Cross-Device User Identification via Behavioral Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to track user behavior across multiple devices and distinguish between different users on the same devices, leading to ineffective targeted advertising and inaccurate analytics.
Innovation Solution
A system that uses behavioral data to create a device graph, propagates contexts, and calculates correlation matrices to score device relationships, allowing for the identification of similar user behavior across devices and distinguishing between users, thereby enabling more precise targeting and accurate analytics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional tracking techniques (cookies, pixels) are used to track user behavior, then advertising targeting can be delivered to individual users on single devices, but the system cannot track users across multiple devices or distinguish between different users on the same device
Solution Approach 1:
The patent segments the tracking system into device-level identifiers and user-level profiles. Each device is assigned a unique identifier (e.g., device graph nodes) while user profiles aggregate behavior across multiple devices. This segmentation allows the system to distinguish between different users on the same device and track individual users across multiple devices, resolving the contradiction between single-device precision and multi-device versatility.
Solution Approach 2:
The patent introduces a new dimension of user identification by creating device graphs that map relationships between devices and users. Instead of relying solely on device identifiers or cookies, the system builds a multi-dimensional view where users are connected to multiple devices through behavioral patterns and contextual data. This dimensional expansion enables cross-device tracking while maintaining user distinction.
2Loss of information
If device-level tracking is implemented, then advertising can be targeted to specific devices, but the system cannot discern between different users operating the same device
Solution Approach 1:
The patent introduces device graphs as an intermediary structure that connects devices to users through behavioral patterns. The device graph acts as a mediator that aggregates signals from multiple sources (device identifiers, behavioral data, contextual information) to infer user identity and relationships. This intermediary approach preserves user distinction information without requiring direct user identification on each device, managing the complexity through structured data relationships.
Solution Approach 2:
The patent changes the parameters used for user identification from device-centric (device IDs, cookies) to behavior-centric parameters (usage patterns, contextual signals, temporal patterns). By shifting the identification parameters from hardware-based to behavior-based metrics, the system can distinguish between users on the same device while maintaining a manageable complexity level through pattern recognition rather than direct identification.
3Measurement precision
If behavioral data from multiple sources is collected to improve user identification, then cross-device tracking accuracy improves, but data processing complexity and computational requirements increase
Solution Approach 1:
The patent implements preliminary action by pre-building device graphs and establishing device relationships before advertising campaigns are launched. The system proactively collects and processes behavioral data to create user profiles and device mappings in advance, rather than attempting to identify users in real-time during ad delivery. This preliminary processing reduces computational complexity during actual advertising operations while maintaining high identification accuracy.
Solution Approach 2:
The patent uses copying by creating simplified representations of user behavior and device relationships in the form of device graphs and aggregated profiles. Instead of processing all raw behavioral data for each advertising decision, the system creates condensed copies (profiles, graphs, scores) that capture essential user characteristics and device relationships. These copied structures enable efficient querying and matching while reducing the complexity of handling full behavioral datasets.
Data Source
AI summary
A system and method is provided to track users across applications and devices, and to distinguish similar or different users among a number of commonly-used devices. Accordingly, systems and methods are provided for analyzing user device information to determine different users among a number of devices. In particular, user's behavior when using the devices is observed, and responsive to that observed behavior, is the determined whether a particular users are distinguishable from other users. For instance, it may be useful to distinguish between users that operate the same devices or exist within a similar network (e.g., a home network, a workplace network, etc.) for various purposes. For instance, it may be useful to distinguish between users for displaying advertising to these users, performing some other operation with respect to a subset of the users.