In democratic theory and practice, it has become a popular view that designed deliberative mini‐publics can effectively counteract failures of representative democratic institutions. But when should mini‐publics be deployed, and how should they be designed? This article develops a framework for thinking about these questions. It argues that when representative democratic institutions ensure the empowerment of inclusions, enable the formation of collective agendas and wills, and are capable of translating those agendas into binding decisions, mini‐publics should be used sparingly and as complementary initiatives; the less representative institutions are able to serve these functions, the more mini‐publics should gain independence and standing to correct these problems. The article shows how this can be operationalised in light of two key institutional design issues – coupling and authority – and discusses some empirical examples that foreground the empirical leverage offered by the suggested framework.