In December 2020, ten states filed a complaint in U.S. Distsrict Court in Texas, alleging Google (Alphabet) of joining with Facebook in forming a digital-marketing monopoly, designed to rig ad auctions. This project was code-named Jedi Blue, and designed to limit Facebook’s own competitive moves in exchange for special treatment in auctions run by Google.
“Google is a trillion-dollar monopoly brazenly abusing its monopolistic power, going so far as to induce senior Facebook executives to agree to a contractual scheme that undermines the heart of [the] competitive process,”
Texas Attorney General Ken Paxton
We have previously reported on the murky nature of Google’s own auction system, and the rampant fraud that can easily be identified within it. For advertisers, there is little we can do but to accept the auction bids and traffic coming in. The best way to limit negative effects is to properly set up our campaigns as best as possible to target the users we want, alongside vigilant monitoring of our traffic, performance and costs. It should also be noted that Facebook is not a defendant in this case.
What is Jedi Blue?
Jedi Blue is the code-name given to the price-fixing deal between Facebook and Google’s parent company Alphabet. Challenged by Facebook’s growing role in the pay-per-click digital advertising field, Google attempted to preserve its search engine marketing monopoly. Google’s problem was exacerbated by the growing use of header bidding.
What is Header Bidding?
Header billing is an ad-sales method, whereby website publishers can circumvent established (read Google) ad exchanges for buying and selling ad space online. Using header bidding, publishers are able to directly reach out for bids from multiple exchanges at once. The exchange works by auctioning available space for advertisements while the page is loading.
By 2016, some 70% of major publishers had been using the tool, and Google was worried that a major rival like Facebook would jump in with their Facebook Audience Network service.
“Need to fight off the existential threat posed by header bidding and FAN,”
Google advertising executive Chris LaSala (2017 internal document)
When Facebook endorsed header bidding in March 2017, Google went into action, and was able to reach a deal in September 2018. This deal is what’s known as Jedi Blue.
What are the Terms of Jedi Blue?
Acording to a draft of the lawsuit, the following stipulations of the agreement between Google and Facebook are included:
- Facebook pays Google a 5%-10% transaction fee. (Google says this fee is standard.)
- Google assists Facebook in recognizing mobile and web users.
- Facebook bids to show ads to 90% of recognized users.
- Facebook receives 300 millisecond timeout to recognize a user and place a bid. Tther participants have a 160 millisecond timeout.
- Facebook agrees to be locked into spending $500 million annually starting in year four.
In exchange for this, Google provided Facebook with special treatment in the form of direct access to their ad bid servers. While normal bidders would typically go through an exchange, and the exchange would then determine the bid winner and send it through to Google’s servers.
By avoiding the middleman, Facebook faces less competition in delivering ads and is able to deliver them at a lower cost. Also, while Facebook is offered a 5% to 10% transaction fee, Google charges a 20% fee to typical exchange users. Google circumvents this seemingly unfair advantage, by claiming that the transaction and exchange fees are different, and by barring Facebook from discussing their terms publicly.
What can we do ourselves?
What Effect Does Jedi Blue Have on Me?
Much like attributing value via uplift modeling, assisted conversions or ad recall, putting a value to the effects of this program are murky at best, if not impossible. On the one hand, Google claims there are at least 25 other companies participating in its program, and that:
“There’s nothing exclusive about [Facebook’s] involvement, and they don’t receive data that is not similarly made available to other buyers”
and perhaps most importantly:
“We don’t manipulate the auction”
On the face of it, this appears accurate. Google does not manipulate the auction directly to Facebook’s favor, save for the nearly double timeout to identify bidders and place a bid. However, only ad buyers within Google’s program have access to this privileged information, which is something buyers on exchanges do not.
Furthermore, the lawsuit also alleges that ad placements which were likely produced by bots were not charged to Facebook. While Google does refund some ad spend that they attribute to bots after the fact, there is no detailed accounting of this, other than a single line-item fee return in the Google Ads console. Other participants in the auctions in question who asked for the same information were denied it.
Data is everything when it comes to successful and improving digital marketing ad performance. The more data you have, the better decisions you can make on where to place your marketing spend.
“It’s like saying CarMax has the ability to tell when they have a junky used car, and they don’t tell you”
Adam Heimlich, CEO Chalice Custom Algorithms LLC
What Can I Do?
Very little, it would seem. Setting auction bids and deciding who places where is out of our hands, so as end-users and marketing experts we’re stuck paying the fees that show up in our console. We also don’t have access to the same, deeper analytics that these special Google partners may have.
Google does strive to limit ad fraud and maximize ad traffic quality for its users. The biggest conflict of interest, however, is that Google both runs the ad auctions it participates in. This limits their incentive to attack fraud, and is also addressed in the lawsuit.
Google, Facebook and the other major online players play a huge role in modern advertising, and are relied on by countless businesses and agencies. Their influence and ease of building a monopoly are rightly a concern. From our positions as advertisers, however, we must be vigilant. Experience, above all else, plays a huge role in preventing unncessary spend and delivering ads to our targeted users. Vigilant monitoring helps us prevent any new tricks or schemes that may be negatively affecting our performance. We use the data that we are given as best we can to deliver results, and can only hope for more transparency from our exchanges and services.