Abstract Background Large system transformation in health systems is designed to improve quality, outcomes and efficiency. Using empirical data from a longitudinal study of national policy-driven transformation of maternity services in England, we explore the utility of theory-based rules regarding ‘what works’ in large system transformation. Methods A longitudinal, qualitative case study was undertaken in a large diverse urban setting involving multiple hospital trusts, local authorities and other key stakeholders. Data was gathered using interviews, focus groups, non-participant observation, and a review of key documents in three phases between 2017 and 2019. The transcripts of the individual and focus group interviews were analysed thematically, using a combined inductive and deductive approach drawing on simple rules for large system transformation derived from evidence synthesis and the findings are reported in this paper. Results Alignment of transformation work with Best et al’s rules for ‘what works’ in large system transformation varied. Interactions between the rules were identified, indicating that the drivers of large system transformation are interdependent. Key challenges included the pace and scale of change that national policy required, complexity of the existing context, a lack of statutory status for the new ‘system’ limiting system leaders’ power and authority, and concurrent implementation of a new overarching system alongside multifaceted service change. Conclusions Objectives and timescales of transformation policy and plans should be realistic, flexible, responsive to feedback, and account for context. Drivers of large system transformation appear to be interdependent and synergistic. Transformation is likely to be more challenging in recently established systems where the basis of authority is not yet clearly established.