SUPPLIER SEGMENTATION BASED ON PERFORMANCE DATA USING HIERARCHICAL CLUSTERING

Abstract

In many manufacturing environments, procurement managers must regularly compare suppliers with significantly different performance profiles, which makes structured evaluation essential. Grouping suppliers with similar performance characteristics can make procurement decisions more transparent and easier to justify in practice. However, supplier performance is typically described by multiple evaluation criteria measured on different scales, and these performance characteristics may change over time. For this reason, analytical approaches are required that can handle both the multidimensional nature of the data and the changes in supplier performance over time. In this paper, an adaptive supplier segmentation approach is developed based on hierarchical clustering techniques. Supplier performance is evaluated using multiple operational criteria, including quality rate, on-time delivery, unit price, lead time, complaint frequency, and operational flexibility. After applying min–max normalization and weighted performance evaluation, supplier similarities are calculated using Euclidean distance, and hierarchical clustering methods are applied to identify homogeneous supplier groups. Several linkage strategies are compared, and the clustering quality is assessed using internal validation indices such as the Silhouette, Davies–Bouldin, Calinski–Harabasz, and Dunn indices. The empirical analysis is conducted on a dataset containing 15 suppliers observed in two consecutive evaluation periods. The results indicate that hierarchical clustering can reveal interpretable supplier groups with clearly distinguishable performance profiles. A comparison of clustering structures across consecutive periods also shows that supplier segmentation evolves as new performance data become available.