As part of a randomized controlled trial conducted in Myanmar between 2016 and 2019, we explore the performance of a maternal cash transfer program across villages assigned to different models of delivery (by government health workers vs loan agents of a non-governmental organization) and identify key factors of success. Measures include enrolment inclusion and exclusion errors, failures in payment delivery to enrolled beneficiaries (whether beneficiaries received any transfer, fraction of benefits received and whether there were delays and underpayment of benefit amounts) and whether beneficiaries remained in the program beyond eligibility. We find that women in villages where government health workers delivered cash transfers received on average two additional monthly transfers, were 19.7% more likely to receive payments on time and in-full and were 14.6% less likely to stay in the program beyond eligibility. With respect to the primary health objective of the program-child nutrition-we find that children whose mother received cash by government health workers were less likely to be chronically malnourished compared to those whose mother received cash by loan agents. Overall, the delivery of cash transfers to mothers of young children by government health workers outperforms the delivery by loan agents in rural Myanmar. Qualitative evidence suggests two key factors of success: (1) trusted presence and past interactions with targeted beneficiaries and complementarities between government health workers' expertise and the program; and (2) performance incentives based on specific health objectives along with top-down monitoring. We cannot exclude that other incentives or intrinsic motivation also played a role.