Context-Based Content Selection for Anonymous Requests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online systems face challenges in selecting relevant content for presentation when requests from third-party systems or applications do not provide user identification, leading to reduced interaction and revenue, as the selected content may not be of interest to users.
Innovation Solution
The online system maintains information about content presentation contexts, including user characteristics, application details, and previous content interactions, to identify users who were presented with similar content in the past, generating characteristics for users who shared common traits in those contexts, allowing for targeted content selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the online system selects content based on user identification information, then content relevance to users is improved, but the system cannot handle requests from third-party systems or applications that do not provide user identification
Solution Approach 1:
The patent introduces context information as an intermediary element that bridges the gap between content selection and user identification. When user identification is unavailable from third-party systems, the system uses context information (device characteristics, application data, previous interactions) to infer user characteristics and select relevant content, thereby maintaining both accuracy and compatibility
Solution Approach 2:
The system performs preliminary actions by collecting and analyzing context information in advance, building user profiles and content preferences before actual content selection occurs. This allows the system to quickly retrieve pre-analyzed data when user identification is provided, and to infer user characteristics from stored context data when identification is missing
2Measurement precision
If the online system requests user identification information from third-party systems, then content selection accuracy is improved, but this increases system complexity and may reduce ease of operation
Solution Approach 1:
The system applies partial action by requesting only the minimum necessary user identification information from third-party systems, rather than comprehensive user profiles. When full identification is not provided, the system uses available partial information combined with context analysis to achieve sufficient content selection accuracy without increasing complexity
3Measurement precision
If the online system collects and maintains detailed context information about content presentations, then future content selection is improved, but information storage requirements and processing complexity increase
Solution Approach 1:
The system extracts only the most relevant and distinctive features from context information, rather than storing complete context data. By identifying and retaining only key characteristics that are most predictive of user preferences, the system maintains high content selection accuracy while significantly reducing data storage requirements and processing complexity
Data Source
AI summary
An online system maintains information identify a context in which sponsored content items were presented to users. A context in which a sponsored content item was presented to a user identifies additional content presented to the user prior to the sponsored content item, and may identify additional content presented in conjunction with the sponsored content item. The online system identifies users to whom at least one sponsored content item was presented in a context and generates characteristics for the context based on characteristics of users who were presented with at least one sponsored content item in the context. When the online system receives a request to present sponsored content items in the context that does not identify an online system user, the online system selects sponsored content items for the request based on the generated characteristics for the context.