Greg and Stephen discuss what you should measure, how often, and what sample sizes you need to reach reliable conclusions. Along the way they discuss butterflies and barnacles, how long it takes customer perceptions to change, and the danger of weaving narratives around volatile survey results.