This paper focuses on the systems-of-innovation (SI) approach and its policy implications. It introduces the topic by briefly reviewing the emergence, development, and diffusion of this approach and, then, spells out the constituents, activities, and boundaries of the SI approach. On this basis, the paper discusses the reasons for public policy intervention in the innovation process and the division of labour between private and public actors in carrying out SI activities. In this context, the paper argues that the notion of optimality is irrelevant in an innovation system context and it discusses the importance of uncertainty for innovation policy and the inevitable selectivity of such policy. A brief survey of the strengths and weaknesses of the SI approach rounds off the paper and points to avenues for future research.