Protein-energy malnutrition has many diverse location-specific causes which make if difficult to solve through uniform interventions implemented through vertical programmes. This paper investigates the role of information in the planning, management and evaluation of several community nutrition programmes judged to be successful. The programmes come from Tanzania (Iringa), India (Tamil Nadu), Dominican Republic and Colombia. The review finds that the initial conceptualization and design of these programmes benefited from the results of earlier surveys and experience with similar programmes in the same or other countries. Strong capacity for operations research is important to assist with a myriad of small but important programme design details and larger mid-term re-orientations. The impact of this information depends upon the flexibility of the programme and receptivity of its management towards a learning-by-doing approach. Information for on-going programme management differs widely and conforms to the overall character of the individual programme. Thus, Iringa employs a simple system based on community growth monitoring, primarily to catalyze intervention planning and action at household and community levels. Tamil Nadu's system is far more complex and is primarily intended to assist in the delivery of centrally planned interventions. Programme evaluation benefited from information generated within the programme, but more rigorous impact evaluation requires stronger designs and more in-depth analysis than is usually provided. Overall, the review suggests that characteristics like community participation, empowerment and growth monitoring are less important in the short/medium term than strong management, a learning-by-doing approach, and the existence of some method for informing programme design and management about community needs and responses to the programme. The former characteristics may well be important for the longer-term sustainability of programmes.