As revenue management research progresses, simplifying assumptions are removed from the underlying mathematical models. In consequence, these models grow, leading to an increase in complexity that may affect both the performance of automated systems and revenue management analysts. In this article, we demonstrate how both hierarchical and dynamic complexity may increase as revenue management models become more sophisticated. For this purpose, we introduce an example of dynamic complexity based on forecast parameterization based on a simulation study. On the basis of a data analysis created in cooperation with Deutsche Lufthansa, we demonstrate the dependence of state-of-the-art revenue management systems on analyst input. We argue that increased complexity endangers the performance of both analysts and automated systems if it is not deliberately managed. Finally, we discuss five possible strategies for responding to increasing complexity: Ignorance, full automation, visualization, result simulation and input transformation. We describe the possible implementation of each strategy and list the opportunities and challenges that each of these response entails for revenue management.