This article is based on the ten-year experience of an operations research project in Bangladesh. It assesses how, and under what circumstances, research-based advice and results of pilot projects contribute to change in large-scale public programs. It discusses project research on issues facing the national family planning program: recruitment and training of field-workers; delivery of injectable contraceptives; management information; field-workers' use of service registers; field supervision; satellite clinics; and contraceptive user fees. These issues are used to illustrate the advantages and disadvantages of a long-term institutionalized project, and to describe the diversity of means for communication with policymakers. The analysis shows that research, policy decision, and implementation can occur in any sequence. Policy advice that disrupts long-standing power relationships and organizational culture takes a great deal of effort to implement. Operations research can produce useful changes in organizational behavior, even when large-scale problems remain.