Passive Video Analytics via Data Collection Agent
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the popularity and user preferences of online video clips are inaccurate and misleading, as they rely on voluntary ratings and do not account for detailed viewer behavior, and are susceptible to manipulation through scripts and browser refreshes.
Innovation Solution
A data collection agent (DCA) is loaded onto a web video player to passively collect and record detailed viewing information, including user interactions and preferences, which is then sent to a central server for analysis and reporting, allowing for the generation of viewership analytics and preference scoring without requiring user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If voluntary rating systems are used to determine video popularity, then user feedback can be obtained, but the data is inaccurate and susceptible to manipulation
Solution Approach 1:
The patent replaces the manual rating mechanism (users voluntarily clicking rating buttons) with an automated passive monitoring system that uses JavaScript code to track actual viewing behavior. This substitution eliminates the opportunity for manipulation while capturing authentic engagement metrics such as time spent viewing, replay frequency, and interaction patterns.
Solution Approach 2:
The system allows viewing behavior data to speak for itself rather than requiring users to explicitly rate videos. The passive monitoring system automatically collects and analyzes viewing patterns, letting the actual usage data determine video popularity and quality without user intervention or subjective input.
2Loss of information
If detailed viewer behavior tracking is implemented, then accurate user preference information can be obtained, but system complexity increases
Solution Approach 1:
The JavaScript-based passive monitoring system serves multiple functions simultaneously: it tracks viewing duration, detects replays, monitors interaction patterns, and identifies fraudulent activity all through a single integrated code implementation. This multi-functionality reduces overall system complexity compared to implementing separate tracking mechanisms for each metric.
Solution Approach 2:
The patent introduces a JavaScript code intermediary that runs in the user's browser and acts as a bridge between the video player and the analysis system. This intermediary passively collects detailed behavior data without requiring complex server-side processing or user installation, simplifying the overall system architecture while maintaining comprehensive data collection.
3Ease of operation
If passive monitoring without user input is used, then user burden is reduced, but the ability to detect fraudulent activity decreases
Solution Approach 1:
The passive monitoring system continuously collects viewing behavior data and provides feedback patterns that reveal fraudulent activity. By analyzing anomalies in viewing patterns such as impossible replay sequences, suspicious timing patterns, or bot-like behavior, the system can detect fraud without requiring user input or intervention.
Solution Approach 2:
The system replaces manual user rating actions with automated behavioral analysis that inherently detects fraud. The JavaScript code monitors actual viewing mechanics such as playback continuity, pause patterns, and interaction timing, substituting subjective user input with objective behavioral evidence that reveals fraudulent activity.
Data Source
AI summary
Various user behaviors are passively monitored and recorded when a user/viewer interacts with a network video player, e.g. a web video player, while watching an online video clip. For one embodiment, a data collection agent (DCA) is loaded to the player and/or to a web page that displays the video clip. The DCA passively collects detailed viewing and behavior information without requiring any specific input or actions on the part of the user. Indications of user preferences are inferred by user actions leading up to viewing the video, while viewing the video, and just after and still related to viewing the video. The DCA periodically sends this information to a central server where it is stored in a central database and where it is used to determine preference similarities among different users. Recorded user preference information may also be used to rate a video itself.