Proposing a New Method for Customer Segmentation Based on Their Level of Loyalty and Defining Appropriate Strategies for Each Segment

The evaluation of customer loyalty have a significant impact on improving business processes. Ordinary methods of customer loyalty evaluation have been designed based on three components including; recency of transactions (R), the frequency of transactions (F) and the monetary value of transactions (M). In this study, it has been attempted to examine some affective factors, including the total number of purchased goods, returned goods, discounts and the average delay of distribution and their impact on increasing quality of assessment be measured. The main objective of the current study is to propose a new model for customer segmentation based on their level of loyalty and to define appropriate strategies for each segment. The data set for this study is obtained for the customers of a food wholesale. The obtained data have been analyzed using Clementine ۱۴.۲ software application using MLP and RBF neural networks as well as the K-means algorithm. The results of the study show that the proposed method provides the highest level of accuracy for predicting the customers’ loyalty. Based on this proposed method, the customers are divided into five clusters (Loyal, potential, new, lost and churn customers) from the point of view of loyalty, with the characteristics of each cluster expressed based on the status of seven factors. Based on these characteristics, appropriate approaches for managing the customers in each segment are proposed.

Keywords:

customer segmentation , Data Mining , evaluation of customer loyalty , food wholesale