30 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS portance to select an appropriate sample of consumers and this can only be done, in the light of up-to-date and adequate data about the potential market for the new product. In the case of complete innovations no market data will, by definition, exist. In such cases it may be necessary to place the product with a sample representative of the whole population and to discover from the results to which sections of the population the product has greatest appeal. A point to be guarded against in interpreting the findings of research of this kind is the effect of novelty, which may well cause respondents to rate the product on first acquaintance rather higher than they would after prolonged use. The possible need for extended usage tests should therefore be borne in mind. At the present time, even with the use of controls, one is often at a loss to interpret results of this kind with any degree of precision. Even where novelty is not a factor, ratings on single placements are often optimistic, reflecting respondents' desire to ascribe to the test product, with which they have been provided free of charge, as good a rating as they can decently muster. The degree of 'exaggeration' in ratings obtained in this way appears to vary from product field to product field. More work needs to be done in order to throw further light on this phenomenon. Before arriving at the final product, a series of tests covering such variables as the form of the product (should it be in the form of a liquid, a powder or a creamS) and the type of packaging (assuming a powder is preferred, should it be an aerosol, puffer pack, loose powder etc.?) may be necessary. Tests involving factorial designs may also be called for to assist in choice of colour, perfume, type or strength, viscosity etc. from a number of possible variants. CONCLUSIONS Product development is a very large subject and I have only been able to touch on a few aspects of it in so far as it impinges on consumer re- search. I hope, however, that I have indicated the paramount importance of using some forIn of consumer research to check progress at each stage to ensure that sufficient is known at the outset about the market in which the test product fails and to use the greatest care in defining the objectives of the research and the universe from which the sample of informants for each test should be drawn. Even after the development and launch of a successful product, con-

PRODUCT TESTING IN THE SPHERE OF PRODUCT DEVELOPMENT 31 stant vigilance is required to ensure that every product on the market is maintained up-to-date both in quality and appeal. Competition does not remain idle and markets change under various influences, sometimes with great rapidity. There is therefore a recurring need for consumer research while a product is still on the market. (Received: 24th August 1966) REFERENCES (1) Ludford, Mrs. S. M. J. Soc. Cosmetic Chemists 18 (1967). (2) Ovens, L. H. in Middleton, A. W. (ed.) Cosmetic Science 103 (1959) Butterworths, London. (3) Pugh, C. Private communication. Introduction by the lecturer Mrs. Ludford and I are saying more or less the same things. There is room for differences of opinion in product testing, but both papers show that existing tech- niques are far from perfect, and that a lot still remains to be done to improve matters. I would like, however, to emphasise a number of points. First of all, product develop- ment does not end when the product is on the market. It begins as an idea in someone's head and ends only when the product which eventually evolved ceases to be on the market. In dealing with product testing there are two main aspects: the development of completely new products which cannot strictly be compared with anything existing at the moment, and the modification, improvement, or development of existing products. It is very important to have in mind from the beginning the test objectives and the need for the most precise definition of them. Stemming from this, the sample of people used for a test depends on the objectives. The interpretation of the results must also take account of the sample used as well as the objectives, otherwise one can be misled. What about the future? There are a number of areas in which development work is going along. Once again, quoting Mrs. Ludford {1), social class is not necessarily a useful variable it would possibly be more useful if we could categorise people by their attitudes to life, or to new products by psychological traits, by dividing people, for example, into extroverts or introverts as yet we have no simple means of cate- gorising people in this way. Existing methods are not entirely satisfactory and tend to involve a full scale survey in themselves. What we want is some means of achieving this end by asking only two or three questions. Another problem is the lack of adequate vocabulary. In all areas of sensory percep- tion people find it difficult to express themselves in precise terms. In product testing we are not very often dealing only with highly intelligent, well educated people we are dealing with a cross section of the whole population. Dr. Belson of the L.S.E., who used to be with the BBC Audience Research Department, undertook research some years ago on the level of understanding of BBC talks. His findings revealed that a high proportion of the population do not know the meaning of many seemingly simple words. We must therefore be extremely careful when asking people questions about particular attributes of products. Improvements in technique may have to be in the