US et al v Google: Day 1

The Monopoly Report is going daily during the trial. Our intrepid founder is in a windowless room without a phone or laptop to get you the good stuff.

US et al v Google: Day 1

This trial is the ad tech version of the Seinfeld finale where all the characters wronged by Jerry and friends come into court one by one and lay out the previous decade of (alleged) malfeasance. Day 1 was a marathon, though a quite efficient one, with opening arguments plus four complete witnesses.

Here’s what happened.

The judge

Judge Leonie Brinkema seems intent on a swift trial. Since a jury trial was eliminated as a possibility earlier, the whole case rests on her shoulders, and she has no problem moving along testimony that’s duplicative, or shortening cross examinations that are overly didactic. 80 years old and she’s learning about header bidding — what a country!

Stuff we learned

This will be a regular feature of our trial coverage:

• Tim Rowe of Gannett explained that the “M” in “CPM” is from the Roman numeral “M” for one-thousand. This is not etymologically correct as the M comes from the French “mille”, which means one-thousand. But the roman thing is so much easier to remember!

• Andrew Casale divulged that his win rate is 0.5% of auctions.

• Asked who is the largest DSP, Casale said “depends on the day, but it is a coin flip between DV360 and TTD”

The government’s case

Opening arguments were thirty minutes per side and included some light Powerpoint. Based on these and the witness testimony, the strategies of the two sides became quite clear.

The plaintiffs’ case tracks their original complaint but with a bit of refinement and simplification. Instead of a litany of code words and secret projects, the case boils down to four allegations:

1. Monopoly on publisher ad serving

2. Monopoly on the ad exchanges

3. Monopoly in the advertising ad network market (e.g. AdWords)

4. Illegal tying of the ad server and the ad exchange

They supported these arguments with both juicy quotes and eye-popping market share stats. On the quotes, these exhibits have not yet been published so I did my best short-hand to capture them, so they may not be 100% verbatim correct. You have been warned.

Here’s a good quote:

Here’s another:

Here’s one about getting AdX demand on other exchanges:

And then the market share stats were fun to see. See chart below for the plaintiff’s estimated share numbers for Google:

Interestingly, the plaintiffs avoided using some of the terms we’ve been hearing like “Bernanke”, “Bell”, etc, but instead used what I consider to be much less well defined and less technical terms like “first look” and “last look” to describe some of the alleged abuses. It will be interesting to see how this develops.

Google doesn’t deny anything, focuses on legal issues

Google pre-released a blog post on Sunday with their legal argument, so not too many surprises on what the said. But it is a bit surprising if you zoom out a bit and see that they are really not spending any times debating the allegations against them. Instead they are 100% focused on legal arguments which try to neuter the case.

Google has three arguments:

1. The market for “open web display advertising” is an artificial definition and does not exist.

2. Google cannot be compelled to deal with competitors so its behavior is legal.

3. Google’s actions had legitimate business goals.

If you were taking a drink every time you heard the phrase “open web display advertising” today, well you would have been kicked out of the courthouse since beverages of any kind were not allowed. But you get my point with this hackneyed cliche, right?

In the cross-examinations of the witnesses (see below) almost all of Google’s points were focused on market definition. They made argument after argument that display is not so different from social, or that display and native and video all go together, or that the same ad could appear on the Washington Post’s website and mobile app, so they were clearly the same thing.

Google cited Ohio v American Express (link) asserting that a two-sided market should only be treated as a single “market” for evaluation based on a single transaction. Thus, they are trying to pull apart the plaintiff’s argument that DFP, AdX, and Google Ads are different markets for this evaluation.

Google also contested the market share estimated shown above, claiming their market share of “display” is only 25% and shrinking. They didn’t show their methodology, but an easy guess would be they include social and native in the denominator.

In this author’s opinion, Google’s opening mixed in very legitimate points with some pretty epic bullshit.

Here’s an example of bullshit (paraphrase):

Here’s an example of something smart. In 2008 an internal Google doc listed their goals as threefold:

• Increase ROI for advertisers

• Increase CPM for publishers

• Show more relevant ads to end users

This is mom-and-apple-pie stuff here. I’m getting choked up. Can I get some more bullshit? OK, then Google showed a slide claiming they are great at interoperating with other parties and including the example of “AdX Direct Tags”, which is a joke, right? That product barely exists and it mostly useless. They also said this (paraphrase):

Then they made me laugh, with a final thought:

Plaintiff’s witness 1: Tim Wolfe from Gannett

What he testified to:

• Gannett makes $330 million from programmatic and pays Google about $15 million for ad tech. Switching ad servers isn’t worth losing revenue.

• Roughly 25-40% is direct sold, rest is programmatic.

• Implementing header bidding increased revenue 15-20%.

What the cross was about:

• Showed Tim a copy of the ads.txt file, everyone in the court a little confused.

• Tried to impinge Tim since Gannett has a private antitrust suit against Google in NY.

Plaintiff’s witness 2: Andrew Casale from Index

What he testified to:

• 83% of impressions that make their way to Index originate at DFP.

• Cannot offer PG deals because he’s not in the ad server.

• While he has huge scale (250B auctions/day) he has a fraction of the data of Google and thus cannot compete as well on pricing and QPS filtering.

• He experimented with low (or zero) take rates and saw marginal increases in win rates. They didn’t really turn this into any kind of an argument against Google.

• Discussion of waterfall, first look, and last look. All were pretty poorly defined and illustrated terms IMHO. But the judge got visibly animated when it became clear to her that Google was allowed to bid after seeing everyone else’s bids (this may have been the most important moment of the day).

What the cross was about:

• Ironically, the cross started by admonishing Casale for being confused about which Google products were which. He was confused that Google Ads was actually Google Adwords. Clearly, someone should have read my Field Guide.

• They read out various points Casale made in his deposition, for example that the phrase “open web display ads” was not a common term.

• Also notably, they got Casale to say that Facebook and Amazon are competitors to him, as part of the “macro market.” This will come back, I’m sure.

Finally, a diagram was offered called Plaintiff’s Demonstrative G that was a slide showing a bunch of exchanges in a waterfall with AdX winning. This diagram was confusingly labeled “First Look” and was conflating that phrase with “Dynamic Allocation”, which is a real thing. I saw several witnesses get confused by this diagram, and Casale ended up on the wrong side of a Perry Mason moment where he basically admitted that he didn’t know what it meant.

Plaintiff’s witness 3: Joshua Lowcock, formerly of UM

Look up “ad agency guy” in the dictionary and Josh pops out. He even brought an English (actually Australian) accent to make his points. I think the judge may have found it especially charming.

What he testified to:

• Direct and programmatic are not substitutes.

• Native, in-app, video are not substitutes for display.

• He saw no increase in latency with the advent of header bidding. He justified this by citing third-party verifier data.

• Same as above, with regard to bot traffic in header bidding.

What the cross was about:

• Isn’t Facebook just a website filled with rectangular ads? Sounds like display to me?

• Isn’t social top of the funnel the same way display is?

• At UM, didn’t you and your team recommend substituting social for display and vice-versa? Why can’t you remember this powerpoint you weren’t involved in writing from five years ago?

Plaintiff’s witness 3: James Avery, Kevel

James charmed the pants off the judge. He was like the Atlas Shrugged character who build ad servers to save America. Very compelling testimony.

What he testified to:

• When he first launched Kevel he couldn’t sign any customers because he didn’t have Google Ads demand through AdX.

• He lost customers like Reddit to GAM because of that demand.

• Pivoted to Native, since those publishers are mostly direct, not programmatic.

• “Open Web Display Ads” —> Yes, that’s a thing.

What the cross was about:

• Pulled up an email thread with Sarah Sluis from AdExchanger about getting AdX demand and asked why he can’t get it using an AdX Tag. James answered because that’s really waterfall.

• Pointed out to James that his official blog has statements that imply it is easy for mainstream publishers to create their own walled gardens and have their own ad server. He admits it is marketing fluff.

• Outlines the “duty to deal” argument, saying all James wants is to get access to AdX demand (which is true).

What’s next?

Tomorrow should be a day packed with testimony and witnesses. Also the exhibits used in today’s opening arguments will be release sometime in the morning.