Despite the fact that our knowledge on how policies are designed has substantially improved during the last two decades, prescriptive literature on policy formulation remains largely disconnected from these new findings. The article examines five major assumptions upon which policy formulation is still predominantly based: (a) there is one way policies are and should be formulated; (b) effective formulation of policies is more about the right application of methods than of the substance of a policy domain; (c) policy formulation is about choosing from mutually exclusive alternatives; (d) problem definition has priority over problem solution; (e) there is a clear distinction between policy formulation, adoption and implementation. This article shows why these assumptions are outdated and that they lead to many practical problems in the teaching of policy analysis. It is argued that policy formulation guidelines and training in policy formulation should be based on current policy design scholarship that stresses, for instance, the importance of local knowledge, deep understanding of actors' perspectives and the need to formulate policy packages. The article concludes with preliminary recommendations on how to move forward, illustrated with concrete examples from practice.