CONSUMER PRODUCT TESTING 771 Then, as mentioned earlier, panelists will be selected according to specific characteristics. In testing a new instant coffee, care was taken to include families who used only instant coffee, families who used only brewed coffee, and those who used both. The "instant-only" users showed they had become accommodated to the flavor of the then insipid but not unpleasant instant coffees, while persons who were familiar with brewed coffee recognized the virtues of the flavor characteristics of the new product. This illustrates selection of panelists according to type of product they use. Frequency of product use can also be a criterion of selection. In another test it was found that frequent users of the current product were strongly against a variant of the product, while occasional users were delighted with the variant. This could mean two products with an over-all increase in product use. More attention is being paid to teen-age products, and the authors have a source of panelists, who have been found to be as discriminating as their parents and often more communicative. Size of panels may be as small as 25 persons or families and as large as 100 persons or families. Since the size of the A.D.L. panels is not large, these tests are called pilot tests. The roster from which panelists are usually selected consists of the families of A.D.L. Cambridge em- ployees who have expressed their desire to participate in product tests. Most of them have lived in the area more than five years. Naturally, then, the first question to face is: Is this a regional panel? The answer generally is "No." If the type of product is used nationally--e.g., mouthwash--then its use properties can be evaluated by a discriminating-communicating panel anywhere. If the product is designed for a specific region in the U.S., this panel can indicate if the product has the qualities it is supposed to have. They may not particularly like the product, but they can isolate its elements. The second question to face is: How do the results obtained from discriminating-communicating panels compare with those from a national and not necessarily totally discriminating panel? To probe this question, a paired comparison of two toothpastes was put through the A.D.L. consumer toothpaste panel and a national panel twice as large. The preference trends and reasons for preference were the same from both panels, but in analyzing the questionnaires, the information from the A.D.L. panel was found to be more definitive. This result was not unexpected.

772 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS But there is an even more important rationale to the use of dis- criminating panels. Because they do discriminate, they provide the basis for a rigorous test. If the product is acceptable to persons who can discriminate, its properties should also be acceptable to persons who cannot or do not discriminate. Prerequisites To plan a proper test--that is, to select the consumer panel, to decide how often the product should be used and for how long, to choose the method of obtaining responses from the panel, to develop instructions for the consumers, and to anticipate the terminology they might use the designers of these tests charge themselves with two responsibilities: to be sure of the purpose of the test and to be sure of the test product's properties. Every test is specifically designed around the product and the test objective. If the product is a model of a concept, the product develop- ment group may wish to know if it matches the concept and, if not, what modifications are needed. If it is to be a new product, is it in its present status acceptable as a whole does it have more positive than negative features ? If a variant of a currently marketed product-type, how do its attributes compare with those of the marketed product is there a posi- tive that could be exploited in advertising ? If it is an improved version of an existing brand, do consumers see the difference and do they con- sider it an improvement do they see it as a major or minor improve- ment ? Once the test objective is defined, extensive effort will be made to define the product. First, if such technical information is not already available, technical analyses or examinations of the product will be carried out. Since the authors' consumer product testing mainly con- cerns foods and other flavorful products, this technical examination, while including observations on pH, color, and viscosity, will mainly be a Flavor Profile. Flavor Profiles are produced by experienced panels, who work under controlled conditions and use standardized techniques for smelling and testing. The Flavor Profile is a tabular record of the product's sensory (aroma and flavor) properties. Having completed the Flavor Profile, the panel members will work to produce a use profile. For a soft drink, they would drink (as opposed to taste) a bottleful of the beverage in much the same way that con- sumers will--for example, gulp it down, pour it over ice, let it warm up in the glass, and drink it from the bottle. The value of a use profile is