A Bayesian Method for Predicting Future Customer Need Distributions

Abstract

This article presents a Bayesian method to predict future customer need distributions based on forecasts of external factors and forecast accuracies. It is critically important to accurately predict future customer needs when strategically deciding what product to develop. However, prediction is challenging because customer preference changes as external factors that influence customer preference change over time. Consequently, customer needs at the time of product design decisions and at the time of new product market introductions may be significantly different. While external factors can be forecasted, forecasts are not necessarily accurate. This paper outlines a Bayesian approach to incorporate accuracies of forecasts in predicting future customer need distributions. An illustrative example demonstrates impacts of forecast accuracies on customer need distributions and on a strategic design decision. In this example, the customer need is 'fuel-efficient vehicle', the external factor is the gasoline price index (defined as gasoline price divided by consumer price index), and the strategic design decision is the selection of the next vehicle from gasoline-powered, hybrid, and electric. Sensitivity analysis is performed in order to study how a strategic design decision changes with forecasts and forecast accuracies. © The Author(s), 2011.

Department(s)

Mechanical and Aerospace Engineering

Keywords and Phrases

Bayes' Theorem; Customer Need; Design Decision; Distribution; Forecast.

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2011 SAGE Publications, All rights reserved.

Publication Date

01 Jan 2011

Link to Full Text

Share

 
COinS