Anomaly Detection Subsegment Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Web commerce analytics rely heavily on sales data, which is insufficient for making proactive decisions and identifying anomalies in user interactions, as it does not provide comprehensive insights into website performance and user behavior.
Innovation Solution
Anomaly detection systems using machine learning models to identify deviations in metrics like traffic and conversions, breaking them down into subsegments based on user attributes, and generating alerts when these deviations exceed predicted ranges, allowing for proactive decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sales data is used to determine product success, then product performance can be measured, but comprehensive insights into user behavior and website performance cannot be obtained
Solution Approach 1:
The patent segments user interactions into multiple metrics including traffic data, conversion data, and various user behavior metrics. By dividing the analysis into different metric categories and subsegments (such as device type, browser type, country, platform, location, website source, and user status), the system comprehensively captures user behavior patterns that sales data alone cannot reveal, thus resolving the information loss while maintaining measurement precision.
2Loss of time
If traditional web analytics are used, then historical sales performance can be tracked, but proactive decisions cannot be made due to insufficient data
Solution Approach 1:
The system performs preliminary actions by continuously collecting and analyzing multiple types of user interaction metrics in real-time, before sales outcomes occur. By establishing baseline patterns and using machine learning models to predict expected performance, the system enables proactive decision-making through anomaly detection, allowing teams to take preventive or corrective actions before sales performance deteriorates.
Solution Approach 2:
The patent implements feedback mechanisms by comparing actual user interaction metrics against expected values generated by machine learning models. When anomalies are detected (deviations from expected patterns), the system generates alerts that provide timely feedback to web commerce teams, enabling them to make informed proactive decisions based on comprehensive real-time data rather than waiting for historical sales results.
3Measurement precision
If multiple user attributes are analyzed to identify anomaly sources, then subsegment contribution can be determined, but system complexity increases
Solution Approach 1:
The patent introduces machine learning models as intermediaries that automatically process and analyze multiple user attributes (device type, browser type, country, platform, location, website source, user status). These models compute expected values for various metrics and their subsegments, then compare actual values against expectations to identify anomalies. This intermediary layer handles the computational complexity internally while presenting simplified anomaly identification results to users, thus maintaining measurement precision without exposing system complexity.
Data Source
AI summary
Systems and techniques may be used for generating an alert for a website based on user interactions at the website using a machine learning trained model. A technique may include receiving user interaction metrics corresponding to the user interactions, determining, using the machine learning trained model, whether a trajectory of a metric of the user interaction metrics is predicted to traverse a threshold, and in response to determining that the trajectory is predicted to traverse the threshold at a particular time, generating the alert. The technique may include evaluating a set of subsegments of the metric, each subsegment of the set of subsegments corresponding to respective metric source attributes, determining, based on the evaluation, at least one subsegment of the set of subsegments that most contributed to the trajectory, and outputting an indication for display, the indication including identification of the at least one subsegment with the alert.