William J. Infosino’s scientific contributions

To be successful in the telecommunications equipment and services market, we must satisfy the diverse needs and requirements of customers. One critical challenge is to direct resources (R&D, manufacturing, and marketing) efficiently across product and service lines. This paper describes an approach to market analysis and product design that applies to large-business customers who need a variety of telecommunications products, such as business services, and complex switching and transmission systems. We cannot use traditional statistical methods of market research on business telecommunications products, because there are often only a few customers. The methodology presented in this paper helps us determine which new products and features to develop and is specifically aimed at large, complex industrial products. We have applied it successfully to products in switching, transmission, and installation services. Using this methodology, AT&T can design telecommunications products that best meet customer needs and that realize substantial savings in development costs.

This paper compares consumer likelihood of purchase ratings for a proposed new product to their actual purchase behavior after the product was introduced. The ratings were obtained from a mail survey a few weeks before the product was introduced. The analysis leads to a model for forecasting new product sales. The model is supported by both empirical evidence and a reasonable theoretical foundation. In addition to calibrating the relationship between questionnaire ratings and actual purchases, the empirical evidence demonstrates the significant effect of alternative promotion/distribution vehicles on new product sales. The model uses questionnaire data to extend the results of a limited market trial to alternative target markets, product specifications, and prices.

The demand for local telephone calls varies among households. This paper develops models, based on limited data from California and Cincinnati, which predict the demand for local calls from household characteristics. The models enable one to stratify a metropolitan area into regions of expected high, low, and medium telephone demand and thus provide a mechanism for efficiently estimating the demand in a metropolitan area from the demand of a sample of telephone customers in the area. Although the data and models are too limited to establish definite causal relationships, the models suggest that the demand for local calls might be related to the number of people in the household and the age and sex of the household head. Furthermore, while there is some ambiguity between the California and Cincinnati results, there is also the suggestion that local call demand might be related to income, the race of the household head, and the telephone density in the wire center.

Author's reply to comments (Morrison, D. G. 1986. Commentary. 385–386 and Pratt, R. W., Jr. 1986. Commentary. 387–388.) about his paper (Infosino, W. J. 1986. Forecasting New Product Sales from Likelihood of Purchase Ratings. 372–384.).