A new statistical procedure to support industrial research into new product development

In recent years, the competitive context in which goods-producing industrial companies operate has been characterized by strong dynamism and growing attention to strategies capable of maximizing customer satisfaction. In order to reach this objective, the importance of paying particular attention to the initial stages of development of a new product is widely recognized and successful companies therefore invest great resources and competences in industrial research. In this paper we propose a new statistical procedure to support the development of new, successful industrial products on the basis of their experimental performances. New product performances are modelled in a stratified one-way Analysis of Variance (ANOVA) framework and a set of non-parametric permutation tests are applied to detect significant differences in performances from which we can obtain partial rankings and a final global ranking suitable for identifying the best product.