The benefits of applying market demand models in engineering design have been well established. Arguably the most common approach has been to apply a random utility model, although any representation of market demand based on customers’ responses to surveys or in market situations can be used to this end. In practice, various methods have been used to collect market data through the use of surveys, typically based either on choices from among sets of alternatives or statements of willingness-to-pay. The applicability of the data collected has been limited, however, by the complexity of the relationships between market research methods and market demand models. The prospects of reusing an enormous amount of available market research data and increasing practitioners’ freedom in applying market research data to forecast market demand compel us to investigate the correspondence between market demand models. In this paper, we build on the recent work of Cook and Wissmann to demonstrate a bilateral correspondence between the popular multinomial logit model and Cook’s S-Model. Results from two independent case studies are presented showing that the utility function in a multinomial logit model may be reparameterized as a value function in the S-Model and vice-versa. The techniques demonstrated in this paper for defining direct comparisons between previously incompatible market research methods and demand models provide the practitioner with powerful new capabilities for validating market models, fusing market models to expand customer-based trade spaces, and presenting results to decision-makers in familiar formats.

This content is only available via PDF.
You do not currently have access to this content.