Supplier Segmentation with Partial Orders, Tiered DEA and Cluster Analysis

Ranking suppliers performance is one of the central issues of supplier management. In supplier selection, most quantitative analysis assumes that any two suppliers are comparable on any criteria. This assumption is then used to rank the suppliers. In practice, however, suppliers are not always clearly comparable, while the aim of the evaluation is to manage the supply base. This can be expressed in terms of not establishing a linear ranking of suppliers, but rather grouping them according to some idea. There are two approaches to this: outranking suppliers or splitting suppliers into groups. The latter is rarely used in supplier evaluation. We will demonstrate and compare the application of Partially ordered set (POSET), Tiered Data Envelopment Analysis (TDEA), and Cluster Analysis on a numerical example of a supplier evaluation problem. Our results show that the methods lead to similar results and that software is available to solve them