This document discusses measurement for quality improvement. It explains that measurement in improvement aims to provide a basis for action to improve processes and outcomes, rather than just estimating parameters. Improvement measures should be simple, specific, and available in real-time. Statistical process control methods are important to separate normal variation from changes resulting from interventions. Examples are provided of run charts measuring improvements in recording BMI for mental health patients and compliance with care bundles. The document advocates making the theories behind improvement efforts more explicit.