When countries have tried to launch ambitious large programs immediately at full scale, these programs typically have existed more on paper than in the field. Large programs that cannot deliver promised services lose credibility among intended clients and family planning personnel (126, 210). In contrast, in a small program satisfied users spread the word about family planning and encourage wider use, helping the program to grow (44 ). Also, a program that delivers on its promises wins advocates for family planning, who can help obtain the resources needed to expand services. Successful small programs typically attract the interest, funding, and donor assistance that they need in order to expand (93 ). As they grow, they apply the lessons learned from experience to become successful large programs (52 ).

In the 1960s and 1970s small studies and field experiments paved the way for the successful larger national programs in Indonesia, South Korea, and Thailand (44 ). Probably the best reported example of starting small and building on lessons learned, however, is the Family Planning Services Project in Matlab, Bangladesh, which began in 1977.

The Matlab project. The Matlab experiment and its extension to the national level have demonstrated that, by making a choice of contraceptives accessible to most people and by emphasizing direct personal communication between service providers and people in their communities, programs can raise contraceptive prevalence and reduce fertility rates among poor, rural people with little formal education. In this pilot project the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) introduced special family planning services, along with maternal and child health services, to residents of some villages in the Matlab subdistrict, a rural area typical of much of Bangladesh. The project was designed so that researchers could compare the experimental villages with an otherwise similar set of villages that received only the usual government family planning services. After the initial design proved to have little impact, the project in 1977 began offering intensive services that corrected for deficiencies in the initial approach.

In the experimental villages, young, married, literate, high-status women, trained to serve as field workers, offered a range of contraceptive choices to women and provided information, counseling, and regular follow-up visits (124). Within 18 months of offering services in the experimental villages, contraceptive use rose from less than 7% to more than 30%, while in the comparison villages contraceptive use rose hardly at all (134). There were doubts that the resource-intensive Matlab approach could be incorporated successfully into the national family planning program because of the large costs involved (110 , 123). When the ICDDR,B Extension Project tested the Matlab approach in other subdistricts as part of the regular government family planning program, however, contraceptive prevalence also rose dramatically in those areas (134). Since then, Bangladesh has incorporated lessons learned from the Matlab Extension Project into its national family planning program (23).

More recently, in Bangladesh, the Jiggasha approach to providing family planning has been developed using extensive baseline research, communication studies, and other quantitative and qualitative measurement. Jiggasha began in 1989 in Trishal, a typical rural subdistrict. In this approach, government family planning field workers and rural community members use traditional communication networks and techniques of community participation to provide modern contraceptive information and supplies. Survey research has demonstrated that the program has increased awareness and use of family planning in Trishal while enabling the field worker to use her time much more efficiently (112, 206).

The Bangladesh family planning experience suggests that, "without any form of governmental duress, the poorest societies in the world, and ones with life expectancies no higher than 50 years, can achieve contraceptive prevalence rates above 50% and total fertility rates below four" (22).