Encrypted Identifier Matching for Privacy-Preserving Attribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
User privacy concerns in mobile and web applications hinder effective data analysis for targeted advertising, as users increasingly opt out of sharing identifiable information, complicating the matching of impression and conversion datasets for attribution, lift, and ranking metrics.
Innovation Solution
An advanced matching system segregates impression and conversion datasets into separate data buckets, uses minimum bucket sizes, constructs user identifiers by concatenating data fields, hashes them, and computes intersections to maximize matching likelihood while preserving privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If user identifiable information is collected for targeted advertising, then advertising effectiveness is improved, but user privacy is compromised
Solution Approach 1:
The patent introduces encrypted identifiers as an intermediary between user data and advertising systems. These encrypted identifiers enable matching of impression and conversion events without exposing actual user identities, thus maintaining advertising effectiveness while protecting user privacy through the intermediary encryption layer
Solution Approach 2:
The patent extracts and removes identifiable information from the data collection process. By using encrypted identifiers that contain no personally identifiable information, the system separates the functionality needed for advertising attribution from the harmful element of user identification
2Object-affected harmful factors
If encrypted identifiers are used for matching datasets, then user privacy is preserved, but matching accuracy deteriorates
Solution Approach 1:
The patent applies preliminary actions by pre-processing data into encrypted identifier formats before matching operations. By constructing encrypted identifiers with specific properties (deterministic encryption, consistent formatting) in advance, the system enables accurate matching while maintaining privacy throughout the attribution process
Data Source
AI summary
Systems and methods herein describe privacy preserving multi-touch attribution. The described systems access a plurality of impression events and a plurality of conversion events, and for each impression event and each conversion event, wherein each impression event and each conversion event are associated with user identifiers, the described systems generates a hashed user identifier based on the associated user identifier, initiates a key agreement protocol comprising a key, generates an encrypted identifier by encrypting the hashed user identifier with the key, and stores the encrypted identifier.