Social Network Notification Relevance Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Social networking systems are ineffective in surfacing dated but relevant information and fail to provide contextually and temporally relevant information to users based on their current circumstances and predicted interactions with other users.
Innovation Solution
A social networking system predicts user interactions and determines relevance scores for information items about target users, using occurrence type values, affinity values, and inferred affinity values to select and communicate relevant information to recipient users through notifications, which can be requested or automatically pushed, based on their interests and interaction plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If social networking systems display recently received social information based on current snapshot, then the information is timely and relevant to current context, but dated but relevant information is not surfaced and users miss contextually relevant information
Solution Approach 1:
The system pre-processes and stores social information with metadata about its relevance to different user contexts and time periods. When a user views a target user's profile, the system retrieves pre-processed information that is relevant to that specific context, rather than only showing recently received information. This allows dated but contextually relevant information to be surfaced when needed.
2Ease of operation
If social networking systems organize and present recently provided relevant information, then the information is current and easy to access, but the system fails to provide contextually and temporally relevant information based on user circumstances
Solution Approach 1:
The system tailors the information presented to each specific user context by analyzing the relationship between the viewing user and the target user. Different users viewing the same target user receive different sets of relevant information based on their specific circumstances, such as whether they are meeting for the first time, are friends, or have shared interests. This makes the information both easily accessible and highly adaptable to local contexts.
Solution Approach 2:
The system pre-calculates and stores relevance scores for different types of information based on user relationships and contexts. When a user views a profile, the system quickly retrieves pre-computed relevant information rather than searching in real-time, maintaining ease of operation while providing contextually adapted content.
3Reliability
If social networking systems provide comprehensive information about target users, then users have more context for interactions, but the system becomes complex and difficult to manage
Solution Approach 1:
The system extracts only the most relevant pieces of information about target users based on the viewing user's context and relationship. Rather than providing comprehensive information about all aspects of a user's profile, the system selectively presents information that is most useful for that specific interaction context, such as shared interests, mutual friends, or recent life events relevant to the relationship.
Solution Approach 2:
The system uses relevance scoring to dynamically adjust which information is presented based on multiple parameters including user relationship type, interaction context, and information recency. This parameter-based filtering simplifies the system by automatically determining what information to show rather than requiring manual configuration of comprehensive information displays.
Data Source
AI summary
To provide more contextually and temporally relevant information to its users, a social networking system may surface relevant information about a target user with whom an recipient user is likely to interact. The social networking system predicts whether the recipient user is likely to be in contact with a target user, either currently or in the near future. If contact is predicted, the system determines information about the target user that that the recipient user may find of interest for their interaction. To determine what information may be of interest, the system determines a relevance score for information items about that target user. The system can then use the relevance scores to decide whether and which information items to surface to the recipient user, e.g., in the form of notifications.