1 Quality Metrics for Business Process Models Irene Vanderfeesten 1 , Jorge Cardoso 2 , Jan Mendling 3 , Hajo A. Reijers 1 , Wil van der Aalst 1 1 Technische Universiteit Eindhoven, The Netherlands; 2 University of Madeira, Por- tugal; 3 Vienna University of Economics and Business, Austria. SUMMARY In the area of software engineering, quality metrics have shown their importance for good programming practices and software designs. A design developed by the help of these metrics (e.g. coupling, cohesion, complexity, modularity and size) as guid- ing principals is likely to be less error-prone, easy to understand, maintain, and manage, and is more efficient. Several researchers already identified similarities between software programs and business process designs and recognized the po- tential of quality metrics in business process management (Cardoso, Mendling, Neuman & Reijers, 2006; Gruhn & Laue, 2006; Latva-Koivisto, 2001). This chapter elaborates on the importance of quality metrics for business process modeling. It presents a classification and an overview of current business process metrics and it gives an example of the implementation of these metrics using the ProM tool. ProM is an analysis tool, freely available, that can be used to study process models im- plemented in more than eight languages. I NTRODUCTION Key in many instances of innovation is the transfer of information and understand- ing developed in one discipline to the other (Kostoff, 1999). A prime example is workflow management itself, a technology based on the importation of process models from manufacturing operations into administrative work. In this chapter we embark on further opportunities for knowledge transfer to the field of process mod- eling and workflow management, in particular from software engineering. In the mid-1960’s software engineers started to use metrics to characterize the properties of their code. Simple count metrics were designed to be applied to pro- grams written in languages such as C++, Java, FORTRAN, etc. Various of these metrics provided a good analysis mechanism to assess the quality of the software program design. Since there is a strong analogy between software programs and business processes, as previously argued in (Reijers & Vanderfeesten, 2004; Guce- glioglu & Demiros, 2005), we believe that software metrics, such as coupling, cohe- sion, and complexity, can be revised and adapted to analyze and study a business process' characteristics. A business process model, regardless whether it is modeled in e.g. BPEL, EPC, BPMN or Petri Nets, exhibits many similarities with traditional software programs. A software program is usually partitioned into modules or functions (i.e. activities), which take in a group of inputs and provide some output. Similar to this composi- tional structure, a business process model consists of activities, each of which con- tains smaller steps (operations) on elementary data elements (see Table 1). More- over, just like the interactions between modules and functions in a software pro-