This document discusses the limitations of using average response time to measure web application performance and introduces Apdex as a better metric. It explains that average response time does not account for outliers and can be distorted by extremely high response times. Apdex calculates a satisfaction score based on thresholds for satisfied, tolerating, and frustrated users, which provides more useful information for evaluating performance and user experience compared to average response time alone. The document recommends using Apdex instead of average response time to make "go/no-go" decisions and better understand the number of users affected by performance issues.