Google’s Privacy Sandbox: Situation Report

For months now, the expected end of third-party cookies has been stirring up the digital marketing ecosystem, as it searches for credible alternatives. Developed in beta mode for several years, the Privacy Sandbox is presented by Google as one of the solutions.

What exactly is the situation? We take a look at this technology.

What is the Privacy Sandbox?

The end of third-party cookies will considerably affect the individual tracking of users as they browse different sites, and consequently the building of audiences.

The Privacy Sandbox is designed to compensate for this by anonymising profiles and creating interest groups. Users will no longer be targeted individually, but as a group, so that they cannot be distinguished. In this way, targeting will remain relevant, while respecting individual privacy.

How does Google’s Privacy Sandbox work?

This is a set of APIs enabling website publishers to ask the browser about the user’s centres of interest or to retrieve information (anonymised and grouped).

Currently compatible with Chrome, these APIs appear to be integrated by Microsoft Edge but rejected – at least in part – by Safari and Firefox on privacy grounds (officially 😀).

These APIs are as follows:

  • Analysis of historical data to deduce users’ centres of interest (API Topics)
    • Topics are categories deduced by the browser based on the pages visited. In Chrome, users can view them, delete those that don’t interest them and deactivate them from the Settings menu.
  • Remarketing (API Protected Audience)
    • Delivery of retargeting ads (for example, you visited a shoe site, and an ad for shoes will be shown to you elsewhere) while limiting third-party tracking between websites.
  • Conversion tracking (Attribution reporting API)
    • Data enabling a link to be established between ad clicks and views (post-click and post-view attributions) or ads views and conversion events (e.g. sales).
  • Shared Storage API
    • Read/write cross-site data in Chrome’s Local Storage. As we understand it, this makes it possible to reproduce cross-site browsing tracking on the written level – which is what the third-party cookie allowed – but the data will be grouped and anonymised on a read level.
  • Private Aggregation API
    • Generate aggregate data reports using Protected Audience data and Shared Storage cross-site data.
  • Fenced Frames API
    • We can see the limits of the French translation. As we understand it, the idea is to be able to embed content in the page in the form of a div or iframe without sharing any data. These techniques are used by many adtech companies and it would probably be useful to offer an alternative that is more respectful of confidentiality.

More information is available directly from the Privacy Sandbox website, including the timetable for the availability of APIs for testing (remember that 1% of third-party cookies are currently refused by Chrome, precisely so that sample tests can be carried out for the market.

In a nutshell: the Privacy Sandbox is a set of browser tools that enable the functions of third-party cookies (audience and conversion tracking) to be reproduced anonymously and on a group of users in order to avoid privacy protection constraints.

However, the CNIL considers that the Privacy Sandbox should be subject to the user’s consent and specific information on this method.

Uncertainties raised by the market

A detailed article by Luciana Uchôa-Lefebvre published in JDN a short while ago echoes the questions being asked by the various players in the market.

Here’s a quick summary:

  • For media buyers, the lack of flows specific to the Privacy Sandbox APIs on the main DSPs is complicating adoption efforts.
  • A regulatory investigation by the Competition and Markets Authority has raised concerns about the potential impact of the Privacy Sandbox on competition. This regulatory uncertainty adds a further layer of complexity to an already turbulent landscape.
  • Publishers are also concerned about the implications of the Privacy Sandbox for web performance. With an expected increase in the number of requests and concerns about page latency, the user experience could suffer…

Other alternatives to third-party cookies

Other alternatives to third-party cookies do exist, but their adoption remains limited. These solutions include :

  • Shared advertising IDs: this is a unique identifier used by different platforms and applications to track advertising activity and online behaviour.
  • Data clean rooms: this is a secure platform that enables advertisers and data providers to collaborate and analyse data without compromising user confidentiality.
  • First party ID: a unique identifier collected directly by a website or application owner from the user’s interactions with its own platform, unlike a third party ID which is collected by third parties.
  • Contextual targeting: an advertising method which consists of displaying ads based on the specific content of a web page or a particular context rather than the demographic or behavioural data of users.

Third-party cookies alternatives

First party data, Customer Data Platform and Attribution can help circumvent the end of third party cookies

The end of third-party cookies has two major effects: building audiences and tracking conversions. As we can see, the alternatives are full of uncertainties and, above all, they do not cover every aspect.

Another, more virtuous and sustainable way of collecting and using data is possible by using several technologies (available on our platforms: Platform X and Adloop):

  • The adoption of first-party tracking, which enables websites to collect consented data directly from users, thereby fostering lasting customer knowledge.
  • Enriching user profiles with CRM or product data to refine targeting
  • The creation of first-party audience segments and their distribution to marketing partners via a Customer Data Platform (CDP)
  • The adoption of campaign performance measurement systems that combine :
    • Attribution/Contribution to assess the real performance of each advertisement
    • Statistical or probabilistic systems to assess the influence of post-printing in the closed ecosystems (walled gardens) that are social networks (techniques linked to Media Mix Modelling – MMM).
    • Full-funnel analysis to measure campaign performance across all campaign objectives

To find out more about these alternatives, please contact us.