Market share predictions: a new model with rating-based conjoint analysis

Hervé Guyon

Université Paris Sud

Jean-François Petiot

Ecole Centrale de Nantes


In today’s highly competitive market, developing new products able to satisfy consumer needs and preferences is a very important issue. Shortening innovation cycles has now become a key challenge to both company growth and profitability. Risk control in product innovation and the reduction of innovation cycles require valid and rapid customer preference measurements in order to identify new products for preference maximisation. Preference measurements can indeed help with predicting market share and reducing the risk of product failure by assessing customer preferences during the early product development stages (Michalek et al. 2005). For several years, considerable research has been devoted to preference measurements and modelling within various fields of research. In marketing, the influence of product attributes on preference has been investigated by means of both compositional and decompositional methods. The typical decompositional method, Conjoint Analysis (CA) (Green & Rao 1971), is a method for measuring preferences, perhaps the most widely applied by marketing researchers (Chakraborty et al. 2002; Huber 2005). According to Huber (2005), 85% of current managers find CA to be the best adapted and most widely used method during product strategy development phases, 60% of all new product development making use of CA; while 90% of managers consider CA to yield reliable results regarding consumer preferences and to constitute the best testing method for improving new product success rates, and 86% of managers anticipate an increase in the number of future CA applications.