Day 2: Everyone's Got Opinions on a Spin

This is the daily coverage of the remedies trial in US v. Google. I’ll be in Virginia writing every day on the courtroom activities.

Day 2: Everyone’s got opinions

First, I bring you the Whack-a-Mole diagram I mentioned yesterday. Isn’t it beautiful!

Judge Brinkema has the quote of the day:

Today saw a parade of witnesses from companies directly or tangentially effected by Google’s monopoly. They were asked their opinions of the various proposed remedies. It was tiresome, to be totally frank. If I wanted to hear the opinions of some guy who works at a holdco I would have stayed in product management. The witness list:

• Witness #3: James Avery, Kevel

• Witness #4: Luke Lambert, Omnicom

• Witness #5: Jay Friedman, Goodway Group

• Witness #6: Jed Dederick, The Trade Desk

I’m not going to go through each witness’ testimony but instead will summarize. After two days we can conclude two contradictory things:

1. Google’s proposed remedies are totally bogus and don’t do much of anything to restore competition;

2. DOJ’s proposed remedies are highly problematic to implement and have serious, if not critical, flaws.

The most interesting thing that’s come out of the testimony thus far is disagreement about what to spin-out and how to do it. Basically, everyone disagrees on this point, and there’s no satisfying answer. I’ll take a moment to summarize the options based on the testimony and reading between the lines.

Option 0: Don’t Spin Out

Supporter: Google

The is the null hypothesis and Google’s remedy proposal. The problem with this proposal, which has been pointed out by every single DOJ witness, is that it makes it almost inevitable that Google could cheat in the auction moving forward. Just because they agree to bid into prebid and establish a monitor does not prevent them from using their position on both the buy- and sell-sides for their own advantage. Call this the “Judge Mehta” option in the case.

Option 1: AdX Spin Out Separately

Supporters: DOJ, Luke Lambert

The DOJ’s proposal is to spin AdX immediately, then wait and see about DFP. Reading between the lines, they think a DFP spin-out is technically and operationally daunting, so they want to kick that can down the road and not force the judge to go out on a limb.

The advantage to this spin is that almost anyone could be a buyer for AdX without a major conflict of interest — unless they own a publisher ad server. Meta, for example, would be an ideal buyer for AdX. They would being FAN demand into the publisher ecosystem, thus making publishers more money. And separated from the ad server, publishers could treat the FAN/AdX demand at an arm’s length just like Amazon’s TAM or any other exchange. Other buyers could include competitive exchanges, DSPs, PE funds, or many others.

The problem with this solution was hinted at by James Avery today when he was asked why Kevel was building its own exchange. Because, he said, “every ad server needs its own exchange.” Which then begs the question, for the astute reader, of what happens to the abandoned DFP that sits within Google without any monetization engine? Google’s lawyers like to remind the court that the majority of DFP customers get it for free — who takes care of those orphans if you spin AdX separately? This is a critical flaw with Option 1.

Option 2: Spin AdX and DFP Together

Supporter: James Avery 

James Avery made the case very concretely while on the stand — he thinks DFP and AdX need to stay together and get spun together. He said specifically, that “AdX separately is a lot less valuable” and would be “not viable.” This actually echos a piece I wrote a couple of weeks ago that asks “What’s AdX Really Worth?” I conclude, not too much.

Wouldn’t spinning the exchange and ad server together just reinforce the monopoly, even without the Google Ads demand? Yes! It would mean that the already quite difficult task of switching ad servers would now be even harder, because you would have to replace whatever unique demand your ad server bundled with demand from a new ad server. And since the market currently only supports one single scaled ad server, hoping that post-spin out there would be a robust market of options is optimistic to the point of delusional. In sum, Option 2 doesn’t really solve anything at all.

Option 3: Consolidate and/or Shut AdX

Supporters: Andrew Casale, Jay Friedman

It’s not entirely fair to put consolidate and shut into the same option, but they have the same rationale. If you consider the market for exchanges in display to be fully fungible because of header bidding, then their relative market share is mostly persistent because of inertia. Whether there are five major exchanges, or four, or three, is in the long term irrelevant to both buyers and sellers.

Casale’s plan, which I facetiously and somewhat unfairly described yesterday, is to have Index acquire AdX, then take the market share and consolidate. Jay Friedman, representing the cynical buy-side, just wants fewer exchanges, and would see AdX just shut down rather than spin.

In addition to being highly unlikely, this option also suffers from the same problem of a neglected DFP being unable to care for its publisher orphans.

Option 4: Profit?

There really is no option 4. Judge Brinkema has her work ahead of her.

The Trade Desk Names Partners with Hidden Fees

Towards the end of the day Jed Dederick, the CRO of The Trade Desk, was being questioned about various random stuff. When asked about the trial and potential outcomes he gave the very nice quote “There’s a sense of hope about it.”

What showed a bit less optimism was his deposition, parts of which were read back by Google attorneys during their cross-examination. In that deposition Jed had specifically, and by name, said he thought SSP partners had “non-disclosed markups,” including Pubmatic, Magnite, and Index Exchange. That’s some tea!

What’s next

Based on Judge Brinkema’s admonishments, the DOJ is speeding ahead with testimony from the expert witnesses. We are going to hear about the potential economic effects of the spin-out scenarios and their technical feasibility. The DOJ thinks they’ll wrap their case Thursday, with Google starting on Friday.

Honorable mentions

• Rembrand got a shout out

• Bunny Studio got a shout out

• Tentatively, closing arguments scheduled for November 17th

• GAM does 8 million QPS