Al Blankenship (of Carter-Wallace) has just given an enthusiastic account of a new technique for evaluating television commercials. Your boss—the marketing research manager for a large food manufacturer—who is in the audience with you, wants you to analyze the technique carefully and make a recommendation on the use of the technique. The transcript of Blankenship’s remarks follows:

Jim stopped in my office one day early this year, bursting with an idea he had to test the effectiveness of television commercials. He told me that in fall, 1976, WCAU-TV had telecast a program which discussed the pros and cons of the proposed roofed-over sports stadium for the city. Viewers were asked to telephone their reactions to a special number to indicate whether they were in favor of or opposed to the sports palace. 

Jim and his group had been assigned the job of keeping a running total of the vote. He had become intrigued, he said, that this sort of approach might be used to measure the effectiveness of television commercials. In a balanced experiment, you could have an announcer, immediately following the test commercial, ask people to telephone in to request a sample of the product. Differences in rate of response between different commercials would measure their effectiveness.

My reaction was immediate and positive. This was really getting close to a behavioral measurement of response to advertising. But it lacked a crucial control. How could you be sure that the same number of people had been exposed to each commercial? The technique required a measurement of the size of audience exposed to each test commercial.

In this situation, I thought of C. E. Hooper, since one of Hooper’s specialties is measurement of audience size. If audience size could be built in as a control, it seemed to me that the technique was solid. I got Jim together with Bruce McEwen, Executive Vice President of Hooper. Bruce was just as excited as I had been. However, following our discussion, I began to cool off. I was afraid that the audience size measurement made the whole thing too cumbersome, and that the cumbersomeness might somehow introduce error. There was something a bit sloppy about the methodology. I did not warm up to the idea, the more I thought about it, that the viewer was going to get a free sample merely by a telephone call. This was not real life. I was afraid that the free offer bit would result in such a high level of response that it would be impossible to differentiate between commercials.

Several weeks later it hit me. What we needed was an easy method of controlling audience size and who received the special offer, and a way to make the viewer pay at least something for his [or her] product. Couponing, properly designed, could
provide the solution.

A simple method was devised, and pretesting was conducted on the couponing aspects to make sure that the price level was right. The entire test procedure required four steps: a screening telephone call, a coupon mailout, a telephone postcall, and measurement of coupon redemption.

The precall is made within a stated time period in advance of the television show that is to carry the test advertising. The respondent is asked about his or her viewing plans for the forthcoming period. The last brand purchased of each of several product groups is asked about. The product group for the brand of the test commercial is included. Immediately following screening, each person stating that he or she intended to watch the test vehicle is sent a special coupon, good for the product advertised at a special, low price.

This coupon is sent in the manufacturer’s envelope, and so far as the recipient knows, has no connection with the survey. This is not a store coupon. To be redeemed, it must be sent to the manufacturer. However, it is made as easy as possible to redeem. A postage-paid return envelope is included, and all the recipient must do is insert the proper coins in a card prepared for this purpose, which includes his name and address. 

The coupon has an expiration date of one week from date of mailing, to prevent responses that are meaningless trickling in over a long time period. The procedure makes it possible to consider coupon responses only from those who viewed the program, which is crucial.

The day following the telecast, a call is made to each person who has said that he or she expected to view the particular program. The only purpose of the call is to determine whether the person has actually viewed the particular show. No question about advertising or about brands is asked.