Improving customer relationship management through data mining based on the GMDH algorithm

Effective customer relationship management has become a major challenge in business competition. Stores need information about who their customers are, what their expectations and needs are, and how their needs should be met. One of the problems of grocery stores in previous years was the lack of sufficient information of managers about the preferences and shopping carts of customers and, thus, the lack of goods needed by customers in the warehouse and improper layout of stores to provide customers with the goods they need. The present study aims to analyze the data of store customers using neural networks and obtain sufficient information from the customer's shopping cart in the grocery store. The results showed that good clustering enables companies to increase their relationship with customers and increase their sales. Also, the layout of each store can be defined according to preferences for the goods needed by customers who visit the store more so that increased comfort leads to increasing their purchases. Prominent goods in each cluster, business units (goods groups) with a share of goods at least twice as high as the mean share in the sample can be distinguished.