Mozilla, the Firefox developer and Meta AI (previously Facebook), have had a long-standing rivalry, with the former calling out the latter on the company’s disastrous privacy and security record. However, Mozilla has recently announced that it is working with Meta by adopting Interoperable Private Attribution, or IPA. IPA is a novel approach that intends to enable conversion measurement – or attribution – for advertising.
What is Interoperable Private Attribution?
The main notion of IPA is to replace per-action ad reporting (e.g., the browser sending data to an advertising group when you click on an ad) with aggregated reports for batches of events, as detailed in the proposal draught. To avoid fingerprinting, websites can produce a “match key” linked to your account or device, which is only accessible by the browser. A few functions are also in place to make it impossible for anyone (including companies or marketers gathering data) to identify consumers who interact with advertisements. It’s comparable to Prio, Mozilla’s tool for analysing how people use Firefox, that was released a few years ago.
Following are the reasons why Mozilla and Meta AI are collaborating:
- The Internet relies on advertising for its survival. Mozilla has been trying to figure out how to apply privacy-preserving advertising technology to the attribution problem so that advertisers can get answers to key questions without jeopardising their privacy.
- Advertisers use attribution, which creates indicators, to determine whether their marketing strategies are effective.
- Publishers can also use related measurement approaches to figure out how they are assisting advertisers. Despite the fact that attribution is essential to advertising, existing attribution approaches offer poor privacy features.
According to the Proposal Draft authored by Erik Taubeneck (Meta), Ben Savage (Meta), Martin Thomson (Mozilla), the following are the main components of IPA:-
Sign up for your weekly dose of what's up in emerging technology.
- Setting match keys: The goal is to create an interoperable API that allows each browser / mobile device used by a single person to generate standardised ad impression and ad conversion reports that can only be joined within a secure multi-party computation using the proposed privacy-preserving measurement protocol and trusted helper servers.
- Event generation: Mozilla offers a new API that allows developers to create standardised source events (ad impressions) and trigger events (ad conversions) from apps and websites.
- Event collation model: IPA recommends logging that is identical to what is already feasible, with the exception of the removal of global IDs and the addition of encrypted blobs.
- Using encrypted match keys to join events: While prior approaches for aggregate attribution measurement have recommended connecting source and trigger events on-device, Mozilla and Meta propose doing so within the context of secure multi-party computation.
- Results of measurement: Mozilla and Meta propose that each site/app have a global privacy budget (per match key). Consider the scenario in which a site/app only generates source reports or trigger reports.
Despite the fact that the proposition appears to be sound, the alliance is unexpected. Last month, Mozilla and The Markup started a pilot study to see how Meta/Facebook used tracking pixels across the web to track web activity. The study’s purpose, according to Mozilla, was to “report on where Facebook is following you and what kind of data they are gathering.” In recent years, the Mozilla-Markup group has campaigned against Facebook several times, and less than a year ago, it began running advertisements on Meta’s platforms criticising the company’s creepy ad targeting capabilities. Unlike nearly every other collaboration news announcement published by Mozilla, the partnership with Meta has not been posted on the official Twitter accounts for Firefox or Mozilla. Only a low-key announcement was posted on the Firefox subreddit.