Organizations that are part of collaborative service networks need to handle increasing amounts of data in their information systems to construct complex customer-oriented solutions from dynamically selected service elements. This brings numerous challenges in today’s highly competitive markets, where companies need to provide customers with services that have a high level of quality, in a time and cost effective manner. Having prior knowledge of the performance associated with specific choreographies of services allows companies to provide customers with services tailored to their specific requests, while maintaining a high quality of the services. This paper presents an approach for service selection and ranking considering customers’ requirements, knowledge of the historic behavior of services, business process constraints and the characteristics of the execution environment. Following the specifications of the proposed framework, a prototype has been implemented targeting the automotive sector. Tests of this prototype illustrate the usability and validity of the proposed approach.