Source-Related User Activity Measurement for Recommendation Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content publishers face challenges in accurately measuring the quality of content recommendations provided to users, which affects user engagement and satisfaction.
Innovation Solution
A system and method for calculating user engagement measurements by collecting user activity data during an activity window, identifying user sessions, and calculating source-related user activity measurements based on interactions with recommendation sources, such as widgets or applications, to assess the quality of content recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If content publishers provide additional recommended content to users, then user engagement and satisfaction improve, but the ability to accurately measure recommendation quality deteriorates
Solution Approach 1:
The patent segments user activity measurement into source-related user activity (specific to recommendation sources) and general user activity. This segmentation allows precise measurement of recommendation quality by isolating activities directly attributable to recommended content, resolving the contradiction between providing more recommendations and measuring their quality accurately.
Solution Approach 2:
The patent implements a feedback mechanism where source-related user activity measurements are calculated and used to assess recommendation quality. This feedback loop enables publishers to measure the impact of recommended content and adjust their recommendation strategies, thereby maintaining measurement precision while continuing to improve user engagement.
2Measurement precision
If content publishers track detailed user activity data to measure recommendation quality, then measurement precision improves, but system complexity increases
Solution Approach 1:
The patent extracts only the necessary source-related user activity data from the broader user activity dataset. By taking out only the specific activities related to recommendation sources (such as clicks on recommended content, time spent on recommended pages), the system achieves precise measurement without the complexity of processing all user activity data.
Solution Approach 2:
The patent establishes activity windows and pre-defines source-related user activity metrics before data collection begins. This preliminary action simplifies the data processing system by setting clear boundaries and definitions in advance, eliminating the need for complex real-time analysis of all user activities.
3Measurement precision
If content publishers implement comprehensive user activity tracking, then recommendation quality assessment improves, but user privacy concerns increase
Solution Approach 1:
The patent applies local quality by measuring only the specific user activities related to recommendation sources rather than tracking all user behaviors. This localized measurement approach maintains sufficient precision for assessing recommendation quality while minimizing privacy intrusion by focusing only on activities directly related to the recommended content.
Data Source
AI summary
Identifying impressions relating to a target publisher that are related to or derived from a user interaction with a content recommendation source. User activity data for multiple users is collected during an activity window. Based on the collected user activity data, an initial interaction by a user with a source is identified and used to establish a source-related user session beginning at a time of the initial interaction and ending after a session period. A set of impressions (e.g., page views) by the user relating to the target publisher occurring during the user session is identified. The identified set of impressions is associated with the user session. A source-related user activity measurement is calculated based on the identified user sessions and associated impressions occurring during the activity window.