CONSUMER TESTING A GUIDE TO PRODUCT DEVELOPMENT 179 efficiency or time. Triangle tests are helpful in determining the magnitude of differences. Having narrowed down the field there are apt to be still other questions concerned with the viscosity, as of a toothpaste the type and level of perfume or flavor in almost all products the oil content the amount of humectant, as in shaving cream the texture as in lotions or creams etc. These factors and many others can be and probably often are decided on the basis of personal judgment. It should not be suggested that this is entirely bad in itself and certainly some screening can be done in this way. However, there are usually key points where the answer can best be ob- tained from the consumer. Since we are talking about the product definition phase, we may need only a preliminary consumer feeling for those characteristics which are still undetermined. Here some sort of a location test may be adequate. This type of test has been described in the literature particularly for the food industry. General Foods has used extensively and has described (6) a testing procedure called a store test. This test derives its name from its location in a large department store where customers in the store are in- vited to sample products in question, usually in pairs, and to express a preference for one over the other. This store test technique has advantages for such variables as perfume or flavor type and level, consistency or viscosity and appearance. One of its principal values is that the technologist can usually provide samples adequate for testing on the basis of laboratory mixes without having to develop final processing, packaging, etc. Flavors, for example, can be tested by adding to prepared foods without necessity of developing fixations to stabilize the flavors for dry storage. The results are quickly obtained and can be used almost as day-to-day guides for laboratory work. These tests are limited, however, to the type of product and product usage which can be carried out in public without embarrassment to the participant in the test. The population sample is not random, being limited to the clien- tele of the store, and more particularly, to those customers who are in or near the area where the test is being conducted. Children are poorly rep- resented, especially during the school year, and working wives can be reached only at limited periods, such as Saturdays or evenings. Nevertheless, this type of test has proven by experience to provide re- markably valid results on many types of food products. Certainly these findings have been more helpful than could be obtained from small captive panels such as groups of fellow workers who are apt to be biased by fore- knowledge of the work or by their total experience with the technology involved in the field of development. That is, captive panels are apt to become "expert" in a very short span of experience. Another type of test which is helpful during this phase of product devel-

180 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS opment is a home use test. V:irtually all product characteristics which can be recognized by consumers •an be tested in the home. This test permits the participant to try out the products under normal use conditions and in complete privacy if this is desirable. The consumer can use the product several times and has a chance to become familiar with it. The technol- ogist can introduce variables at will, obtain an expression of consumer preference for any number of variables which he wishes to evaluate. The home use test is generally more expensive and time consuming than the store test described above. The product must have been presented in a form which is a close approximation of finished product development, so that the consumer will not be dealing with an obviously crude product form, shape or appearance. Stability to long storage and commercial pack- aging is of course not required for valid test results. Having selected the general characteristics of a product by the above rather preliminary consumer tests, the technologist, combining this infor- mation with the knowledge he has also been obtaining about costs, storage stability, patent infringements or patent protection possibilities, is now in a position to show his management an approximate mock-up of the product to be developed and agree upon a quite specific definition of the desired attributes of the eventual finished product. LaBORaTORY DEVELOP•4ENT Knowing the nature of his problems in some detail, the technologist can now proceed with laboratory work to settle specifically upon the details of ingredients, quantities and performance characteristics of the product which has been defined above. In the definition he has usually come close enough to a finished product to be able to describe rather accurately the attributes which will meet the requirements of his merchandising people or management. At the end of this phase of development, he should know precisely what ingredients and laboratory manipulations are required and should be able to provide small quantities of samples which both he and merchandising personnel agree is the product they would like to produce and sell. Here again, consumer testing is helpful in refining all characteristics of the product to achieve the maximum practical consumer acceptance. At this stage it may very well be imperative to check out product character- istics against competitive products on the market. Again, the technologist is faced with choices of ingredients or procedures where the only decision that counts is the one which can be given by the typical consumer. Loca- tion tests using stores, or mobile sampling units which can go wherever people are apt to congregate are also useful here for the same reasons noted above. The technologist will also want to test his most advanced samples in the home. At this stage, in addition to paired comparisons of variables