In many countries the need for education systems and schools to improve and innovate has become central to the education policy of governments. School inspections are expected to play an important role in promoting such continuous improvement and to help schools and education systems more generally to consider the need for change and improvement. This article aims to enhance our understanding of the connections between school inspections and their impact on school improvement, using a longitudinal survey of principals and teachers in primary and secondary education. Random effects models and a longitudinal path model suggest that school inspections in particular have an impact on principals, but less so on teachers. The results indicate that the actual impact on improved school and teaching conditions, and ultimately student achievement, is limited. Schools in different inspection categories report different mechanisms of potential impact; the lack of any correlation between accepting feedback, setting expectations and stakeholder sensitivity and improvement actions in schools suggests that the impact of school inspections is not a linear process, but operates through diffuse and cyclical processes of change.