Variation in component properties and dimensions is the most widely recognized cause of product nonconformities. However, conventional statistical methods, like statistical process control (SPC), are ineffective in controlling mistakes. The distinction between mistakes and variation becomes increasingly important as the target nonconformity rates approach extremely low values as substantiated by Motorola's experience. Product complexity increases the likelihood of nonconformities due to both variation and mistakes and is thus a root source of nonconformities. We have shown that assembly complexity, quantified using design for assembly analysis, is highly correlated with nonconformity data in two widely different industries. These correlations and the ability to easily measure assembly complexity permits rapid comparison of the potential nonconformity rates of alternate design concepts.
