The myth of accelerating business processes through parallel job designs

This paper discusses postulated advantages of parallel job designs compared with sequential designs in business processes. Until now, the literature suggests an overall dominance of parallel designs. An analysis of relevant application domains, in which parallel tasks are prevalent in order to accelerate processes, offers two remarkable insights: (1) parallel tasks can but do not have to speed-up business processes; (2) coordination efforts may reduce or even invert potential performance gains. Thus, this paper thoroughly evaluates interdependencies between parallel designs and coordination efforts through simulation experiments. Several process patterns for order processing in a real-world example will be examined. The results offer noteworthy characteristics of parallel designs which extent the existing knowledge: (a) gains from parallelization are reduced with increasing process variability, (b) small increments in coordination efforts neutralize gains from parallelization, (c) in high work load situations, the resource capacity for the coordination task is a severe bottleneck, and (d) if the findings (a)–(c) are carefully considered, only the parallelization of multiple tasks leads to significant performance gains.