Dedicated vs. shared resources in organizations: Modifying the Design Structure Matrix (DSM) to support consolidation decisions

Abstract

A key organization design principle is to group interdependent elements in the same clusters. This principle is the basis for the Design Structure Matrix (DSM), which can be used to automate the identification of clusters (e.g., teams). However, it may sometimes be necessary to depart from this design principle. We focus on one such situation, where the expected value of consolidating resources is deemed to be higher than the reduction in coordination costs from grouping resources in a modular cluster. We propose a set of criteria that can be used to assess consolidation potential as well as consolidation costs. We describe a solution for collecting subjective assessments of the value of consolidation and for incorporating this information into the DSM. This approach makes it possible to continue the use of the already established MDL clustering algorithm. We demonstrate the approach utilizing a simple data set with hypothetical values.