The national education reform agenda has rapidly expanded to include attention to continuous improvement research in education. The purpose of this analysis is to propose a new approach to "developmental evaluation" aimed at building a foundation for continuous improvement in large-scale school improvement networks, on the argument that doing so is essential to producing the intellectual capital needed to replicate effective practices and desired outcomes throughout these networks. We begin by developing a "rationale" for developmental evaluation, both to illuminate the need and to discuss its coordination with other forms of evaluation. We continue by proposing a "logic" of developmental evaluation to support analyzing networks as learning systems. We then use that logic to structure a "framework" for developmental evaluation to support evaluators, network executives, and other stakeholders in analyzing and strengthening the foundation for continuous improvement in a given network. Our analysis suggests that building a foundation for continuous improvement among a large number of networks is an educational reform agenda unto itself, one that must be supported and sustained if these networks are to succeed at the level expected under current accountability regimes.