STOCHASTIC MATRIX APPLICATION TO CONSUMER PREFERENCE PREDICTION

Abstract

This study shows the simplification of brand switching by a given population to one or more other brands or alternatives. It was observed that over time consumers’ preferences changed from one brand of a product to another. These changes were as a result of several factors ranging from taste to cost, availability of product, availability of substitute or similar products, durability, branding, etc. Stochastic matrix, as a common tool in consumer preference analysis, was used. This class of matrices has as its entries the probabilities that a consumer would transit from one state, that is, switch from one brand of product to another. This matrix was obtained from data on past consumer behaviour. The population of consumers was categorized to see the pattern of brand switching and its effect of production. Seven products labeled A-G were used to study the switching patterns and were generalized by means of factorials. The obtained model was then used to predict the volume of consumption of a particular product over a period of ten months.