “general search engines.” DOJ asserts that Google only competes against Bing, Yahoo, and DuckDuckGo in general search, which Google counters that it competes against specialty search sites like Amazon, Yelp, and TripAdvisor, especially for commercial queries.
Microsoft CEO Satya Nadella tried to bolster DOJ’s argument by calling the internet “Google’s Web,” and Apple’s John Giannandrea stated his belief that Amazon and Expedia are not search competitors to Google.
But statements from competitors probably aren’t sufficient for market definition. To a large degree DOJ’s “general search engine” market definition argument fell to one of its economist witnesses, MIT’s Michael Whinston, who listed a series of criteria of a general search engine.
But after listing several of those traits — including “catering to habit formation,” “less time/energy used to recall a specific site,” and “depth and breadth of results” — Whinston admitted under cross-exam that these criteria were also true of specialty sites like Amazon or TripAdvisor.
Furthermore, Whinston seemed to overlook Google’s share of valuable shopping queries (which are only about 20% of all queries) — where Amazon has a lead:
Google can lose on this market definition aspect of the case and still prevail — but Whinston’s testimony has to persuade Judge Mehta in order for the rest of their case to even have a shot.
2. Is The “Marketing Funnel” Relevant and Rigid, or Outdated and Flexible?
DOJ also claims that Google has market power in “general search text advertising,” and called multiple witnesses from the advertising industry — including ad leaders at Home Depot, Booking, JP Morgan Chase, and Universal McCann — to assert that search advertising is distinct from other forms of online advertising.
The “marketing funnel” was presented as a visual representation of the idea that ads at the top of the funnel (say, display ads for brand awareness) are a distinct market from intent ads at the bottom of the funnel (like Google search ads).
[Aside: if a trial watcher took a drink every time the “marketing funnel” was mentioned during this case, they’d be facedown on the courthouse floor by noon.]
The problem for DOJ is that many of the same witnesses who testified that search ads are unique — including the witnesses from JP Morgan Chase, Universal McCann, and an economist from Columbia Business School — also admitted under cross-exam that their firms believe that the marketing funnel is outdated, overly rigid, and too linear for the fluid “journey” that customers go on today.
Again, Google can lose on this part of the case and still prevail — but DOJ’s reliance on the “marketing funnel” to establish this part of its market definition was undermined by their witness’ cross-examinations.
3. The Relative Importance of Defaults, Scale, Data, and Recipes is as Clear As Mud
Moving on from market definition, the heart of the case is whether Google’s search default deals — with Apple, Mozilla, and Android OEMs like Samsung — give it an insurmountable advantage in search that no one can hope to beat.
While DOJ argues that Google’s high share of search queries is largely due to the data and scale that default deals provide it, Google agrees that defaults are valuable — but argues that they are trumped by the quality of the search engine. Each side offered arguments bolstering its case, as I summarized here:
The picture that emerges from DOJ’s witnesses is disagreement about the reasons for Google’s success. Google’s Hal Varian argued that it was more about the “quality of Google’s search recipes” than the amount of data Google has — and we’ll surely hear more Google-called witnesses speak to their investments in search quality. And Neeva’s Sridhar Ramaswamy acknowledged that a search engine only needed 2.5% of queries to have sufficient data and scale.
On the other hand, Microsoft’s Mikhail Parakhin described a “feedback loop” of more users leading to more clicks leading to better quality.
Meanwhile DOJ witness Antonio Rangel spoke to the power of defaults — but Microsoft’s Jonathan Tinter acknowledged that 75% of Microsoft Edge users switch their default search engine from Bing to Google. This was also true of the Bing defaults on Verizon Blackberry Devices, all Nokia devices, and Microsoft’s own Surface Duo phone.
So how powerful are defaults, really?
Watching this debate play out, it’s very difficult to see how Judge Mehta will be able to sort out a definitive, technical answer to how much of Google’s search success is due to engineering, scale, data, and default deals. DOJ’s case rests on a simple argument that Google’s default deals were more important to Google’s success than any other factor — but that is far from clear.
4. The Two Faces of Microsoft
Microsoft would be the biggest beneficiary of a DOJ win in this case, as Bing would be well-poised to gain search share if a court placed restrictions on Google’s search deals. Not surprisingly, four different Microsoft witnesses testified for the government.
But the trial revealed that Microsoft tells Wall Street one thing and courts another.
As tech journalists Casey Newton and Kevin Roose noted on their podcast, only a few months ago Microsoft CEO Satya Nadella said that Bing’s AI advancements could help it beat Google. But he told the court the opposite — that Microsoft was a victim in search who has no hope of catching up.
As Newton says, “to whom was he telling the truth?”
Nadella wasn’t the only one. While Microsoft’s Mikhail Parakhin claimed Bing can’t compete with Google’s “feedback loop”, he was also forced to acknowledge under cross-exam Bing’s “negative feedback loop,” with Microsoft failing to invest in and properly staff its Bing engineering team.
You can’t begrudge Microsoft for seeing this case as a big opportunity — but its claims in court contradicted its actions and public rhetoric, raising questions about the real truth.
5. Were DOJ’s “Victims” Compelling or Lacking?
DOJ’s opening statements at trial identified three “victims” of Google’s search default deals: DuckDuckGo, the search startup Neeva, and an “app search” service called Branch Metrics. Each company testified for DOJ — but holes emerged for each:
- DuckDuckGo’s Gabriel Weinberg said search defaults were too hard to change, and endorsed “ballot screen” remedies — but acknowledged under cross-exam that DuckDuckGo’s market share in Europe barely budged when consumers were given a ballot-screen choice.
- Neeva’s Sridhar Ramaswamy blamed Google’s search deals for Neeva’s demise — but acknowledged that Neeva charged consumers $5/month for a service that most customers were accustomed to receiving for free — and that Neeva was acquired for two times its seed investment (hardly a victim).
- Branch believed its partnership with Samsung sputtered due to the latter company’s search default deals with Google. But Branch’s Alex Austin said it never saw itself as a competitor to Google — and both Austin and Samsung’s Patrick Chang testified that Branch’s failures had a variety of factors unrelated to Google.
6. Apple’s Search Deals Are the Result of Both Negotiation and Testing
In its questioning of Google’s Joan Braddi about Google’s search default deals with Apple, DOJ attempted to suggest that Google muscled Apple into unfavorable contract terms — and that Google was in the driver’s seat of those negotiations.
But Apple’s John Giannandrea and Eddy Cue stated clearly that they partner with Google because they have tested and verified that its results are the best. And Braddi herself said Google’s negotiations with Apple involved give and take from both parties, like any business negotiation.
While some have alleged that Apple’s deal with Google discourages Apple from launching its own search engine, not a single witness testified to Apple ever considering doing so. And if Apple’s search default deals were solely about picking the partner willing to pay the most, Nadella’s testimony revealed Microsoft would have paid even more, but Apple rejected Microsoft’s offers because it saw Bing as inferior to Google.
7. There are Boring Explanations for SA360 Not Having all of Microsoft’s Desired Features
In the most obscure part of the trial, the state AGs are alleging that Google intentionally and illegally refused to integrate Microsoft’s ad platform features into Google’s own cross-platform search engine marketing (SEM) too, SA360.
Setting aside the legal question of whether Google owes any legal obligation to Microsoft here, from a factual perspective this claim hit the rocks when:
- SA360 Product Managers Ryan Krueger and Amit Varia described how their engineering team had to prioritize adding new features based on advertiser demands — not Microsoft’s;
- Home Depot’s Ryan Booth acknowledged that SA360 didn’t even incorporate all of Google’s own ad platform features; and
- Booth acknowledged that SA360’s competitors at Adobe and Marin don’t support all of Microsoft’s features either.
The idea that Google slow-rolled Microsoft over SA360 out of spite or competitive harm hasn’t been clearly proven — and we’ve seen plenty of plausible alternative explanations.
So why is this issue even being litigated at all? Well, most of the state AGs’ original case — on search “self-preferencing” — was tossed out before the trial, and this niche SA360 claim was all that remained. It’s been odd to spend so much time on an obscure claim that benefits only Microsoft — but I suspect it’s a function of state AGs wanting to preserve their seat at the “antitrust trial of the century.”
Peripheral Issues
There are a few other issues that have arisen in the plaintiffs’ case that have sparked a lot of debate and commentary, but which I believe are ultimately peripheral to the heart of the case. These include:
- Google executives’ “hot documents” on advertising. Some spicy emails and memos arose during the testimony of Google’s Jerry Dischler, Michael Roszak, and Adam Juda — several of which were focused on the inner workings of Google’s ad auctions and pricing. While these may end up playing a role in the Judge’s advertising market power analysis, they’re not central to the search default deal issue at the heart of the case.
- Chat history. DOJ quizzed several Google witnesses about their internal chat history policies, after suggesting in their opening statement that Google has been systematically hiding evidence. While these kinds of issues have resulted in sanctions in past cases, they’re also not central to the core legal questions around search defaults.
- Trial transparency. As a follower of this trial, I would love to see more testimony and exhibits — and I appreciate the media’s role in pushing for greater transparency. There’s a temptation by some Google critics to view “the trial as the remedy,” but some observers have put the trial’s optics and theatrics ahead of the core legal dispute.
Instead of closing arguments, the parties will file post-trial closing briefs with the judge a few weeks after the trial is complete. But given the burden of proof placed on DOJ, it’s not too early to assess the overall story they told.
And even if Judge Mehta ends up agreeing with the one or both the plaintiffs’ market definitions, it’s difficult to see a clear story from the government’s case that Google’s search defaults are not only the key driver of Google’s success, but ought to be forbidden outright.
Chamber of Progress (progresschamber.org) is a center-left tech industry association promoting technology’s progressive future. We work to ensure that all Americans benefit from technological leaps, and that the tech industry operates responsibly and fairly.
Our work is supported by our corporate partners, but our partners do not sit on our board of directors and do not have a vote on or veto over our positions. We do not speak for individual partner companies and remain true to our stated principles even when our partners disagree.
