Clustering System Group Customers through Fuzzy C-Means Clustering

Like other economic sectors, it is important to identify, satisfy, and attract profitable customers in the software industry. Organizations have decided to analyze customer behavior and keep the most valuable customers satisfied due to competitive conditions and customer attraction costs. This applieddescriptive study was conducted on the dataset of System Group customers, including 26620 records in 2017. The dataset was analyzed to extract key factors such as the quality of being strategic, the number of software systems, contract sum and customer lifetime. For this purpose, the cross-industry standard process for data mining (CRISP-DM) was employed along with thefuzzy C-means (FCM) clustering algorithm to classify customers and identify profitable and loyal ones. Then the existing data were clustered, and the resultant clusters were evaluated. Finally, the dataset was divided into four major clusters. The first, second, third, and fourth clusters included the special customers (140 members), loyal customers (1800 members), ordinary customers (8960 members), and low-value customers (15720 members), respectively.