Kenneth Hartmann : abstract
|Multivariate Statistical Methods In Marketing
|The battle for markets has intensified dramatically in recent years. Marketers are no longer able to view their customers as a single homogeneous group characterized by a small number of demographic variables. Instead, they must develop strategies to appeal to numerous customer segments with distinct demographic and behavioral characteristics and different sets of wants and needs. Advances in multivariate statistical techniques and computer software have provided powerful tools for analyzing and modeling these complex relationships. This presentation will discuss multiple discriminant analysis, which is a widely used technique for predicting which group a customer belongs to based on a set of independent variables. The groups are subsequently profiled in terms of these variables to highlight important differences. Applications include classifying customers based on likelihood of buying a product, credit risk, brand loyalty, or lifetime value. A short case study will illustrate how discriminant analysis works in practice.