Google is auditioning candidates to succeed the third-party cookie

The race to replace the third-part cookie is set to start in earnest.

Google is ready to test some of its recently announced “Privacy Sandbox” proposals with other exchanges and demand-side platforms to see how its plans to replace third-party cookies with less data-invasive solutions will actually work within advertising auctions.

Google’s announced its intention to run these “bid request signal experiments” late April in a post on GitHub, a software development collaboration platform. Google then commenced contact with external ad tech companies about taking part in the test. (We couldn’t find out the precise timelines of the tests, but please do let us know if you can help.) 

“This is an early-stage concept and we don’t have more details to share right now,” said a Google spokesperson. “We plan to publish updates and progress in Github as part of the process.”

The move marks yet another step in Google’s two-year countdown towards its intention to end support for third-party cookies in Chrome. Google’s Chrome team said when making that announcement in January that the intention was to encourage advertising companies, publishers and other browser providers to create a new set of privacy-focused open web standards.

Google launched the Privacy Sandbox in August to put forward ideas about how behavioral advertising and ad measurement on the web would work in a future without cookies and has invited industry comment via GitHub or through the World Wide Web Consortium’s Improving Web Advertising Business Group. The group had 163 participants at the time of writing, from companies including Google, Apple, Facebook, Axel Springer, The Washington Post, Criteo and The Trade Desk.

In the GitHub post, a member of Google’s “RTB group” said it is beginning to explore how user targeting could work without cookies in real-time bidding auctions by using a technique called Federated Learning of Cohorts, which uses machine learning algorithms that run on the device to group people into audience segments based on behavior such as browser history. The idea is to improve privacy by letting advertiser target groups (or “FLoCs) of users based on common interests, rather than using a pseudonymous identifier for each individual user. 

“Exchanges could offer a uniform RTB interface for aggregated audience targeting to bidders regardless of the data source,” said Google in the GitHub post.

Google also wants to explore exchanges could adapt its Turtledove proposal for retargeting purposes. 

“Exchanges can explore the separation between contextual and user interest components in the RTB data flow, along with hosting the ad selection logic that needs access to both components (for instance, brand safety checks for a product retargeting ad) in a sandbox,” says Google in the GitHub post.

Google said exchanges could also work to apply a “privacy budget” that applies constraints on how much information about a user can be revealed in a bid request. The idea is that any application programming interface calls that violate the budget will result in an error or be replaced with a more privacy-preserving alternative. 

Finally, Google also wants ad tech companies to experiment with its “Trust Token API,” a cryptographic token that would allow, for example, a publisher website to display they have a high likelihood of real and valid user to buyers — but in a way that wouldn’t increase the ability of third parties to track those users across sites.

“We plan to evaluate the viability of these and other proposals via small-scale real-world experiments conducted by exchanges and bidders,” Google concludes in the GitHub post. “We welcome advertising industry partners to provide input to these RTB experiments that minimize user data sharing and better protect user privacy.”

Tom Kershaw, chief technology officer of Rubicon Project and a member of the World Wide Web Consortium’s Improving Web Advertising Business Group, said the proposals were welcomed but he emphasized that they are still in the very early stages.

“These are experiments with a capital E,” said Kershaw, who is also chairman of Prebid.org ad tech industry organization, which is contributing to the Improving Web Advertising Business Group as a collective. “We definitely support these initiatives but there is not a single proposal on the table close to being adoptable right now. A ton of work needs to be done.”

Wil Schobeiri, chief product officer and chief technology officer at MediaMath, another member of the Improving Web Advertising Business Group, also reiterated the proposals are a long way from evolving into technical standards.

But, he added, “It’s great, given the implications of what’s at stake, that the advertising arm of Google is attempting to implement what the browser arm is doing. It’s a really important step in the process.”