Google is no stranger to antitrust claims and lawsuits.

But a recently updated lawsuit, brought by the Attorney Generals of 15 US states and Puerto Rico, has brought up some very interesting details about some of the inner workings of Google’s major revenue – ad sales.

Note: In addition to this lawsuit, Google is also facing a federal antitrust lawsuit, a demand for increased regulation in Turkey, a lawsuit brought by 38 US States about antitrust behavior, and a lawsuit about Android and the Google Play Store’s payouts to developers.

Although I knew of this lawsuit, I didn’t become aware of some of these juicier details until I read this thread from Twitter by user @fasterthanlime.

Here are a few that are particularly interesting:

Header Bidding

A technique called “header bidding” allowed website owners to offer ad space to a variety of ad exchanges – this way they could “rent” space to whoever would pay the most.

This would mean website owners could offer the same ad space to Google’s AdSense and other advertising platforms. They would only show the ad that paid the site owner the highest amount. Google’s own internal communications called header bidding “an existential threat”.

Google responded to this threat through a series of anticompetitive tactics. First, Google…allow[ed] publishers using its ad server to route their inventory to more than one exchange at a time. However, Google secretly made its own exchange win, even when another exchange submitted a higher bid. Google’s codename for this program was Jedi—a reference to Star Wars…Google deliberately designed Jedi to avoid competition, and…harmed publishers. In Google’s words, the Jedi program “generates suboptimal yields for publishers and serious risks of negative media coverage if exposed externally.”

Pg 5 – emphasis mine

Google’s Walled Garden

Google has projects devoted to creating a closed ecosystem out of the broader public (currently open) internet.

In this system, Google’s goal is to limit publishers’ ability to “know” their users. This would make Google the sole arbiter of that information. Of course, this would give even more power to Google and their ad exchange.

One of these was called Project NERA, which was designed to “successfully mimic a walled garden across the open web [so] we can protect our margins.”

A core component of this was…

Chrome Tracking

I’ve written about alternatives to Chrome quite a bit. I also wrote a blog entry in October of 2018 when Chrome started to log users into the browser when they logged into a Google service.

Turns out Google is using Chrome logins to track users on non-Google sites. Of course, they are using that information to advertise more effectively and deny that audience information to 3rd-party websites.

(This also explains Google’s interest in no-cookie ad tracking.)

To get publishers to give Google exclusive access over their ad inventory, Google set publishers up for a lose/lose scenario. First, Google started to leverage its ownership of …Chrome, to track and target publishers’ audiences in order to sell Google’s advertising inventory…Then, Google began to steer users into [logging into Chrome] by using deceptive and coercive tactics. For example, Google started to automatically log users into Chrome if they logged into any Google service (e.g., Gmail or YouTube)…If a user tried to log out of Chrome in response, Google punished them by kicking them out of a Google product they were in the process of using…On top of this, through another deceptive pattern, Google got these users to give the Chrome browser permission to track them across the open web and on independent publisher sites like The Dallas Morning News. These users also had to give Google permission to use this new Chrome tracking data to sell Google’s own ad space, permitting Google to use Chrome to circumvent reliance on cookie-tracking technology. The effect of this practice is to rob publishers of the exclusive use of their audience data (e.g., data on what users read on The Dallas Morning News), thereby depreciating the value of publishers’ ad space and benefitting ad sales on Google’s properties (e.g., YouTube).

Pg 95 – emphasis mine

Google and Facebook Collusion

In 2018 Facebook and Google reached a deal to stop directly competing against each other in ads. This agreement was called “Jedi Blue”.

As a result of their bidding agreement, Facebook significantly curtailed its header bidding initiatives and would instead bid through Google’s ad server. In return, Google agreed to give Facebook a leg up in its auctions. In an internal Google memo titled “FAN deal discussion,” Google memorialized that “FAN requires special deal terms, but it is worth it to cement our value.” The parties agreed upfront on when and how often Facebook would bid in auctions, and when and how often Facebook would ultimately win.

Pg 77

Those advantages include:

No exchange fees

But with the Jedi Blue agreement, Google made Facebook a large-scale concession and let FAN circumvent exchanges and bid directly into Google’s ad server. Instead of paying exchange fees, Google charged Facebook a lower 5 to 10 percent fee and prohibited Facebook from speaking publicly about its special lower pricing terms.

A speed advantage over other advertisers

Google subjects other marketplaces competing for publishers’ inventory in Open Bidding to 160-millisecond timeouts. Competitors have actively complained that 160ms is not enough time to recognize users in auctions and return bids before they are excluded. By comparison, Google nearly doubled timeouts, extending them to 300 milliseconds.

Giving Facebook access to contractual relationships with publishers

Google further induced Facebook to help Google “kill HB” by letting Facebook have direct billing and contractual relationships with publishers. This term was advantageous to Facebook because Google prohibits other exchanges and networks in Open Bidding from having such direct relationships.

Giving Facebook access to information on potential spam

Google further induced Facebook to help it shut down competition from header bidding by informing Facebook which impressions are likely targeted to spam (e.g., impressions targeted to bots, rather than humans). Facebook does not have to pay for those impressions. Other networks have asked Google for the same information, but Google has refused.

Helping Facebook identify users

Indeed, since signing the agreement, Google and Facebook have been working closely in an ongoing manner to help Facebook recognize users in auctions and bid and win more often…Google and Facebook have integrated their software development kits (SDKs) so that Google can pass Facebook data for user ID cookie matching. They also coordinated with each other…through the adoption of Unified Pricing rules…The companies also have been working together to improve Facebook’s ability to recognize users using browsers with blocked cookies, on Apple devices, and on Apple’s Safari browser[.]

And there’s more, like guaranteed “Win Rates” for Facebook ads.

This deal has been ongoing since 2018. Google claims publicly that “All participants in the unified auction, including Authorized Buyers and third-party yield partners, compete equally for each impression on a net basis”.

But Wait, There’s More!

These are the only major new points. If you want more information check out this Twitter thread, or the legal documents themselves.