Anomaly Detection Model for Cost-Per-Click Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately detect and remediate anomalies in digital recommendation costs, such as changes in bidding strategies or data outages, leading to inefficiencies and potential losses.
Innovation Solution
A performance anomaly detection model is established to detect anomalies in cost-per-click data and other performance metrics by analyzing historical trends and seasonality, and automatically initiating alerts and remediation actions when thresholds are exceeded.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional anomaly detection methods are used, then the system can identify performance issues, but the detection accuracy is low and false alarms are frequent
Solution Approach 1:
The patent transforms the anomaly detection approach by changing from static threshold-based parameters to dynamic parameters that adapt to seasonal patterns and historical trends. The system learns optimal detection thresholds from historical data, adjusting sensitivity based on time-of-day, day-of-week, and seasonal variations in performance metrics.
Solution Approach 2:
The system performs preliminary learning and analysis of historical performance data to establish baseline patterns before actual anomaly detection begins. This preliminary action includes training machine learning models on historical cost-per-click data to understand normal variations, enabling more accurate future detection with reduced false alarms.
2Productivity
If manual monitoring and remediation processes are used, then the system can respond to anomalies, but the response time is delayed and efficiency is reduced
Solution Approach 1:
The system implements self-service automation where the anomaly detection model automatically identifies issues and triggers remediation actions without human intervention. When anomalies are detected, the system autonomously initiates corrective measures such as pausing underperforming ads or reallocating budgets, eliminating manual monitoring and response delays.
Solution Approach 2:
The system establishes a closed-loop feedback mechanism where detection results automatically trigger remediation actions, and the outcomes of these actions feed back into the detection model for continuous improvement. This feedback loop enables rapid iterative optimization of both detection accuracy and remediation effectiveness.
3Reliability
If comprehensive performance monitoring is implemented, then the system can detect more issues, but the complexity of the system increases
Solution Approach 1:
The patent segments the complex monitoring task into distinct components: data collection from multiple third-party servers, historical trend analysis, seasonal pattern recognition, anomaly detection, and automated remediation. Each component is handled by specialized modules, reducing overall system complexity while maintaining comprehensive monitoring coverage.
Solution Approach 2:
The anomaly detection model is designed as a universal system that handles multiple performance metrics (cost-per-click, impressions, conversions) across different time scales and seasonal patterns. This multi-functional approach consolidates what would otherwise require separate monitoring systems for each metric and pattern type.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The technology disclosed herein relates to utilizing a performance anomaly detection model to identify performance metrics (e.g., cost-per-click data) that are above an anomaly threshold. For example, communication sessions can be established with one or more servers hosted by one or more third-parties for receiving performance metrics for a first entity. The performance metrics received for the first entity can be used by the performance anomaly detection model, which can be trained using historical performance metrics (e.g., of the first entity, of the first entity during particular time periods, of the first entity for particular geographical locations), for anomaly detection. Based on one or more anomaly detections, one or more notifications or particular displays can be provided to a user device.