Is Google’s Reckoning Finally Here?
The first big antitrust trial of this century is nearing its end. Looking back on the evidence we've seen — and ahead to Judge Mehta's huge task.
For the past two months, in a courtroom in Washington D.C., a judge named Amit Mehta has been overseeing the Google search antitrust trial, in which the government is seeking to address the firm’s ostensible market power. The stakes are massive, and not just because this trial is the culmination of years of political pressure and investigations from the Federal and state government, Congress and global enforcers. It is because of Google’s importance, and what it could mean for the rest of corporate America.
Google is more than just how 90% of us look for things, it is more than just a gateway to the internet, or a firm with 15 products that each have more than 500 million users. Its algorithms and business choices shape how we ask questions, how we communicate, and what political choices we make. Google organizes the news industry, the entertainment industry, and answers our questions about medicine. As one witness put it, it is no longer the open web, but the Google web. Or as Scott Galloway once said, “Google is God.”
This antitrust case, brought by the Department of Justice Antitrust Division, is the first real challenge to the firm’s immense market power, from which its political and economic power flows. And so how Judge Mehta hears the evidence, interprets the law, and how both Google and the government present their arguments matter.
It’s also been said that “the trial is the remedy,” and that’s because firms do worry about their reputation being harmed if they act badly and it gets exposed. Early on in this trial, that form of remedy was limited by frequent closures of the courtroom. But our coverage helped start an important conversation about the openness and transparency of proceedings. In the weeks since that conversation began, we’ve seen Judge Mehta take steps to open up the courtroom to the public, end most closed-door questioning of witnesses, and allow more internal documents and corporate data to see the light of day. We have helped, in some way, make antitrust trials more public, which is good for everyone.
With that said, yesterday was Yosef’s last day covering the biggest antitrust trial of the 21st century since he has another pre-existing professional commitment starting up. Don’t worry, Big Tech on Trial will continue. Lee Hepner, an antitrust lawyer with the American Economic Liberties Project, will be taking over to continue providing coverage of the final days of trial. Lee has also recently been filling in for Matt Stoller with the BIG newsletter, where he’s written some fascinating deep dives on labor unions and algorithmic pricing software.
For Yosef’s last piece, we thought we could zoom out a bit and attempt to provide a bird’s-eye view of what we’ve heard through the first nine weeks of trial. Consider this a rough overview of what the parties have argued over the last couple of months, with links to articles and trial exhibits if you want to go deeper on any of these points.
As a reminder, both the DOJ and the States have brought their claims under the law that prohibits monopolies, which is Section 2 of the Sherman Antitrust Act. This law requires them to prove that Google has monopoly power in a specific market or set of markets. This sounds obvious and pedantic, but defining a clear market is at the heart of any antitrust case. The law also requires the government to show that Google has maintained its monopoly through conduct that unfairly thwarted competition, versus just making a better product that consumers prefer using.
First, we’ll present the government case, then Google’s case, then share a few observations.
The World According to Federal Enforcers
The Antitrust Division at the Department of Justice has said Google has monopoly power in three separate markets all involving its search engine. These markets are called “general search services,” “general search text advertisements,” and “search advertisements.” Google has high shares in all of these purported markets, so the evidence has largely been focused on proving that these markets are properly defined antitrust markets. It’s a very nerdy exercise, but one where a lot of cases fail.
What is general search?
General search is what you and I think of as a normal search engine, Google, DuckDuckGo, or Bing — a starting point for asking questions, or what in the trial was called a “one-stop shop” for all kinds of queries. That’s how the government defines search. Google contested this definition and claimed it competes query-by-query with everyone from Expedia to Yelp.
The DOJ supported its market definition with analysis from its expert witness, MIT Professor Michael Whinston, as well as testimony and documents from different general search engines (including Google) that suggest general search engines view each other as their biggest competition. (See slide 6 from the demonstrative slide deck used with Whinston if you’re interested in more examples of this testimony.)
There are some intuitive points here on how to define a general search engine. A key part of the case is what browsers and smartphone developers — like Mozilla or Apple — consider putting as the default search engine. They have considered Microsoft’s bids for default placement of Bing — but they would never consider making a specialized vertical provider like Amazon or Yelp the default search engine. That is, the DOJ argues, proof that general search is a properly defined antitrust market.
Another core point of dispute is scale. A search engine is a learning engine, taking the results from one search and using them to improve the next time a user makes a similar query. Without enough data or users, which comes from access to distribution, argues the DOJ, a search engine won’t be very good.
The case from enforcers is that Google’s strategy is to deny scale to its rivals to prevent them from developing quality competitive products. Testimony from Microsoft executives highlighted the importance of scale in building a successful search engine — what Microsoft CEO Satya Nadella described as a “vicious cycle”. And despite significant venture funding and the leadership of a former Google executive, Neeva, a smaller rival, went out of business because it couldn’t obtain enough users.
Advertising Markets
But general search isn’t the only market. There’s also search advertising, which the government argues is a discrete market. This one’s a bit trickier. Google sells a lot of advertising, but it is obviously not the only firm selling ads. You can buy TV ads, or newspaper ads, or billboards. So in what sense is Google an advertising monopoly?
Well, the antitrust enforcers argue that Google controls a specific and essential kind of ad, which is to say, the ad provided to consumers in response to a search query that communicates commercial intent. Being able to target a consumer when he or she specifically asks about buying a product can be extremely valuable, and it is, according to the DOJ, a unique type of advertising. You cannot substitute other forms of ads for search ads, and that’s one definition of a market.
Marketing has a bunch of weird jargon in it, which turns out to be important. Search ads offer the ability to target consumers when they are about to buy, which is to say, lower down in what is known as “the purchase/marketing funnel,” meaning the way consumers go on a journey from learning about a product to choosing to buy it, and the different ads necessary at each stage. And the DOJ says that distinguishes search advertisements from other kinds of digital advertisements, which means it fits the market definition required to win a case.
Another way to show monopoly power is through what’s called “direct evidence,” which relates to the ability to control prices without facing competitive consequences. Here, the government explored Google’s power to manipulate prices in advertising auctions. Enforcers elicited testimony from Google advertising executive Jerry Dischler about Google’s potential ability to increase revenue by raising advertising prices from 5% to 15%. In an email Dischler sent to a colleague, he discussed “shaking the cushions” on ad launches to meet revenue goals, which to the government was clear direct evidence of monopoly power.
Defaults and $26 Billion of Monopoly Maintenance
Ok, so that’s the market definition piece. What about the second part of the case, which is showing that Google not only has high market share, but is actively locking rivals out of the market?
Here the key question is what are known as “default” deals that Google has with various partners like Apple and Samsung, which, through their control of phones and browsers, can put search engines in front of large numbers of users. When you open the Safari browser on your Apple iPhone, for instance, there’s a search engine placed there by default.
Defaults are powerful, as DOJ’s behavioral economics expert Antonio Rangel explained early on in the trial. Even if users know they can switch their default, most users won’t — and Google recognized early on how important defaults could be. Google’s power over defaults prevents rivals from being able to compete on the merits. Google pays a lot of money — $26 billion in 2021 — to make sure it is the default search engine in as many “search access points” as possible, thereby denying competitors meaningful opportunities to obtain users and scale. Even Google recognized that search engine defaults threaten competition, as its general counsel wrote to Microsoft in 2005.
According to the DOJ, Google’s default search engine deals not only prevented other general search engines like Microsoft, DuckDuckGo, and Neeva from being able to effectively compete, but they also prevented a young company called Branch — which offered in-device search capabilities similar to Spotlight Search on iPhones — from ever being able to get off the ground.
DOJ’s examinations of Apple witnesses focused on Apple’s development of its own search technology, suggesting that Apple may have built its own competing search engine if it weren’t for the massive payments it received from Google — reported by the New York Times to be roughly $18 billion in 2021.
The alleged harms of this monopoly maintenance are what you’d expect. The government argues it allowed Google to raise prices for advertisers without regard to the prices of ads on other digital platforms, and has allowed Google to forgo quality improvements in privacy and other areas that it would have otherwise pursued. Consumers are also deprived of the potential for a higher-quality search engine that could emerge if there was healthier competition.
Destroying Evidence
There’s also an element of this case that goes beyond Google’s power and conduct in the market, allegations of hiding evidence. These efforts included widespread use of history-off chats to hold substantive business communications, as well as adding in-house attorneys to emails to shield the emails from discovery under attorney-client privilege. Google also instructed its employees to avoid using certain words and phrases that might trigger scrutiny from regulators. In the government’s opening statement, DOJ’s lead attorney stated outright, “They turned history off, your honor, so they could re-write it here in this courtroom.”
Any trial is about credibility, so if the judge thinks that Google cannot be trusted, he should be more likely to rule against the search giant. The history-off chats are the subject of a still-pending motion for sanctions against Google, but the DOJ also wants Judge Mehta to infer that they demonstrate Google’s anti-competitive intent.
More Enforcers! The States Get Their Say
Google isn’t just facing the federal government, but a set of state attorneys general who have brought additional claims. The case initially brought by the States included the claim that Google had weakened “Specialized Vertical Providers”, like Yelp and Expedia, by limiting their visibility on search results pages so that Google could more prominently feature its own verticals. This claim was tossed out before the trial began, but the States have pressed another unique claim forward.
The States are focused on some of the technology underpinning search advertising. To place advertisements on Google and other search engines, many advertisers use search engine marketing tools that allow them to coordinate their advertising campaigns across both Google and Bing. Perhaps it’s not surprising that Google operates the largest of these search engine marketing tools, SA360.
The States argue that Google has harmed competition by delaying the implementation of important features on SA360 for placing ads on Bing. The focus has been on a feature called auction-time bidding, which allows advertisers to maximize the efficiency of their ad spending by adjusting their bids on advertising auctions in real time. SA360 implemented this feature for Google Ads in 2019, but they still haven’t rolled it out for Microsoft Ads.
This delay has not only harmed Microsoft, but it has also deprived advertisers of the benefits of having a competitive search advertising market — and the States have tried to suggest that this kind of delay was intentional. A much smaller search engine marketing tool, Skai, was able to successfully develop auction-time bidding for Microsoft Ads all the way back in 2020.
The World According to Google
Ok, now let’s go to the defense. Google, as you would expect, is contesting every part of the government’s case.
First, Google argues that it doesn’t have monopoly power in any properly defined relevant market. Instead of the markets alleged by the government plaintiffs, Google argues it fiercely competes against all kinds of specialized vertical providers for search users, for example, Amazon for shopping-related queries and Expedia for travel-related queries, as well against all kinds of digital platforms for search advertisers.
To make this case, Google relied heavily on the market definition analysis of its economic expert, Dr. Mark Israel of Compass Lexecon (last seen in the Antitrust case against JetBlue-American Airlines). Dr. Israel opined that the DOJ and States “defined away” Google’s strongest competition in both search and advertising. Throughout the trial, Google employees testified they perceive fierce competition for different kinds of queries from companies like Amazon, Expedia, and even TikTok.
Google also contested the idea that its search ads are a unique form of advertising, arguing advertisers shift their ad spend between all kinds of digital platforms based on the expected return on investment, and that the purchasing funnel is an outdated concept that over-simplifies how the real world works. Google, because of this competitive pressure, aggressively innovates in both responding to search queries and serving advertisements, which is why search and search ads are growing, hardly the sign of purportedly monopolized markets. If advertisers are paying higher prices for search ads, it is only because they are receiving higher value.
Google also disputed the States’ claim that they intentionally delayed the roll-out of features for rival Microsoft Ads on SA360. It took years for Google to build auction-time bidding for its own advertising auctions, and it has been actively working on building the feature for Microsoft Ads without any kind of intentional delay.
So that’s market definition, and the States’ case. Now let’s go on to what may be the core of Google’s defense, which is that its default deals are not actually harming competition. First, Google has argued that it won default status with partners like Apple and Mozilla not just because it paid money, but because it has the best products on the merits. These companies chose to make Google the default over bids from Microsoft and other rivals, because defaulting to Google search provided users with the best experience. Users are still able to easily change their default search engine if they want to, and other search engines pay to be featured as default options for different browsers. These aren’t exclusive deals designed to maintain a monopoly, they are reasonable contracts won fair-in-square in the marketplace.
Second, Google has argued that scale is not very important in the quality of search, so the idea that it is thwarting competition by denying rivals’ access to user data doesn’t make sense. According to one Google expert, Virginia Tech computer science professor Edward Fox, user-interaction data explains only 2.9% of the quality gap between Google and Bing. Google is the best because of its innovations, investments, and superior engineering. Google just has special sauce.
Aside from dismissing default placement and scale, Google further justified its conduct by arguing that its deals produced what are known in legal jargon as “pro-competitive benefits” — basically the deals have been good for competition overall even if they also have the effect of excluding others from the market. The main idea is that other firms used the money they got from Google to develop new products, enhance existing products, sell their offerings to consumers at lower prices, and offer security upgrades. In other words, Google’s default deals have fueled intense competition in “outside markets” like internet browsers and mobile devices.
Without Google’s control of Android, there wouldn’t be as many Android device-makers making cool smartphones, and so competition with Apple’s iPhone would be much less intense. The free licensing of open-source Android software is only possible because of Google’s distribution and revenue-sharing agreements with partners, which provide Google with the revenue to continue investing in the Android system.
Fundamentally, Google’s innovations have benefitted consumers and advertisers, as evidenced by the growth over time in search query volume and digital advertising spending.
Going Forward
While Big Tech on Trial will continue its coverage, this will be Yosef’s last article. For Yosef, attending and covering this trial has been an incredible learning experience — and we are both so thankful to all of you who subscribe to BIG for making it possible. It was really important to shine a light on this trial, and that mission became even more important early on in the trial amidst its “unprecedented secrecy.”
We are also very grateful to all of the Big Tech on Trial readers, whose interest and engagement have provided the motivation to follow the proceedings as closely as we could.
Yosef is going to shy away from making any final predictions about how the case will turn out. We’ve seen some positive signals for DOJ’s case, but Judge Mehta keeps his cards close to his chest — and there’s too much the case can turn on to have a high degree of confidence in either direction. Getting the facts straight is tricky enough. On top of that, Judge Mehta will have to resolve unsettled legal questions, make sense of conflicting analyses from highly-credentialed economists, and potentially weigh all kinds of effects against each other to determine the net effect of Google’s conduct. If he does rule against Google, then what? Judge Mehta will have to then do another procedure to deal with remedies. And regardless of how he rules, this case could very well wind its way up to the Supreme Court on appeal.
In his piece in late August previewing the trial, Matt highlighted the role of an iconoclastic judge named Harold Green in breaking up AT&T in the 1980s, which happened in the same district court as this trial. In many ways, big decisions about our economy can come down to which judge gets assigned to which case, and how he happens to think about technology and markets. We don’t know how Judge Mehta will rule. But we do know that he has a huge task ahead of him, and how he approaches it will shape our future on the internet and beyond.
That’s all we have for now. Keep an eye out for Lee’s articles covering the final days of the trial.
Incredible coverage, thank you both for devoting the time to this complex matter.
It seems blatantly obvious that a company cannot grow so explosively by sheer innovation (secret sauce, n/w/s), but the obvious must survive our hopelessly byzantine legal system.
Any chance the three of you could find time to do a roundtable on:
1) In a perfect world, I would reorganize/breakup Alphabet Inc. by BLANK.
However, since we have yet to perfect this world,
2) the current legal possible ways to reorganize/breakup Alphabet Inc. I would employ are BLANK.
Thank you for all your hard work!