Service-oriented architectures are widely discussed as a design principle for application and enterprise architectures. Nevertheless, an adequate granularity of services has not yet been researched sufficiently from an economic perspective. The finer the granularity to realize the functions of a process, the higher the number of services is, and the more effort has to be directed towards composing them. In contrast, very coarse grained services bear the disadvantages of higher implementation costs and lower reuse potential (e.g., in different processes). The aim of the decision model proposed in this paper is to determine an adequate granularity of services from an economical perspective. Thus, degrees of freedom, which often exist for the choice of granularity after a domain analysis, can be leveraged to realize a cost-efficient solution. We illustrate the applicability and practical benefits of the decision model with an example from the context of a financial services provider.