Mozilla released Few days ago which is working hand in hand with Facebook to implement IPA technology (Interoperable Private Attribution), which allows advertising networks to receive and process statistics on the effectiveness of advertising campaigns, but always respecting user privacy.
It is mentioned that IPA introduces the concept of a match key, which is generated relative to the user, but is write-only. In connection with this key, initial (source) and final (target) events are generated, the information of which is transmitted in encrypted form to third-party arbiter servers that are not associated with advertising networks and are trusted by the user or the browser.
These servers allow ad networks and advertisers generate a query that show aggregate statistics that link beginning and ending events at different sites.
In order for this to be feasible when processing statistics without disclosing data about specific users, use differential privacy cryptographic mechanisms and confidential computing Multi-Party Computation (MPC), which allow several independent participants to perform calculations without receiving information about other people's data (websites, browsers and advertisers individually do not have access to all the information) and perform operations on data encrypted.
Besides it developers have an API which includes three basic functions which are:
- Assign a match key (navigator.setMatchKey)
- Generate a source event (navigator.generateSourceEvent)
- Generate a final event (navigator.generateTriggerEvent).
Source events are actions such as showing ads and user clicks on ads. Final (target) events characterize actions performed on another site after the user's response to the advertisement, such as purchasing a product, registering, and downloading an application.
Since these events take place in different places, to assess the effectiveness of advertising, ad networks track user activity when visiting different sites, using methods such as identification cookies, which is perceived as a violation of privacy.
The mechanism Proposed IPA makes it possible to link initial and final actions without leaking information about the user: mapping keys can only be assigned and are processed only in encrypted form, so statistics can be tracked, but it is not visible which user it is.
As a result, with the help of IPA, advertisers will be able to evaluate the results of advertising campaigns without violating confidentiality, and site owners will be able to analyze ways to optimize advertising placement.
We've sought to apply privacy-preserving ad technology to the problem of attribution, so advertisers can get answers to important questions without harming privacy.
Attribution is how advertisers know if their ad campaigns are working. Attribution generates metrics that allow advertisers to understand the performance of their ad campaigns. Related measurement techniques also help publishers understand how they are helping advertisers. Although attribution is crucial for advertising, current attribution practices have terrible privacy properties.
To improve privacy protection, it is also possible to add random noise to the aggregation results and to apply the "Privacy Budget" principle, according to which the browser provides information related to sensitive data only up to a certain point, if the limit of API calls is exceeded, the emission of more information is blocked.
The developed API has three main objectives:
- Preservation of confidentiality: limitation of the total amount of information about individual users published during a certain period of time and protection against abuse (attempts to circumvent restrictions through registration of a large number of domains, tracking attempts and additional identification).
- Added measurement support of all the main types of ad conversions, including analysis of statistics on views, clicks, activity after re-entry, comparison of conversions between users who saw and did not see ads, evaluation of performance on different sites, interaction of links with ads of a user in different browsers and on different devices.
- Stay competitive: the availability of statistics for all players in the online advertising market and the absence of obstacles in connecting new participants.
Finally, if you are interested in knowing more about it, you can consult the details in the following link