Many policies and programs based on informational interventions hinge upon the assumption that providing citizens with information can help improve the quality of public services, or help citizens cope with poor services. We present a causal framework that can be used to identify leaks and blockages in the information production and dissemination process in such programs. We conceptualize the “information pipeline” as a series of connected nodes, each of which constitutes a possible point of blockage. We apply the framework to a field-experimental evaluation of a program that provided households in Bangalore, India, with advance notification of intermittently provided piped water. Our study detected no impacts on household wait times for water or on how citizens viewed the state, but found that notifications reduced stress. Our framework reveals that, in our case, noncompliance among human intermediaries and asymmetric gender relations contributed in large part to these null-to-modest results. Diagnostic frameworks like this should be used more extensively in development research to better understand the mechanisms responsible for program success and failure, to identify subgroups that actually received the intended treatment, and to identify potential leaks and blockages when replicating existing programs in new settings.