The Best Arguments Google Can Buy?

Judge Brinkema abruptly terminates a Google witness' testimony because it wasn't adding anything else to the case.

In one internal memo, Google considered opening up its publisher ad server (DFP) in an attempt to expand traffic on Google’s Ad Exchange, where it takes a 20% fee. Alas, they did not.

I was really excited for the first full day of Google’s case-in-chief on Monday, after beginning in earnest last Friday afternoon. I wanted to see Google give us its best stuff, and take the government’s case head on. In the same way a neutral football fan watching a game wants to see big plays, a competitive, talented back and forth, and compelling moments, I wanted to see a compelling legal battle play out in court after two weeks of seeing the government, in many ways, run up the score.

I wanted Google to make it interesting. I know they have the ability, and I know there are arguments for them to make.

Google’s legal team is in many ways, a dream team, full of the most expensive big name lawyers corporations can buy. They are extremely talented, and save for an exhibit that went awry to begin their case, the quality control, poise, and raw ability has been apparent throughout. While the government has phenomenal lawyers, especially its core team that has worked tirelessly, and performed brilliantly on this case, on average, one could make the case that with Bill Isaacson, Jeannie Rhee and Karen Dunn, Google has an edge on star power and talent.

For those who are lawyers and have practiced at big firms, the pedigree and expectations of attorneys like these does not require much of an explanation. For those outside this cloistered paramilitary division of corporate America that gets deployed into cases like these to defend companies in existential legal battles—understand that these are the best of the best, from the best firms, the best law schools, and of whom we would have the highest expectations.

And so it is entirely baffling to me, and to others, the fact that, as of Monday evening, two days into Google’s case, and potentially 48 hours from the end of it, as well as the end of this trial—Google does not seem to have a case to match the government blow for blow. Or if it does, is not presenting it.

Before I had the chance to speak to some others during a break, I was thinking hard about what I was seeing and hearing. Could Google really be doing this? “Am I missing something?”, I thought to myself, while watching direct examination of Google witnesses who did not seem to be adding any important evidence to this case, or at least evidence that was meaningfully applicable to the core legal issues at play. During the break, others were saying the same things. By the afternoon, amazingly, Judge Brinkema removed any doubt—telling Google’s lawyers that the Google witness on the stand, Marco Hardie, in essence, was adding nothing, leading to the abrupt end of this direct examination.

Here I’ll give the usual summary of what happened on Monday. It is shorter than usual, because as I’ve alluded to, there was not much of substance or consequence that happened. Below that I’ll give some legal analysis on the core legal issues in this case, and talk about what Google might be trying to do.

The day began with Nitish Korula, a former Engineering Director of Ad Serving & Ad Quality at Google. It is interesting to see Google on direct examination and the government on cross—the roles reversed at long last.

Mr. Korula was asked if a publisher is forced to use AdX—he said no, in fact that they must be vetted, pounding once again the pretext of “brand safety” that makes its obligatory daily appearance in Google’s case.

Mr. Korula then said that “This is our product” and that “we want to make sure it is quality.” This testimony, taken together with the daily “brand safety” recital (my colleague Arielle Garcia has humorously included it on a crossword puzzle she made for watching this trial) from Google, seemed aimed at arguing that disintermediation (the potential remedy, or alternative to how Google currently operates) could have negative and even unintended consequences on the ad tech ecosystem.

This is a common refrain from big tech companies facing legal challenges on competition issues—to point to “security and safety” as a sort of catch all talisman to ward off antitrust challenges. In other words, allegedly self-preferencing or anticompetitive behavior is necessary, as only the firm and the firm alone can keep its platform safe, whereas rivals could introduce security challenges without the firm’s gatekeeping in place.

For example, Apple is notorious for this—alleging its closed down App Store and the “vetting” therein is necessary to protect users, leaving it free to self-preference and implement anti-steering rules that disallow developers from collecting payment or distributing apps anywhere besides through the Apple App Store (which of course also happens to have a 30% fee due to Apple for every sale).

Mr. Korula then walked the Court through how one would use Google Ad Manager (GAM) if they were an advertiser looking to set up a campaign. It was probably unintentional, but I chuckled at the irony of Mr. Korula during this demonstration saying, “this is what you would see if you search Google Ad Manager with your favorite search engine,” when showing an exhibit of the GAM webpage. (Google Search was declared an illegal monopoly last month…)

This testimony seemed aimed at highlighting Google features and protections offered to advertisers. The government did not really show us any of this, so it was interesting to actually see under the hood. It’s also debatable how relevant it is. I do not believe any witness said they were confused as to how to operate Google’s products.

The other point here, given the extended amount of time spent highlighting all of the brand safety controls, may be to highlight how a potential remedy that results in disintermediation could complicate publisher and advertiser concerns over the safety of content and placement of ads.

If it was not clear before, it’s certainly clear now: The government is alleging illegal “tying” of three separate product markets (publisher ad server, ad exchange, and ad network), and Google wants to make this a “refusal to deal” case.

Judge Brinkema interjected to ask Mr. Korula if a publisher using Prebid in Google Ad Manager 360 can always set their own price, to which Mr. Korula said “yes.” It’s always noteworthy when the judge asks questions. I think Judge Brinekma’s questions, more than anything else, reveal that she fully understands the issues in this case.

I noticed Google’s lawyers were running a lot of notes to the podium during this direct examination. This might just be how Google’s legal team operates on direct examination (after all, they just started) or it might also mean things were coming up during Mr. Korula’s direct examination that required some course correction or new or changed lines of questioning. We’ll never know.

As the morning became midday, it became evident Google is not really interested in going toe-to-toe with the government in their case-in-chief. The content, the arguments, the fire—it’s just not there, and it doesn’t seem like they have any particular sense of urgency. Google must think it can win on its proposed findings of fact and conclusions of law, which were filed pre-trial, or on appeal. In other words, Google wants a narrow legal battle on the law, that takes place in filings, and is staying away from the facts and troublesome evidence as much as possible.

Mr. Korula was asked a vague question about “the role innovation plays” in Google’s adtech, and received an equally vague answer. Mr. Korula later talked about Google using AI to identify ads that are more likely to be successful to benefit publishers and advertisers, and other innovations.

Two observations here: rival ad exchanges and publishers are conspicuously absent from Google’s case. From the witnesses we’ve heard from, Google’s lawyers are focused on advertisers (not ad agencies, or ad exchanges) and not publishers. Publishers, particularly the ones the government presented, offered righteous anger at Google. Competing ad exchanges seem to simply not be a concern of Google’s—in business, or in this case. After all, advertisers are their prized customers because they provide the bulk of the revenue. On the other hand, ad exchanges and ad agencies, to an extent, are rivals they’re just not interested in dealing with.

The other thing I jotted down, is that pervasive and at times overwhelming fumes have bellowed from Google’s case-in-chief that smell just like the fumes of the “Apple-CEO-Tim-Cook-to-be-Obama's-State-of-the-Union-guest” early 2010s era. With your indulgence I’d like to offer some brief color here, as we are nearing the end of the trial, and this has become apparent, and in many ways, quite important to understanding what’s going on in this courtroom when Google’s lawyers speak. The way Google talks about itself, or at least the ways its lawyers talk about Google—radiates this Obama-era framing of neoliberal corporatism and technocratic centrism, that still views the monolithic leviathan that is mega-cap Big Tech firms as scrappy garage startups that America’s downtrodden, tired, poor, and huddled masses yearning to breathe free, ought to look up to as wondrous beacons of innovation, levity, and progress.

I know it’s been a very long ten or so years, so it’s very fair to not be able to recall it well—but listening to Google’s lawyers in court all day for the last several weeks, and particularly as Google has taken up its case-in-chief, I really do feel like I’ve been transported back to ~2013. There’s something to this: Recall Google’s top lawyer, Karen Dunn, led the presidential debate preparation for Kamala Harris, and played the same role for President Obama in 2012, and Hillary Clinton in 2016. Jeannie Rhee, of the same law firm as Ms. Dunn, and another top Google lawyer in this case, represented Hillary Clinton in her email server case, and worked in the Obama Justice Department under Eric Holder. As for Mr. Holder, Kamala Harris tapped him to vet candidates for her running mate. As Robert Kuttner wrote over the summer in The American Prospect: “Of all the people available to review and recommend possible candidates for Kamala Harris’s running mate, why did she choose Eric Holder? It’s a reminder that the corporate wing of the Democratic Party, which served the country so badly under Clinton and Obama, may be down somewhat under Joe Biden, but it is never out.”

If you have the ear for it (and one need not have perfect pitch to pick up on some of these notes) there’s a way these sorts of people talk about the issues in this case (or don’t talk about them) and of companies like Google, and issues of corporate power more broadly, that paints this world in unbridled optimism, infinite possibilities, and wonder and awe at what these companies have to offer. Maybe back when the newest iPhone used to amaze people on a yearly basis with its innovation, or when Google first came out with innovative collaboration tools, or its revolutionary search engine, and platforms like YouTube, these framings were appropriate. But it’s 2024, and the US government just spent two weeks laying out everything that is wrong with the reality of what Google has become.

Google’s lawyers, as was absolutely and undeniably clear by the end of the day, are operating as if these issues simply do not exist—as if the government is to blame for putting down the early 2010s tinted glasses and rolling up its sleeves at long last. Legally, this takes the form of Google choosing to pursue a “refusal to deal” case, instead fighting (and choosing to simply not see) a tying case—essentially an end run around a view of the industry it (and its lawyers, and the broader body politic around Big Tech) do not want to embrace.

Returning to the trial—for the rest of Mr. Korula’s testimony, he presented an interesting alternative explanation of the “waterfall” in which it is in fact the case rivals merely “bid” yes or no in the first instance, in other words, whether they want to bid on an impression or not, before getting to the actual dollar amount. This runs counter to the government’s alleged waterfall, which usually was used to depict this system as one in which Last Look allows Google to win bids with a penny more than whoever the highest bidder was. Mr. Korula distinguished that it is in fact advertisers, not Google, who are doing the bidding. Distinguishing and propping up the role of advertisers has been a key theme in Google’s case.

On cross examination, the government came in heavy. In fact Judge Brinkema asked the government lawyer to lower his voice, whereas witnesses and sometimes lawyers have only ever been asked to raise their voices—a pretty routine occurrence in this trial.

Cross examination focused on discrepancies in testimony regarding Google’s product pages, and focused heavily on an internal “SWOT” (Strengths, Weaknesses, Opportunities, Threats) analysis done by Google that identified AppNexus as the only considerable rival that the company saw—along with a “weakness” that Google feared publishers would perceive Google as acting “unilaterally” in its own interests, leading to a lack of trust.

The government highlighted a number of internal chat messages from Mr. Korula, including one eyebrow-raising message where he asked another Google employee “If the only change for AdX direct was retagging, but ‘AdX direct’ still existed as a concept for antitrust, would you still be concerned?” The central question here obviously being why Google engineers are discussing antitrust in the first place.

The rest of the day involved testimony from Sarah Stefaniu, an Account Manager at Google, Kendall Oliphant, Chief, Contract Program Office of the Communications Division at the US Census Bureau, Marco Hardie, Industry Director (Sales) for Government and Advocacy at Google, and a deposition read-in of Ben John, VP of Engineering at Microsoft.

To be frank, this afternoon of testimony did not add very much to this trial. Ms. Stefaniu continued a pattern of testimony from Mr. Korula in the morning involving extended walkthrough demonstrations of how advertisers can use Google’s products, screen by screen. Ms. Oliphant was given an extraordinarily long direct examination about the COVID-era digital ad campaign efforts inside the US Census (an advertiser) remarkable seemingly only in that an internal review at the Census Bureau of the 2020 Census ad campaign showed digital ads to have been less successful than anticipated, and that Google’s broad suite of ad tools was useful.

Following this testimony, Marco Hardie took to the stand, and at long last, Judge Brinkema stunningly called time on Google’s march to nowhere, saying what I and many others had felt all day—this testimony is adding very little to the key legal issues at hand. Google took the witness down from the stand, and the day ended more or less on this note.

So what is Google trying to do with this approach?

I figured it would be worth looking at the central switch of sorts on which this case will turn: is this a tying case, or a refusal to deal case?

My next article, with an eye towards the finish line, will get into how Judge Brinkema might decide on which analysis to use. But for now, what are each of these about?

Companies are generally under no antitrust obligation to sell or license their products to, or provide their assets for use by, another company. (Competition And Monopoly: Single-Firm Conduct Under Section 2 Of The Sherman Act : Chapter 7, justice.gov). Refusal to deal is one of the most unsettled area of antitrust law, and there is hesitancy to applying it, although it certainly has been applied in the past. In one case from the 1950s, the only newspaper in a town refused to carry advertisements from companies that were also running ads on a local radio station. The Supreme Court found that the newspaper's refusal to deal with businesses using the radio station strengthened its dominant position in the local advertising market and threatened to eliminate the radio station as a competitor. Other cases like Otter Tail Power Co. v. United States and Aspen Skiing Co. v. Aspen Highlands Skiing Corp add further color to refusal to deal analysis. In Otter Tail, the Supreme Court found that a power company refusing to sell electricity at wholesale to towns looking to replace the company with their own service was done “solely to prevent municipal power systems from eroding its monopolistic position.” In Aspen Skiiing, the Supreme Court found that a firm operating three of four mountain ski areas in Aspen, Colorado, violated section 2 of the Sherman Act by refusing to continue cooperating with the firm that owned the fourth ski area in offering a combined four-area ski pass.

In Eastman Kodak Co. v. Image Technical Services, Inc., Kodak and independent service operators (ISOs) serviced Kodak copying equipment. ISOs sued after Kodak began limiting their ability to obtain replacement parts. The Supreme Court observed that although "[i]t is true that as a general matter a firm can refuse to deal with its competitors," that right "is not absolute; it exists only if there are legitimate competitive reasons for the refusal."

As you can see, the bar in refusal to deal cases is pretty high. It involves not only a company refusing to deal with a rival, but doing so with little in the way of a cognizable business justification. That is to say, there is a threshold need to prove that you were refusing to deal in order to hurt a rival.

Alternatively, in tying analysis, tying occurs when a firm forces customers to buy one product in order to purchase another. The products are bundled or “tied” together. This practice restricts consumer choice and limits competition. In a fair marketplace, businesses compete on price and on how good their products are. If an illegal tying arrangement is in place, a seller can use its strong market power on a popular product to force customers to buy a second, lesser product. (The Antitrust Laws, doj.gov)

Tying analysis sounds a lot like the way Google has bundled its ad exchange, AdX, and publisher ad server, DFP. It’s a big part of the government’s proposed theory of the case, and it’s a tough hill to climb for Google to rebut it. You can see why Google would prefer the government have to fight a different, harder battle, and make this about refusal to deal.

Is this case a “refusal to deal” case, as Google prefers? Or a “tying” case, as proposed by the government? We can certainly pick that apart further at a later date. For now, your guess is as good as mine, and Judge Brinkema’s guess is the one that matters.

Well we’re in the home stretch. I’ll be back with what might be the penultimate day of this trial!