The elements of this vector, which is specified more precisely in the next section, indicate which services a given customer has chosen in a specific subscription period. Moreover, suppose that in order to find association across products, analysis of customers’ usage data by hierarchical clustering method has resulted in a dendrogram shown in Fig. 1. In this case, it can be inferred that customers choose three different groups of services. Correlations between pairs of services are shown as a matrix in Table 1. Admittedly, cross correlations verify the result of clustering.