Concurrent formation of part families and machine cells can be achieved by algorithms based on the part-machine matrix. Although, part features form the basis for the formation of the part-machine matrix, they are not directly considered in the process of forming part families and machine cells. The goal of this paper is to understand the characteristics of machine cells in terms of part features. The relationship between part features and machine cells is captured as a decision tree. The decision tree also becomes a tool for post-clustering analysis including evaluation of different clustering methods and for to classify parts that may appear to belong to more than one group. The simplicity of the decision tree, the number of parts classified by each leaf of the tree, and the effect of pruning on the tree are used as metrics to evaluate the clustering methods.