Marc Wouters

2025

This thesis presents a practical framework for identifying IT cost optimization potential at the depart-
mental level using K-Means clustering. By combining a multi-dimensional set of key performance
indicator, including Tool per Employee, Cost per Employee, Hardware Cost Ratio, and Inactive
License Ratio—this study uncovers heterogeneous resource consumption patterns that traditional
single-metric analyses often miss. The research applies this methodology to a real-world case study
within a large multinational enterprise, analyzing comprehensive internal data to reveal distinct
departmental profiles. The findings successfully identify four unique clusters: a High-hardware/High-
inactivity cluster with significant potential for immediate cost savings, a High-cost/High-tool cluster
characterized by high IT demand and effective resource utilization, a Low-cost/Low-hardware cluster
with minimal IT infrastructure, and a High-hardware/Low-inactivity cluster with a balanced resource
allocation. By translating these statistical insights into actionable, cluster-specific recommendations,
this research provides a data-driven, replicable analytical framework that bridges the gap between
algorithmic pattern recognition and practical IT governance.