The purpose of this paper is to examine leadership practices in school systems that are implementing data-driven decision-making employing the theory of distributed leadership. With the advent of No Child Left Behind Act of 2001 (NCLB) in the US, educational leaders are now required to analyse, interpret and use data to make informed decisions in all areas of education, ranging from professional development to student learning. The emphasis on data-driven decision-making practices to bring about improved student outcomes is relatively a new feature of the education reform landscape and thus requires educators to learn and develop new competences. Leadership is one crucial bridge that can support and direct these new learning efforts. Using qualitative data from a case study of four urban school systems, the authors' findings indicate that: (1) leaders at all levels co-constructed the vision and implementation of productive data-driven decision-making by creating an ethos of learning and continuous improvement rather than one of blame; (2) in order to give data relevance, leaders also distributed decision-making authority in a manner that empowered different staff members to utilise their expertise; and (3) the school systems directed their resources on building human and social capacity mainly by focusing on modelling and knowledge brokering amongst their staff. The paper concludes with a discussion of research and policy implications based on the findings. (Contains 1 table.)