GOLDER, PETER NEWMAN; PHD

                         UNIVERSITY OF SOUTHERN CALIFORNIA, 1994

                         BUSINESS ADMINISTRATION, MARKETING (0338)
 

                         A firm's ability to compete in new product markets is vital to its profitability and long-term survival.
                         Therefore, it is important to understand the genesis and development of these markets. Diffusion of
                         innovation theory and models based on this theory have traditionally provided this understanding in
                         marketing. A review article (Mahajan, Muller and Bass 1990) and meta-analysis (Sultan, Farley and
                         Lehmann 1990) summarize the many contributions of diffusion research, however, they only minimally
                         address some of its limitations. Closer examination of diffusion research suggests that the cumulative
                         effect of these and other limitations is that diffusion models may not capture the underlying
                         phenomenon of sales growth, they may only be useful for selective data and they may produce
                         inaccurate sales forecasts. This dissertation seeks to improve our understanding of the sales growth of
                         consumer durables through an explanatory approach. Through this approach, I uncover many novel
                         insights. First, I present generalizable findings about the period of sales before take-off. No other
                         research has even addressed this important period in the sales history of new products. Second, I
                         develop approaches of predicting the year of take-off. Third, I evaluate the primary theoretical
                         determinants of sales growth. Fourth, I develop a model that incorporates these determinants and is
                         capable of forecasting the complete sales history. The proposed model of sales growth has several
                         advantages in comparison with diffusion models. First, sales forecasts are more accurate. The
                         improvement is greater when forecasts are made three years ahead rather than just one year ahead.
                         Second, parameters are more stable when the period of analysis is varied to consider different years.
                         Third, an explanation for variation in parameters across categories and over time is presented. The
                         primary implication of these advantages is that the proposed model is more likely than diffusion models to
                         be capturing the underlying phenomenon of sales growth.
 

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