This paper argues that there are many challenges to designing and implementing water and sanitation interventions that actually deliver economic benefits to the households in developing countries. Perhaps most critical to successful water and sanitation investments is to discover and implement forms of service and payment mechanisms that will render the improvements worthwhile for those who must pay for them. In this paper, we argue that, in many cases, the conventional network technologies of water supply and sanitation will fail this test, and that poor households need alternative, non-network technologies. However, it will not necessarily be the case that specific non-network improved water supply and/or sanitation technologies will always be seen as worthwhile by those who must pay for them. We argue that there is no easy panacea to resolve this situation. For any intervention, the outcome is likely to be context-dependent. An intervention that works well in one locality may fail miserably in another. For any given technology, the outcome will depend on economic and social conditions, including how it is implemented, by whom, and often on the extent to which complementary behavioral, institutional and organizational changes also occur. For this reason, we warn against excessive generalization: one cannot, in our view, say that one intervention yields a rate of return of x% while another yields a return of y%, because the economic returns are likely to vary with local circumstances. More important is to identify the circumstances under which an intervention is more or less likely to succeed. Also for this reason, when we analyze a few selected water and sanitation interventions, we employ a probabilistic rather than a deterministic analysis to emphasize that real world outcomes are likely to vary substantially.