Data mining approach for customer segmentation in b2b settings using centroid-based clustering

Big data and advanced analytics in organizations are dominant in customer-centric departments such as marketing, sales, and customer service. For company, designing marketing strategies using customer segmentation is useful to improve business revenue. Clustering algorithms able to deal with large data set to recognize patterns and identify customer segments. In this paper, different clustering algorithms will be compared, specifically centroid-based clustering K-Means, CLARA, and PAM with Fuzzy C-Means clustering. The purpose of this research is to find optimum number of clusters using clustering algorithm with the best validation measure score. Dataset is acquired from Tech Company in Indonesia that provide machine with Point of Sale system for food and beverages merchants, since the company in B2B settings. Among three clustering methods, K-Means have the best validation measure score. After compared to Fuzzy C-Means, K-Means outperforms FCM based on time complexity and quality of clustering. Cluster analysis is done to identify customer information. Therefore, this research able to deliver an insightful understanding about customer characteristics using big data analytics and provide an effective Customer Relationship Management Systems.