Prototyping a data-driven customer segmentation utilizing machine usage data for product portfolio management

Abstract

For companies to have well-fitting product portfolios, product portfolio management requires profound knowledge of its customers. Data-driven decision processes like customer segmentation offer more objectivity than established qualitative assessments. In contrast to existing research, which focuses on customer segmentation based on buying behavior, this paper prototypes a usage-behavior-based customer segmentation for a machine tool manufacturer for sheet metal processing, utilizing machine usage data. Additionally, self-service reports of the customer segmentation are generated. The design choices regarding the data selection, pre-processing, and clustering algorithm selection are verified based on the segmentation results and real-world applicability of the generated reports.