While monitoring systems are not new, applying them at scale to business metrics is revolutionary. Using statistics and machine learning algorithms, business monitoring systems can analyze millions of factors that affect business metrics over various time intervals. They continuously detect anomalies, trends, and correlations and present individuals with a handful of the most relevant insights. Unlike prior generations of alerting mechanisms, these systems excel at separating signal from noise: they quickly learn what business users consider relevant and deliver only those insights. Imagine a dashboard that contains 100,000, 1 million, or even 1 billion metric combinations. Better yet, imagine one that notifies you in real time when any of those metrics deviates from its normal range or changes in combination with other metrics or attributes during a given time period, indicating an issue that affects key business outcomes. Imagine not needing to log in to a dashboard, hunt for relevant patterns, and navigate to root causes. The system does all this for you. The business monitoring revolution has just begun. Today, these systems detect anomalies, identify correlations, and display potential root causes. Soon, they may suggest remedies, predict change, and suggest ways to optimize processes to avoid issues in the future. In essence, the systems will automate things that humans can’t do and augment what humans can with real-time recommendations and suggestions.

- Wayne Eckerson in Business Monitoring Systems: Using Machine Learning to Analyze Business Metrics

June 15, 2020 (Report)