Out with the old: A Bayesian approach to estimating product modification rates

Drawing on the competitive dynamics literature and incorporating the principles of organizational learning, this study introduces an analytical tool to understand the likelihood of major product modifications using Bayesian logit models in the U.S. automotive industry. We develop and test theoretical hypotheses regarding the effects of the brand’s product portfolio size, brand’s product modification dynamism, competitor product proliferation, and competitor product modification rates on a focal product’s modification likelihood. Our results indicate that as the number of products in the brand’s portfolio increases, firms are less likely to modify existing products. Conversely, an increase in previous product modification rates in both the brand and the industry signals that firms are more likely to consider major product modifications soon. This study contributes to the literature by investigating, on a longitudinal basis, the dynamic nature of products through major alterations.