Questioning the purely business-orientated, technology-centred use of the term ‘innovation’, research on social innovation has gained momentum in recent years. While significant progress has been made as regards the theoretical foundation of social innovation, the ‘ecosystem’ facilitating respective innovative activities remains rather vague and ambiguous. Most research informing our understanding draws on related but distinct concepts, such as Regional Innovation Systems (RIS). Following on from this debate, this article reflects on common features and differences between social innovation and other forms of innovation, and the resulting requirements for a Social Innovation Ecosystem (SIES). Drawing on data from the two European research projects, SIMPACT and SI-DRIVE, the article reflects on SIES from the perspective of RIS as an analytical framework, and strategic and management concept. It is argued that, due to a multiplicity of social innovation activities and their local embeddedness, no best solution for SIES exists. We posit that establishing a SIES necessitates (1) a mode of governance that integrates actors from civil society, and the social, economic and academic field; (2) social innovation hubs, labs and transfer centres as intermediaries that accelerate social innovation activities; and (3) the integration of different modes of innovation in transformational innovation strategies.