Analog Devices, Inc. (ADI) manages spare parts across four front-end fabrication sites through largely decentralized site-level practices, a strategy that creates limited network-wide visibility and inconsistent inventory logic. To address these challenges, this capstone develops a cluster-based segmentation framework to improve how spare parts are classified and managed across sites. The framework begins by considering six initial indicators—lead time, supplier pool, demand pattern, demand growth, criticality, and price—which are summarized into three composite indicators capturing supply risk, demand risk, and price exposure. It then segments spare parts into six risk-based clusters and applies Data Envelopment Analysis (DEA)-based scoring to prioritize clusters and guide differentiated inventory policies. The results show that different clusters are driven by different underlying risk factors, supporting the need for differentiated inventory policies rather than a one-size-fits-all approach. For example, some clusters are primarily driven by supply exposure and cost, while others are driven by demand variability. A volume-based ABC validation further shows that high-volume items are distributed across multiple risk profiles rather than concentrated in low-risk segments. These findings suggest that aligning inventory policies with the primary drivers of risk can improve consistency in decision-making and enable more scalable inventory management across sites.
