This article suggests an approach which allows the costly analysis of processes (e.g., in serviceoriented architectures) for benchmarking to be partially automated, so that the performance indicators, as well as qualitative differences between processes become apparent. The approach is based on using appropriate ontologies, which make the process models both syntactically and semantically comparable. In this article, we present a conceptual model for this new approach to process benchmarking, a framework, as well as a software prototype for analyzing and comparing individual process models. We provide an overview of our multi-method evaluation methodology and delineate the technical, conceptual, and economic evaluation perspectives with their respective outcomes. This analysis allowed us to determine whether our approach is generally suitable for generating novel and useful information on different process models that describe the same problem domain.