CHATTERJEE, RABIKAR; PHD

                         UNIVERSITY OF PENNSYLVANIA, 1986
 
                         BUSINESS ADMINISTRATION, MARKETING (0338)
 

                         The pattern of market penetration of an innovation and the factors underlying the diffusion process have
                         been an important subject of study in marketing. This dissertation develops an analytical model of the
                         innovation diffusion process in a heterogeneous population, applicable for high-involvement
                         innovations that are durable in nature. The model employs a Bayesian decision theoretic framework to
                         predict timing of adoption at the individual level, incorporating a potential adopter's preference structure
                         and the dynamics of uncertain perceptions about the innovation. Individual level predictions are then
                         aggregated to yield the penetration curve. In particular, the analytical development of the model
                         considers heterogeneity with respect to initial perceptions, attitude toward risk, susceptibility to
                         information, and the trade-off between price and performance. The impact of variability in information
                         about the innovation's performance is captured by modeling an individual's 'path' to adoption as a
                         stochastic process. We develop a parsimonious basis for classifying potential adopters in terms of their
                         relative timing of adoption. The aggregate-level diffusion model is very flexible in its ability to
                         accommodate a wide variety of possible patterns of diffusion. The model is employed to investigate the
                         effects of heterogeneity in the population, the true performance of the innovation, and the pattern of
                         information on the rate of diffusion and the shape of the diffusion curve. A pilot study is conducted to
                         illustrate an approach to calibrating the model and to provide a preliminary test of its key implications. The
                         results provide tentative support for the predictive performance of the model, and encourage a full-scale
                         application. The model is next employed to derive normative implications for promotional policy. It is
                         shown that, in general, promotional expenditures should decline over time. The model also provides
                         guidelines for selective targetting of promotional effort. The extended model (considering several
                         non-price attributes) provides a basis for determining the attribute(s) to be emphasized in the firm's
                         communication program. Potential managerial applications of the model include (a) pre-launch prediction
                         of the penetration curve, (b) segmentation of the potential adopter population, and (c) planning
                         communication strategy.

Social Systems Simulation Group P.O. Box 6904 San Diego, CA  92166-0904 Roland Werner, Principal Phone/FAX  (619) 660-1603