Defining user profiles is the basis of the advertising business: companies in the sector use all kinds of methods to collect as much information as possible in order to refine as much as possible the personalization of their campaigns.
However, at the same time, user preference is gradually going in the opposite way, that of preserving privacy… which forces companies to innovate and develop new technologies.
In this context, recent news can be understood as Google’s intention to stop support for third party cookies in its Chrome browser: “The aim is to make the network a more private and secure site for Internet users, while at the same time supporting advertisers”.
User-Agent: A definition
The User-Agent is an element of the HTTP protocol that browsers send to the servers we connect to, showing information about the browser itself and about our operating system (software and version number of both, if the system is 32 or 64 bits, sometimes even which language it is configured in, etc).
Actually, a User-Agent is just a string; let’s see here an example (from a Firefox 68 installed on Windows 10):
Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:68.0) Gecko/20100101 Firefox/68.0
This feature of the User-Agent found its utility since Netscape implemented it in the 90’s, since it allowed websites to adjust their code depending on which browser was being used to access them.
However, it has also generated several disadvantages: many minority browsers have ended up ‘pretending’ to be one of the main browsers to ensure correct visualization (Vivaldi, for example, pretends to be Google Chrome in its User-Agent). In the words of Yoav Weiss, a Google engineer, “User-Agent detection has become an endless source of compatibility problems”.
Not to mention that it is a fundamental element of the ‘canvas fingerprinting‘ technique, which allows us to collect information about our ‘fingerprint’ on the Internet, in order to link our browsing activity to a particular individual. But the latter is not limited to the User-Agent: it also takes into account our particular combination of browser plugins, time zone, installed fonts, monitor resolution, etc.
There is a lot of information that we leave behind when we surf the Internet and if you want to see the detail, with EFF’s Panopticlick tool, you can test your browser and see how much information it offers (without you knowing) to others.
What does Google intend to do now?
Now, Google has decided to put on the table a change in the use of this feature: a gradual change throughout this year, starting by downgrading and freezing the User-Agent first in order not to reflect the browser updates and thus be able to unify all versions of the Chrome browser.
This would involve adopting a generic text string that doesn’t reveal much information about the user or their computer: just that they are using a Chromium-based browser, and whether they are on a mobile or desktop computer. It is not planned to remove the User-Agent to ensure backwards compatibility with outdated web platforms, but it would be achieved in this way that it would lose much of its usefulness for advertisers.
Google is not looking to simply knock out the User-Agent. What it is really looking for is to start introducing the technology that will eventually replace it. This technology is called “Client Hints” and will allow servers to request specific data about the computer via HTTPS and browsers to answer those requests based on the configuration previously approved by the user (who can refuse to provide any specific data).
Apple (developer of Safari), Mozilla (developer of Firefox) and Microsoft (developer of MS Edge) have expressed support for Google’s proposal, but have not yet given details of when or how they will join this initiative.