Targeted Notification Selection for Social Networking Engagement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Social networking services face challenges in increasing user participation and engagement, as traditional blanket marketing efforts result in high bandwidth consumption and unwanted messages, with limited success, leading to reduced user interaction.
Innovation Solution
A system and method to selectively identify and notify users who are most likely to engage with a social networking service by analyzing user activity, past behaviors, and preferences, using a machine learning process to determine an engagement score and target specific users with tailored notifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If blanket marketing efforts are used to reach all subscribers, then user participation may be increased, but bandwidth consumption increases substantially and user experience deteriorates due to unwanted messages
Solution Approach 1:
The patent applies local quality by transitioning from uniform blanket marketing to personalized targeted marketing. Each user receives customized notifications based on their individual engagement scores, activity levels, and preferences. The system calculates unique engagement scores for different user segments and delivers tailored messages only to those most likely to respond, thereby reducing overall bandwidth consumption while maintaining or improving user participation rates.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting marketing notification parameters based on user behavior data. The system monitors user activity levels, engagement patterns, and preferences, then modifies notification parameters such as timing, frequency, and content type. Engagement scores are recalculated based on changing user parameters, allowing the system to optimize bandwidth usage by sending notifications only when the probability of positive user response exceeds certain thresholds.
2Area of stationary object
If blanket marketing efforts are used to reach all subscribers, then coverage is maximized, but user experience deteriorates due to spam and unwanted messages
Solution Approach 1:
The patent applies local quality by segmenting the user base into distinct groups based on engagement scores and delivering differentiated notification strategies to each segment. High-engagement users receive different notification types and frequencies compared to low-engagement users. This personalized approach ensures comprehensive coverage of all user segments while minimizing unwanted messages by tailoring communication to user preferences and behavior patterns.
Solution Approach 2:
The patent implements feedback mechanisms by continuously monitoring user responses to marketing notifications and using this data to refine engagement scores and future notification strategies. The system tracks user interactions, adjusts engagement scores based on observed behavior, and modifies subsequent notification targeting accordingly. This closed-loop feedback system reduces unwanted messages by learning from user responses and adapting targeting criteria to minimize spam perception while maintaining effective coverage.
3Loss of energy
If selective notification based on engagement scores is implemented, then bandwidth consumption is reduced, but system complexity increases due to machine learning processes
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing engagement scores for all users before marketing campaigns are launched. The machine learning models are trained in advance on historical user data to establish baseline engagement patterns. During actual campaign execution, the system simply retrieves pre-computed engagement scores and applies straightforward filtering logic, rather than performing complex real-time calculations. This preprocessing approach significantly reduces the computational complexity during notification delivery while maintaining effective bandwidth optimization.
Solution Approach 2:
The patent utilizes copying by creating simplified surrogate models of user engagement behavior that can be quickly evaluated. Instead of running full machine learning inference for each notification decision, the system uses pre-trained models to generate engagement scores that serve as simplified proxies for complex user response predictions. These score copies enable fast, low-complexity decision-making during campaign execution while capturing the essential patterns learned from extensive training data, thereby reducing system complexity during operation.
Data Source
AI summary
Systems and methods for identifying a selected set of users of a social networking service which, upon issuing a notice regarding a re-post activity of an item of content, will likely result in heightened user interaction with the service. Upon receiving a notice of a re-post action by a first user, the social networking service identifies a first set of users of the service that have posted that item of content. Scores are associated with the users of the first set of users, the scores indicating a likelihood of a user that, if notified of the re-post action of the first user, will result in heightened user interaction with the social networking service. A subset of users of the first set of users are identified according to their associated scores and a notice of the re-post action by the first user is sent to the users of this subset of users.