Dynamics in International Market Segmentation of New Product Growth

Abstract

Prior international segmentation studies are static, i.e., they identify segments that are stable over time. This paper shows that country segments in new product growth are intrinsically dynamic. We propose a semiparametric hidden Markov model to dynamically segment countries based on the observed penetration pattern of new product categories. The methodology allows countries to switch between segments over the new product's life cycle, with time-varying transition probabilities. Our approach is based on penalized splines and can thus be flexibly applied to any non-stationary phenomenon, beyond the new product growth context.For the penetration of six new product categories in 79 countries, we recover each country’s dynamic membership to segments over the life cycle. Our findings reveal substantial dynamics in international market segmentation, in particular at the beginning of the product life. Finally, we exploit the dynamic segments to predict the national penetration patterns of a new product before its launch and show that our forecasts outperform forecasts derived from alternate parametric and/or static methods. Our results encourage multinational corporations to adopt dynamic segmentation methods instead of static ones.