Simulating Variability of Rework Cost and Market Performance Estimates in Product Redesign

When considering the redesign of an existing product, designers must consider possible engineering and marketing ramifications. Ideal changes capture a large portion of the market and have a low risk of change propagation that results in reduced cost to the manufacturer. Engineering change tools such as the Change Prediction Method and market research models such as Hierarchical Bayes Mixed Logit allow designers to estimate the cost of the redesign process and market shares of preference. Variability in the inputs of the Change Prediction Method (impact and likelihood values) results in a range of redesign cost values. Assumptions regarding model form and the randomness used in model fitting also lead to variations when estimating market performance. When the variability associated with these techniques is considered, focus should shift from a point-estimate to a region-estimate. This paper explores the region-estimate produced for proposed redesigns when considering rework cost and market share of preference.