Refactoring a service-oriented software application is crucial to establish competitive, secure, and reliable software architectures. While migrating monoliths into (micro-)service architectures is well-studied, the evolution of a service-oriented architecture – particularly, the integration of well-established API patterns – falls short, leaving practitioners alone with assessing the impact of their design decisions when modifying existing architectures. In this article, we address this shortcoming and study the impact of implementing API refactorings in various ways from the developers’ and clients’ perspectives. We employ an existing framework for adding refactorings either internally in the refactored service(s), adjacently (in the same application but another service component), or externally and compare ad-hoc with parametric implementation. We measure the effects of using these different implementation techniques in terms of code complexity and its performance. In this way, our work provides the grounds to guide software architects in establishing well-designed and competitive service-oriented architectures upon evolution.