Day 5: Nothing happened but here are some cool charts

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

Day 5: Nothing happened but there are charts

Today we heard from two interesting witnesses about the technical feasibility of the DOJ proposed remedies. I guess I’m not being fair by characterizing today’s newsletter as “nothing happened,” but compared to the fireworks yesterday, it is a bit of a dull update. To spice things up a little, I’m going to highlight some charts that I referenced previously but just came online (shout out to Arielle at Checkmyads for all the indexing).

Chart 1: AdWords and DV360 Spend All their Budget on AdX

This chart from the expert report of Professor Robin Lee shows that 89% of AdWords display spend ends up on AdX and 63% of DV360’s does as well. Keep in mind that the DV360 number is just for display, so if you add in the YouTube spend (which we learned is now 3-4x the display spend) you realize that DV360 is essentially a powerful tool for buying Google inventory, and not much else.

Chart 2: AdWords and DV360 Combined Represent 67% of all Display Ads

A second chart prepared by Professor Lee shows that not only do the Google tools send all their demand to AdX, but also they make up 2/3 of the entire market! Important caveat this is open web display, so PMPs and other deal types are possibly (?) excluded.

Chart 3: Overall Google Spend Migrating to YouTube

Google’s expert witness Lerner (who has not yet testified) produced this chart, showing that ex-search, AdWords is increasingly spending all its budget on YouTube. I believe this is the first time this data has been shown. “Indirect Open Web Display” is $600-700 million per month which puts AdX gross revenue from AdWords at $7ish billion per year ($650×12×90% to AdX).

Also noteworthy how in-app had very little growth during a period when AppLovin reached a $200 billion market cap. That’s a miss!

Does anyone know what “Other Google O&O” is mostly? Maps is my guess, which would put it at a cool billion/month.

Chart 4: DV360 Spend by Type

This is the chart I tweeted about earlier this week, showing DV360’s spend by type. “Open” display is flat at ~$4 billion run rate while “Open Prog. Direct” is growing and should cross that number soon, so let’s call display $8 billion total. YouTube is a monster, at over $10 billion. Add it all up and eyeball it to about $20 billion. Note: this is more than in my twitter thread because I didn’t have the chart handy and was going by my notes.

Witness #11: Michael Racic, President of Prebid.org

Now that the charts are out of the way, let’s get to the witnesses. Racic is the head of Prebid and he spent his testimony talking about how Prebid operates and governs itself, and why it is a good future owner of the proposed open sourced DFP “final auction logic,” whatever that is.

There really wasn’t much controversy here and I think the judge would be predisposed to accept that Prebid will be a good home for this, in the event that remedy comes to fruition.

Witness #12: Goranka Bjedov, Expert in Software

This expert witness had long stints at both Google and Facebook and managed many complex software migrations and other projects. She gave opinions about how long some of the proposed remedies would take (time from start of those initiatives, not elapsed time):

• Develop APIs for migrating data: 18 months

• Open sourcing the final auction logic: 24 months

• AdX migration: 18 months

• DFP migration: 24 months

Overall she was very credible and gave meaningful examples of projects she had worked on, such as the Instagram migration to Facebook’s cloud. I had to leave a bit early so I didn’t see the cross-examination, but I assume they are going to try to make her look like she’s never seen a computer in her life.

Who won week one?

The DOJ case is almost wrapped up and will likely finish Monday. Even though several of their witnesses were contradictory, I think they got their key points across. Namely, that Google’s remedies are insufficient, and that their proposed remedies are feasible. I don’t think they convinced the judge that their remedies are necessary, however.