Simultaneous Location and Scale Monitoring of Key Performance Indicators

This paper presents new methods for simultaneous location and scale monitoring of key performance indicators (KPIs) of a service. Leveraging the Nonparametric Combination of Dependent Tests (NPC) approach, which provides a robust framework for combining dependent tests, we aim to enhance monitoring effectiveness by integrating information from diverse sources to detect anomalies that may impact service quality. Our nonparametric approach offers flexibility in handling various data distributions encountered in real-world environments. High effectiveness of the new methods is shown in simultaneously monitoring both the location and scale. Among the evaluated methods, the Wilcoxon–Klotz method emerges as the top performer, leveraging the inherent flexibility of NPC. Our results underscore the significant potential of combined monitoring methods in practical settings, providing a reliable and versatile approach to ensure effective monitoring of KPIs. A practical application to a call center KPI is discussed.