A Market Basket Analysis Based on the Multivariate MNL Model

Abstract

The following research is guided by the hypothesis, that products chosen on a shopping trip in a supermarket are an indicator of the preference interdependencies between different products or brands. The bundle chosen on the trip can be regarded as an indicator of a global utility function. More specific: the existence of such a function implies a cross–category dependence of brand choice behavior. It is hypothesized, that the global utility function related to a product bundle is the result of the marketing–mix of the underlying brands. To investigate the determinants of the choice for a certain bundle, a market basket forecast model is adopted from Russel and Petersen (2000) which uses a multivariate logistic function. The target of this paper is to apply a multivariate logistic approach to estimate a market basket model and to make a comparison between the results of the parameter estimates for a Canadian data set with a German one, which leads to a cross–cultural study. To our knowledge the adoption of this model type to a German data set is shown the first time. The estimation technique is derived from models of spatial statistics and will be explained here in much more detail than in Russel and Petersen (2000). The structure of the chosen product categories allow to discover the impact of certain marketing–mix variables and cross national comparison of market basket choice respectively product bundle buying behavior.