This paper considers the distinction between two sets of factors that affect fertility -- community supply of family planning services, and constraints on parent demand for births -- and assesses how it guides the empirical design of evaluation of family planning programs. Evaluating the effect of program activities on fertility requires that certain determinants of parent demand for births be held constant. Increases in supply are associated with reductions in "cost" of birth control. The supply-demand framework can clarify the segments of the population for which the program is likely to be more or less cost-effective, and how the effectiveness of various program activities may change with the scale of the program and interact with other program elements and population characteristics. Evidence from Thailand demonstrates that family planning expenditures are subject to diminishing returns and that marginal returns differ across elements of the program and segments of the population. To improve the efficiency and equity of family planning programs, policymakers need estimates of the cost-effectiveness of family planning programs at the "margin" of various "mixes" of program activities, within distinguishable segments of the reproductive-aged population.