User Identity Correlation via Activity Pattern Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Internet advertising faces challenges in accurately identifying users across multiple devices due to the inconsistency of non-login IDs, leading to poor targeting profiles and frequency capping issues, which affects advertising revenue and campaign performance.
Innovation Solution
A method and system that analyze user activity logs to identify correlating activity information across different identifiers, such as cookies and device profiles, by filtering out popular activities, building an inverted index, enumerating potential user groups, and scoring them based on common websites and inverse document frequency, to assign identifiers to a common user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If non-login IDs (cookies, device IDs) are used to track users, then user identification is simplified, but identifier consistency deteriorates due to cookie churn and device updates
Solution Approach 1:
The patent introduces an intermediary system that collects and analyzes multiple signals (cookies, device IDs, IP addresses, browsing behavior) to infer user identity. This intermediary layer reconciles the inconsistency of individual identifiers by cross-referencing multiple data points and behavioral patterns to determine whether different IDs belong to the same user.
Solution Approach 2:
The system dynamically changes the parameters used for identification by switching from relying on a single static ID to using multiple volatile parameters (cookies, device IDs, IP addresses) in combination with behavioral parameters (browsing history, activity patterns). This allows the system to adapt to identifier changes while maintaining user identification accuracy.
2Measurement precision
If multiple identifiers are tracked to improve user identification accuracy, then profiling quality improves, but system complexity increases
Solution Approach 1:
The patent segments the identification process into distinct components: signal collection (gathering cookies, device IDs, IP addresses), signal processing (analyzing browsing behavior and activity patterns), and user inference (determining user identity by correlating signals). This segmentation allows each component to be optimized independently while managing overall system complexity.
Solution Approach 2:
The system creates a universal user identification framework that works across multiple devices and platforms. By developing a multi-functional approach that handles various identifier types (cookies, device IDs, IP addresses) and behavioral signals within a single system, the patent achieves comprehensive user tracking without proportionally increasing complexity.
3Productivity
If user activity data is extensively collected to improve targeting profiles, then advertising effectiveness improves, but data processing requirements increase
Solution Approach 1:
The patent extracts only the most relevant features from the collected activity data for user identification and profiling purposes. Instead of processing all raw activity data, the system identifies and extracts key behavioral patterns and signals that are most indicative of user identity and preferences, reducing processing requirements while maintaining effectiveness.
Data Source
AI summary
Systems and methods for identifying users according to their activity are disclosed. The identification of a user includes accessing a user activity log having a plurality of identifiers and corresponding activity information for each identifier, determining identifiers having correlating activity information, and assigning identifiers having correlating activity information to a common user.