Continuous innovation of service networks is important for the success of modern service based economies. Over the last years, several analysis and design methods have emerged for engineering SOA applications. However, none of these methods can effectively cope with the increased level of complexity and dynamics that come with service network innovation. In addition, and even more problematic, existing methods fall largely short in assisting application designers and business managers in evaluating and predicting the impact of design decisions on the performance of the service-enabled applications and the end to end business processes at the network level. In this paper we use system dynamics simulation to predict the impact of design decisions for software services and human-operated services on service network performance in the automobile industry. We show that a single design decision at the software level influences 10 out of 12 performance indicators at the firm and the inter-organizational business network level.