Google’s Privacy Sandbox
Last Updated: August 26, 2021
The announcement that Google will begin migrating the digital marketing model by phasing out third-party cookies in Chrome appears to respond to a trend on the web that points towards a hyper-privatization stage in the development of Internet, best represented by Google’s Privacy Sandbox.
This change will cause marketers to lose the ability to create detailed profiles of individual consumers, as it would erode micro-targeting practices. Because of this, companies will have to pivot their strategies to a customer-centric approach that generates value to users in exchange for their consent.
What will replace third-party cookies in the context of digital advertising? There are several components within the framework of the Privacy Sandbox:
According to the Electronic Frontier Foundation , this API is designed to function as a privacy pass using anonymous tokens that cannot be mixed or matched with other tokens, making it difficult for Google to track users. This API combined with a “privacy budget” is an effort by Google to thwart fingerprinting without hampering the legitimate application’s ability to access useful information.
Google has argued that hindering device fingerprinting would be to the detriment of legitimate apps / websites that actually require useful information obtained via APIs. Therefore, the proposal is to limit the amount of data that websites can collect. For this to work, each website would have an allocated budget, and every time a website exceeds the allocated budget the browser (Chrome) would cut access. A couple of questions that come up at this point that remain unanswered are: How would the budget be allocated to each website? What would be the parameters to consider when assigning a budget? The only thing that is clear is that websites that offer services that require access to powerful APIs (such as video calls) would have to obtain the consent of users to “go over budget.”
This feature should also be used in conjunction with the Trust API so that advertisers can mark their ads with metadata (ie Destination URL, Report URL, additional impression data – the latter possibly means “Unique ID”) to obtain information about the conversion rate of the displayed ads. Under this scenario, a report URL would report on whether an ad was “converted” (that is, whether the user actually purchased the advertised product or service).
This is a mechanism that will prevent HTTP apps from accessing IP addresses or any other potentially identifiable information. To avoid correlation of requests using an IP address, browsers could only allow Turtledove (concept explained below) for ad networks that use IP blinding techniques.
First party assets
This mechanism would allow the owners of different domains to declare all the domains as part of the same first part. For example, Google.com, Google.co.uk, youtube.co.uk belonging to the same legal entity.
Federated Learning of Cohorts (FLoC)
The FLoC would favor an “interest-based” model over the current 1: 1 model. This component is based on Google’s federated learning technology, which allows users to build their own local learning machine without sharing all their data at once. This would allow grouping the browsing history of users instead of following each of them throughout their online journey. In other words, the FLoC would track browsing patterns and generate similar user groups or “flocks”. The “flocks” would then be assigned into groups, where each user would be matched with a unique group identifier.
Private Interest Groups (PIGIN)
This is the component that would track the different interest groups that a user is believed to belong to. The key responsibility of the browser is to find out which set of group memberships can be revealed, while preventing tracking and respecting privacy.
Non-server dependent API to keep track of which people are in which stakeholder groups. With this API, sites could ask the browser to join a specific interest group for a defined period of time. This API responds to the growing concern fueled by the current practice of identifying people when they browse websites. For this API to work, the browser has to retain the information (not the advertiser), and advertisers cannot combine interests with any other set of data (such as “what page a user is visiting”) that can identify the user.
These components remain speculative, and it is not yet clear which components will make it to the pilot. What is clear is that Chrome will begin to phase out third-party cookies at the end of 2022, following in the footsteps of Safari and Firefox, who have already limited third-party cookies in a similar way. The game changer is Chrome’s market share, which is much broader than the advertising market share of any other browser.
Now is the time for advertisers, publishers, and marketers to experiment with strategies to begin maximizing addressability and personalization without relying on third-party cookies.
To learn how CookiePro can help you build customer trust, stay compliant and make the move away from third-party cookies a competitive advantage, request a demo with our team today.