Social Graph Segmentation for Ad Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Social networking systems face challenges in effectively utilizing social graphs to generate personalized and relevant advertisements that leverage user interactions and relationships, often resulting in ads that are not tailored enough to individual users.
Innovation Solution
The social networking system analyzes social graphs to identify concept nodes associated with applications that are connected to user nodes and their friends, selecting concept nodes based on social relevance, edge types, and advertising sponsorship to generate customized advertisements that reference user interactions and friend activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the social networking system uses general advertisement methods, then the advertisement delivery is simple and fast, but the advertisement relevance to individual users is low
Solution Approach 1:
The patent segments the social graph into concept nodes and user nodes, creating distinct categories for applications, games, and media. This segmentation allows the system to generate targeted advertisements by selecting specific concept nodes relevant to individual users based on their interactions and friend activities, thereby improving advertisement relevance without overwhelming system complexity through structured organization.
Solution Approach 2:
The system performs preliminary analysis of the social graph to identify concept nodes associated with applications that are connected to user nodes and their friends before generating advertisements. By pre-processing and storing relationships between users, friends, and concept nodes, the system prepares relevant advertisement data in advance, enabling personalized ad delivery when users access the platform.
2Measurement precision
If the system analyzes deep social graphs to personalize ads, then ad personalization improves, but processing time increases
Solution Approach 1:
The system performs preliminary analysis of the social graph to identify concept nodes associated with applications that are connected to user nodes and their friends before generating advertisements. By pre-processing and storing relationships between users, friends, and concept nodes, the system prepares relevant advertisement data in advance, enabling personalized ad delivery when users access the platform.
Solution Approach 2:
The patent implements a degree threshold parameter that limits the depth of social graph analysis to only necessary connections (friends of degree 1, 2, or 3). This partial action approach processes only the relevant portion of the social graph needed for personalization rather than analyzing the entire graph, significantly reducing processing time while maintaining adequate personalization accuracy.
3Productivity
If the system references multiple friend activities in ads, then user engagement increases, but ad complexity increases
Solution Approach 1:
The system selectively references friend activities by applying a degree threshold that limits inclusion to friends within a specified number of degrees from the target user. This partial action approach includes only the most relevant friend activities in advertisements, maintaining user engagement through social proof while avoiding excessive ad complexity by excluding distant or less relevant connections.
Solution Approach 2:
The patent applies local quality by differentiating between different types of friend connections (degree 1, 2, 3) and assigning different levels of relevance to each. The system prioritizes activities from closer friends (degree 1) over more distant friends (degree 3), creating advertisements with varying intensities of social reference that maintain engagement without uniformly increasing complexity across all ad elements.
Data Source
AI summary
In one embodiment, a method includes detecting, for each of a plurality of third-party applications corresponding to a respective plurality of pages associated with an online social network, one or more interactions with the page corresponding to the third-party application by one or more first users of the online social network, wherein each of the one or more first users are connected on the online social network to a second user of the online social network. The method also includes calculating, for each of the plurality of third-party applications, a value representing a social relevance of the third-party application based on the number of interactions with the page corresponding to the third-party application by the one or more first users. The method also includes selecting one of the plurality of third-party applications based on its calculated social relevance value. The method also includes sending, to a client device of the second user, an advertisement for the selected third-party application.