Publication date: 01 Sep 2020
At the beginning of 2020, Google announced its decision to eliminate third-party cookies from Chrome browser by 2022. Google is planning to phase out third-party cookies in two years and offers to use HTTP State Tokens and API instead. The project is called Privacy Sandbox. Why should we care about it?
- Chrome browser is one of the most popular browsers in the world chosen by 65% of internet users.
What are third-party cookies, and what is bad about them?
Cookie files are sent from a website to a browser that allows storing information about a user, like this user’s preferred language. It is the main tool for setting online ads.
Third-party cookies are data aggregated by websites and companies which don’t have any direct relationship with consumers. They often buy first-party data from owners, aggregate large amounts of data and sell them to all interested advertisers.
Third-party cookies elimination is intended to motivate publishers, agencies, and other participants to help Google create a new set of open web-standards which are oriented towards protection of user privacy. This measure is also intended to give users control over their personal data online.
How will Google be phasing out third-party cookies?
After third-party cookies elimination, Google wants ad targeting, measurement and fraud prevention to happen according to the standards set by its Privacy Sandbox. Cookies are to be replaced by five APIs. Advertisers will use each API to receive data about issues like:
- fraud prevention,
- aggregated reports,
Currently (August 2020), Google is still working on its Privacy Sandbox project. Google has proposed many features, but they are still far from affecting all processes and aren’t ready yet. No actual platform or code exists for marketers to properly test and assess.
Analysts from hybrid.ai have submitted several ideas on how it all might work and how it might affect advertisers.
How will targeting be implemented?
Targeting based on interests of users and their socio-demographic profile will no longer be available. Or, other IDs will be used to send information. In the latter case, it will take little time to configure the system so that it could accept and interpret data. Otherwise, targeting will rely on functionally that will be available.
For example, Google offers a special API for interest-based targeting. The browser uses machine learning algorithms to organize users with similar interests into cohorts based on their browsing activity. The information is made available to websites via Client Hints (will replace User Agent strings). It means a shift from single-user targeting to targeting on a cohort level.
What will fraud prevention look like?
According to data from Google, users will be required to fill out a CAPTCHA only once, algorithms will further determine whether this user is a real-life human. It will allow reducing the amount of fraudulent traffic. There are, however, some concerns left regarding how Google will be interpreting “human-like” behaviour.
How will impressions be calculated?
If a supply-side platform (SSP) can send IDs, no difficulties will arise regarding calculation of impressions, as no cookies are needed. How the calculation will be done depends on the SSP alone.
Main conclusion: Advertisers are unlikely to be affected by these changes. Some tools will be replaced with other tools which will probably be easier to use and less costly.