On-device Experimentation via Secure Anonymization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing experimentation systems face challenges in ensuring data security while maintaining consistency and reliability in content distribution tests, often violating data security constraints or failing to ensure test group exclusivity across multiple login sessions.
Innovation Solution
Implementing an on-device experimentation component that runs in a secure environment, using a randomization unit to anonymize and attribute user feedback signals to test groups, ensuring that protected identifiers are not shared with other-party systems and maintaining consistent assignment across login sessions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If user identifiers are shared with other-party systems for content distribution testing, then test group assignment consistency can be maintained across multiple login sessions, but data security constraints are violated
Solution Approach 1:
The system segments the identifier handling process into two distinct parts: a secure local component that generates and protects the original user identifier, and an external experimentation system that receives only anonymized identifiers. This segmentation allows the original identifier to remain protected on the user device while still enabling consistent test group assignment through the anonymized version that is shared with the experimentation system.
Solution Approach 2:
An anonymized identifier acts as an intermediary between the user's private identifier and the experimentation system. The anonymized identifier preserves the necessary linking capability for consistent test group assignment across sessions while preventing direct exposure of the original user identifier to external systems, thus maintaining data security constraints.
2Adaptability or versatility
If multiple content distribution applications are installed on user devices, then content distribution testing capabilities are enhanced, but ensuring test group exclusivity and data security becomes more difficult
Solution Approach 1:
Each content distribution application maintains its own secure experimentation component instance on the user device, segmenting the testing environment. This allows multiple applications to conduct experiments simultaneously with their own test group assignments while the device-level secure environment ensures that anonymized identifiers from different applications cannot be correlated to compromise test group exclusivity.
Solution Approach 2:
The secure experimentation component is implemented locally on each user device rather than centrally on servers. This local implementation ensures that each application's experiment data and anonymized identifiers remain isolated at the device level, preventing cross-application data leakage while still allowing each application to benefit from consistent test group assignment across its own multiple login sessions.
Data Source
AI summary
Technologies for on-device experimentation include embodiments that receive a request to provide digital content for display in a slot of a user interface display screen of a first device. By a secure execution environment, an identifier received with the request is anonymized. The request is determined to be associated with a content distribution test. The anonymized identifier is assigned to a test group associated with the content distribution test. The secure execution environment receives user interface event data generated by the first device in response to the content distribution test. The received user interface event data is attributed to the test group. An association of the user interface event data with the test group and the content distribution test is provided by the secure execution environment to a second device different than the first device while the identifier and the anonymized identifier are not provided to the second device.