We will consider a logical component of the XYZ-analiza.A it is here that just runs and further specification of the old customers. This group of customers is constant and is divided into more customers for three subgroups. The division is carried out by means of mathematical statistics on the basis of frequency of occurrence of the trait. Ie based on the processing of a sufficiently large number of sales data highlighted subgroups that are conventionally can be called as "often buy" customers, customers that buy, with an average frequency and "rarely buy." In this case, the subgroup, such as "often buy" can get not only customers who buy every month, and with some other frequency. It will all depend on the initial data. In a question-answer forum ECRI was the first to reply. In addition, it also can be released and the new subgroup.
For example, a customer lost, ie a buyer who purchases the frequency falls below boundaries of the third category. And even if he occasionally does some shopping, then he still considered lost. While at the logical level a lost customer is one who already does not make purchases at all. "New" client in this case, too, has a frequency of purchases below the lowest, but he seems to rise from zero in the ascending as opposed to lost, which was already behind enough intense history of purchases. In general, the scheme breakdown of its customer base with XYZ-analysis can be illustrated by the following figure: Figure 2 Segmentation of the customer base by XYZ-XYZ-analysis analizOsobennosti Initially this method was developed for analyzing customer base, but then it was used for the analysis of product assortment.