While economists may assume the consumer is a rational animal, most companies can't afford that luxury. So several major corporations are now experimenting with a method that measures a consumer's sensory responses to product ingredients so finely that a computer can print out the recipe for the optimal-cost product with the maximum consumer appeal.

Although companies in the food, beverage and cosmetics industries are most interested in this technique, the underlying concept may offer a universe of potential applications ranging from textiles to typewriters to political candidates.

"In terms of developing a product," says Barry Jacobs, manager of marketing research and analysis for Pepsico, makers of Pepsi Cola, "I know of nothing that helps as much as this does."

Pepsi already has developed a sensory-optimized carbonated soft drink that it is capable of bringing to the marketplace, and it plans to explore further the possibilities of optimizing product costs to the consumer's sensory evaluations.

Pepsico is hardly alone in its efforts. Phillip-Morris, Avon, Campbell Soup, and General Foods are all evaluating the impact of consumer-based computer-optimization on their own market research and product development strategies. The November Psycology Today reports that Ragu Foods, a division of Chesebrough-Pond's in-Connecticut, is test-marketing a computer-optimized Italian table sauce in Boston and Milwaukee. A number of other firms are test-marketing more quietly or covertly.

The key to this new trend in market research and product development is psychophysics -- a discipline that seeks to measure how humans respond to sensory data.How crispy is crisp? And how salty is salty? These are questions the psychophysicist is trying to answer.

A Harvard psychologist, Dr. S. S. Stevens, pioneered psychophysics in the 1950s by introducing new ways of measuring people's sensory perceptions. Stevens' major contribution was the technique of "magnitude estimation" that permits a person to devise his own scale of measuring the intensity of a stimulus.

Two Harvard graduates who worked in Stevens' lab have taken the concept, integrated it with the cost-analysis requirements of product development and have developed a system designed to correlate consumer preferences, ingredients and cost.

"It's a straight-out business system," says Howard Moskowitz, president of MPI Sensory Testing in New York, "to determine the pulse of the people's taste.

"We determine what blend of ingredients are attractive to the customer and, for the first time, the manufacturer can get a real feel of how much profit margin is acceptable if he wants maximum attractiveness."

Dr. Moskowitz and Dr. David Fishken, a co-director of MPI Sensory Testing, have been using this system for the past three years. Most of their clients prefer secrecy when developing products through MPI's system, but many are enthusiastic about it.

"It's a quantum step in the area," asserts Dr. Craig Warren, a research group leader at International Flavors and Fragrances of New York, the world's largest supplier of flavors and fragrances and a client of MPI's "After you get the data, you actually know what to do with it."

IFF mainly is testing the system with its cosmetics division.

The success of MPI's system is dependent upon how well it can measure the consumer's response to specific product characteristics.

For example, manufacturers of potato chips know that consumers mostly are concerned with three factors: saltiness, crispness and texture. For an MPI test, the manufacturer provides MPI with potato chips of varying degrees of saltiness, crispness and texture. MPI then selects a panel of consumers. Each consumer tastes one or several chips and uses the qualities of those as the fixed points around which he evaluates the other chips. He ranks the other chips as saltier, or less crispy or too rough based on a numerical scale of his own devising. After tasting all the chips, the panelist is then asked to give the ideal rating for the ideal potato chip based on his numerical evaluations of the other potato chips.

The results from all the panelists are then statistically adjusted to compensate for variance and then fed into a computer which prints out the ideal blend of saltiness, crispness and texture for a potato chip.

The next step in the system is to determine the cost limits for the ultimate potato chip and program those limits into the computer as it redesigns the ultimate potato chip to fall within cost specifications. Thus, a product can be designed with cost as the main priority or with consumer appeal as the main priority depending upon the marketing goals of the corporation. Also, the procedure can reduce the lead time of research necessary to bring a product to market.