The impact of structuredness on error probability of process models

Recent research has shown that business process models from practice suffer from several quality problems. In particular, the correctness of control flow has been analyzed for industry-scale collections of process models revealing high error rates. In the past, structuredness has been discussed as a guideline to avoid errors, first in research on programming, and later also in business process modeling. In this paper we investigate the importance of structuredness for process model correctness from an empirical perspective. We introduce definitions of two metrics that capture the degree of (un)structuredness of a process model. Then, we use the Event-driven Process Chain models of the SAP Reference Model for validating the capability of these metrics to predict error probability. Our findings support the importance of structuredness as a design principle for achieving correctness in process models.