Efficiency Of Generated Performer Networks In Collaborative Business Process Models

Business process models (PMs) model activities along the execution of business processes. A success factor of process activities is collaboration providing the major part of implicit process knowledge. Applying social network analysis (SNA) on PMs can help understanding the formation of successful organizational structures around PMs. In that context we want to investigate, how the topology of process participants (performers in the following) can be formalized respectively optimized and what are drivers for the efficient interaction between performers and PM. Our approach is to simulate multiple performer networks (PNs) for a corpus containing 9 university admission PMs. Different Social network generators are used in order to generate realistic and extremely structured organizations. The performers in the PN are given random capabilities and according to them, they are associated to the PM functions. For each generated PN, its efficiency is measured by the number of iterations the performers need to finish their tasks at the process functions along the process flow. We describe the topological patterns of the most efficient PNs as we assume them to be success determinants for collaborative business processes. We found optimal parameters to generate efficient PNs for the given corpus. The best efficiency is reached by teams with on average 5 global connected colleagues. This number has the greatest influence on the average PM performance. We also found too many (global) connections between performers to decrease their efficiency and that teams need hierarchy to be efficient in a process.