On the Necessity of Presenting Consumer Preferences as Prediction

真理、真諦。唐˙盧藏用˙衡嶽十八高僧序：「然而年代悠邈，故老或遺，真詮緬微，後生何述。」
On the Necessity of Presenting Consumer Preferences as Prediction

By Alfred Politz and W. Edwards Deming
(Journal of Marketing, 18(1),July, 1953, 1-5


If marketing research is to perform a useful function, it must be put finally in the form of predictions for the use of management. The form of the prediction may be, for example, the likely increases or decreases in sales that would result under certain circumstances from this or that change in quality, style, package, color, or price, of particular product, or from some particular method of advertising it. These predictions are to be used by management as aids in coming to a decision concerning the future quality, style, package, color, or price of an article, or the method of advertising it.

DECISION INVOLVE RISKS

A decision may turn out to be wrong and lead to loses. It can be wrong even when it is made with the aid of good marketing research in the form of predictions, although certainly a decision has a better chance of being the right one if it is made with the aid of good predictions. Otherwise, there would be little use for good predictions, and little use for good market research.

Likewise, a prediction may be wrong, even it is made with good data, although certainly a prediction has a better chance of being correct it is is made with the aid of good data. Otherwise, there would be little use for good predictions, and little use for good data.

Again, predictions involve risks, but the research man must nevertheless make them. It might seem smart to avoid the risk of prediction by confining the research to the mere measurement of a present situation. This viewpoint, however, is illusionary; it does not solve the problem. Until the data of of a marketing survey has been put into the form of predictions, they cannot contribute to the decion that management must make.

The principle that knowledge have temporal spread is well known in the sciences. The aim here is to translate this principle into the language of marketing research, so that marketing research may be more effective for the use of management.

Every decion entails the expectation of a specific result. Therefore, every decision, if it is rational, depend






,