A Bayesian framework for predicting customer need distributions

Predicting future customer needs is critical when selecting a concept for a new product. Customer need prediction is challenging because customer needs may change as external factors that influence needs change over time. This paper proposes a Bayesian framework to predict future distribution of customer needs by incorporating forecasts of external factors and their corresponding accuracies. The framework is demonstrated by an illustrative example in which designers predict future distribution of a customer need (“Fuel Efficient”) based on forecast of an external factor (gasoline price index) and the accuracy of the forecast. The benefit of incorporating forecasts of the external factor on concept selection and a sensitivity analysis of concept selection on the accuracy of the forecast are demonstrated in the illustrative example.Copyright © 2010 by ASME and General Motors Company