In July 2024, Mozilla introduced a technology called Privacy-preserving attribution with a new version of its Firefox browser, which is designed to track the effectiveness of online advertising. It is enabled by default in Firefox 128.
This quickly caught the attention of online privacy advocates, leading to headlines like “Now Mozilla Is Selling User Data.” The uproar attribution technology was so intense that Firefox CTO Bobby Holley had to explain to Reddit users what Mozilla had actually done and why.
It’s time to take a closer look at what Privacy-preserving attribution technology is, why it’s needed at all, and why Mozilla decided to introduce it now.
Google Ad Topics and Facebook Link History*
Let me start with the background. As you may remember, the developers of the world’s most popular browser, Google Chrome, have been attribution technology hatching plans to completely disable support for third-party cookies since 2019.
This technology has been widely used for tracking users’ actions on the Internet for three decades now. So, on the one hand, it is the basis of the online advertising industry, and on the other, it is the main tool for violating users’ privacy .
As a replacement for third-party cookies
Google introduced its own development called Ad Topics some time ago . Within the framework of this technology, user tracking occurs based on the history of the Chrome browser and the history of the user’s interaction with applications in Android. It was assumed that thanks to the implementation of Ad Topics, support for third-party cookies in the browser would be disabled in the second half of 2024.
Another major player in the digital advertising industry, Meta**, which also relies on third-party cookies, has come up with its own version of user tracking.
What we end up with is that by disabling third-party cookie support
Those players in the online advertising market who have some large, fully controlled digital space gain an advantage: the world’s most attribution technology popular browser and OS in the case of Google, the most popular social network in the case of Meta**. Smaller players end up in a dependent position.
The question arises: is it possible to develop some mechanism that will simultaneously allow advertisers to track the effectiveness of advertising and at hungary whatsapp data the same time does not involve mass collection of user data? The answer to this question is the Privacy-preserving attribution technology.
Prio private aggregation system
To better understand the history of this technology, we need to start a little further back. In 2017, Stanford University cryptographers Henry Corregan-Gibbs and Dan Boneh published a paper . In it, they described a privacy-focused system for collecting aggregated statistics that they called Prio.
To put it very simply, Prio is based on the following mechanism. Let’s say you are interested in the average age of a number of users, but you want to preserve their privacy. You set up two (or more) piggy banks and ask each user to count out the attribution technology number of coins corresponding to their age and, without showing it to anyone, randomly distribute these coins into different piggy banks.
Next, you pour the coins from the piggy banks into one pile, count them and divide by the number of users. As a result, you get the desired value of the average age of users. At the same time, if at least one of the piggy banks works honestly (that is, does not tell anyone what goes into it), then it is impossible to determine how many coins exactly an individual attribution technology user put into the piggy banks.
General operating principle of Prio
The main stages of information processing by the Prio system. Source
On top of this basic mechanism, Prio has a lot of cryptography that protects information from interception and ensures the reliability attribution technology of the data received. It prevents users from slipping answers into the system for one benefit or another that could spoil the overall value. However, the main idea is to use two or more aggregators that collect random shares of the necessary information.
Prio’s algorithms have another important feature: they provide much higher system performance compared to previous approaches to reliable anonymized data collection – 50-100 times, according to the authors of the work.
DAP — Distributed Aggregation Protocol
Mozilla became interested in Prio back in 2018. The first result of this interest was the development of the experimental Firefox Origin Telemetry system based on Prio. Notably, this system was designed to privately collect telemetry on the effectiveness of the browser’s fight against advertising trackers.
Then, in February 2022, Mozilla introduced the attribution technology IPA (Interoperable private attribution) technology developed jointly with Meta**, which, apparently, eventually became the prototype of PPA, Privacy-preserving attribution.
In May 2022, a zero draft of the Prio-based DAP, Distributed Aggregation Protocol, was published . The authors of this draft were representatives of the non-profit organizations Mozilla and the Internet Security Research Group (ISRG) , best known for the Let’s Encrypt project that democratized the use of HTTPS, as well as two Cloudflare employees.
In parallel with the work on the protocol
General principle of operation of Distributed Aggregation Protocol
Schematic explanation of the basic working principle of the DAP protocol. Source
Finally, in October 2023, Divvi Up and Mozilla announced that they were working together to implement the DAP protocol in the Firefox browser. As part of this work, a system of two aggregators was created, one of which operates on the Mozilla side, and the other on the Divvi Up side.
How Privacy-preserving attribution (PPA) works
It is this latter system, jointly created by Divvi Up and Mozilla, that is ultimately used by Privacy-preserving attribution technology at the moment. This is just an experiment involving a limited number of sites.
The mechanism of its operation is generally as follows:
At the same time, neither the site nor the aggregation system has information about who these users were. What else they did on the Internet, and so on.
Why is Privacy-preserving attribution (PPA) technology needed?
In the Reddit thread mentioned at the beginning of this post. Firefox’s CTO explained what they were trying to achieve by introducing Privacy-preserving attribution with the new version of their browser.
Mozilla’s point of view is roughly as follows. Online advertising, at least at this stage of the Internet’s development, is a necessary evil. And in general, it is understandable that advertisers want to be able to measure its effectiveness.
In fact, Privacy-preserving attribution technology is an experimental example of such a tool that allows advertisers to receive the feedback. They need without collecting and storing information about everything that happened to users.
If the experiment shows that this technology can satisfy the needs of advertisers. Then privacy advocates will have a serious argument that they can use in future conversations with regulators and lawmakers. This will prove that total surveillance on the Internet can indeed be avoided. Therefore should be limited by law.
Disable third-party cookies now
Coincidentally, almost immediately after the hype around Mozilla’s new technology, Google announced that it was abandoning its plans to disable third-party cookies entirely. It looks like these nasty cookies are here to stay — a bit like Microsoft’s struggle to bury its Internet Explorer .
The good news is that, unlike Internet attribution technology Explorer. Which is really hard to root out of Windows, every user can deal with third-party cookies on their own. All modern browsers allow you to easily disable them – we have detailed instructions on how to do this .
It’s worth keeping in mind that opting out of “cookie opt-out” doesn’t mean the end of Google Ad Topics. The company plans to continue this experiment. So I recommend disabling this thing too: here’s a detailed description of how to do it in Chrome and Android.
And if you use the Facebook mobile app, you should also disable “Link History” using one of our other instructions .
Also, to block advertising trackers you can and should use the “Protection from data collection” function in our Kaspersky Standard. Kaspersky Plus and Kaspersky Premium subscriptions .
Finally, we recommend using our free Privacy checker service. Where we have collected instructions on setting up privacy for the most common applications, services. Social networks for different operating systems.
As for Privacy-preserving attribution technology, this technology seems quite useful. If you think otherwise, here is a simple instruction on how to disable it in Firefox. But personally, I prefer to support the development of this technology. So I am not going to disable it in my browser.