This paper analyzes some of the problems of using diffusion models to formulate marketing strategies for new products. Though future work in this area appears justified, many unresolved problems limit its application. There is no theory for adoption and diffusion processes; such a theory is outlined in this paper. The present models are too restrictive and fail to include the variables that a marketing manager has at his disposal. Six models that remove some theoretical and methodological restrictions are presented. The marketing implications for the assumptions in the models are discussed, and estimation problems are considered. (Author)