Abstract
This paper examines various interpretations of key characteristics. Starting with Juran's dictum of focusing on "vital few"and not on "trivial many", it looks at a representative sample of extant definitions and examples. Causality and sensitivity are then discussed at some length. It is argued that it is the weighted sensitivities that determine what characteristics are key and that literature on design of experiments has supported this view all along. Finally, a statistical interpretation of key characteristics is offered as a unifying framework to capture many of the recent interpretations of key characteristics.