The need to stem the rising tide of non-communicable diseases (NCDs) including diabetes has been recognised at the highest levels through the UN political declaration. Diabetes care services are largely unavailable in the primary care setting in most developing countries and where available the services are unstructured, with poor record keeping, stock outs and frequent disruption in supplies. With no systematic monitoring of care and programme implementation, treatment outcomes are poor and are consequently associated with a high economic burden. Systematic evaluation of programmes through cohort monitoring has been shown to be effective in large-scale interventions for two chronic infectious diseases-tuberculosis and HIV/AIDS. Can the same simple tool of cohort monitoring be applied to improve diabetes care delivery in the developing world? Pilot projects show it is possible, but scale up and expansion would require investment in information technology. In a scenario where systems for NCD are just beginning to be set up, it makes sense to learn from and build further on the initial pilot programmes.