The general sense of the group was that the possible topics for discussion were extremely wide ranging since the industrial market for Market Research was being defined as the non-domestic, or mass commercial market. In this context, many topics discussed were relevant to only a limited part of this wide market and care had to be taken therefore that discussions were as specific as possible .
In this paper I want to discuss three methods of industrial sampling which can be used in some of the situations in which difficulties can arise. The conclusion I should like to draw from these examples of alternative sampling methods for industrial research is that the classical approach is not necessarily the only, or the best, way of solving a particular problem. In such cases it is very often possible to devise a different approach which overcome the need for a relevant sampling frame or whatever other component of the classical approach is not available.
The purpose in presenting this paper is to show how a pioneer method of sampling was carried out, how it was validated and the implications of this validation on tourism studies. Before going into the details of the method it might be of help if some of the background to this particular study was sketched out. The tourist resort concerned was Guernsey and at this juncture it would be appropriate to extend grateful acknowledgements to The States of Guernsey Tourist Committee and to The British Tourist Authority for their permission to give this paper and thanks for their help in preparing it.
A frequent problem in marketing research is how to collect information on the consumers of a product, which has a small market share. If a product is used by 5 out of 100 housewives one will need 4.000 interviews to reach a sample of 200 users. Many products have a market share which is still smaller which makes it even more economically unrealistic to do research by traditional sampling methods. This problem is especially acute when a new product is launched and information is needed on the consumer reactions. The problem of reaching such small segments of the market for research purposes can be solved by abandoning the traditional sampling methods and using instead the flow of distribution. This is done by sampling products instead of respondents and letting the products localise the consumers. The method is to select a systematic random sample of products before they are distributed and attach questionnaires to the products. The questionnaires are to be answered by the consumers and sent to the research institute. We call this type of research: product-sampling or distribution-radar.
The applications of mathematics in the development of sampling methods are generally well known and appreciated. However, mathematics is now engaged in other problem-solving roles which will be just as important to marketers in the future as sampling is and was in the past. Already, powerful analytic devices such as Markov Chains, Mathematical Models, Simulation Techniques, Monte Carlo Methods, and other techniques have been applied to marketing decision making. It is the purpose of this series of bulletins to present at occasional intervals non-mathematical expositions of some of these applications. The first Bulletin will be concerned with a discussion of Markov Chains.