Reduction of variation is a crucial ingredient for quality and productivity improvement. An economical and effective approach is to conduct statistically designed experiments, not only to identify controllable factors that affect the mean level (location) of the response, but also to identify and learn about the ones that affect the dispersion. This article explains how control factors can affect the response variability, how they can be identified and estimated, and gives two well‐documented examples that show how this knowledge can be used to improve products and processes.