Sponsor (July-Dec 1951)

Going bevond these generalizations, these testers were after a cross-section of typical American markets. Cincinnati (455.000 pop. I was chosen as a "miniature metropolitan" market, complete with commuters and a substantial number of white-collar workers. Dayton (210.000 pop.) and Peoria. 111. i |l)5.0(l(l pop.) were picked to represent the medium-sized, strong urban market self-contained, with their own diversified industries. Terre Haute. Ind. 1 65.000 pop. I rounded out the quartet of test cities as a typical rural center ( researchers wanted to sample the reactions of people living on surlounding farmland). The lest media to be used throw a further complication into choice of markets. Since researchers follow the dictum -never do anything in a test market that can't be duplicated nationally— each market chosen must have both newspapers and radio stations to be representative. Selecting markets from a media viewpoint can be highly complex. Here are the points to watch in buying radio or TV time for tests. according to Carl H. Henrikson. Jr.. former research director of J. M. Mathes & Co. and now v. p. and general manager of S-D Surveys. Inc. ( New York market research specialists I . 1 . \ competitor should not have a program in one of your test markets and not in the others; such unequal competitor influence represents an uneuntrolled variable. 2. All the stations used should have about the same "listener lovalty' in *'He who works with his hands is a laborer; with his hands and his head, a craftsman ; with his hands, his head, and his heart, an artist. He who works with his hands, his head, his heart, and his feet ... is a salesman." LOIS INGALLS, Sales Promotion Manager, I. Miller & Sons, IS. Y. their respective markets. Matching a schedule on a powerful network or independent station against a similar schedule on a small secondary station isn't scientifically fair. 3. Take into account nearln largecoverage stations which exert a strong effect on listening in the market city or cities. Such outside "overlap" should be the same for all test markets to avoid bias. 4. Match announcements against announcements, programs vs. programs. / eople ue// Setter than 1 after •» • • and You Can Cover Central New York with ONE Radio Station ACUSE 570 KC NBC Affiliate. WSYR-AM-FM-TV— the Only COMPLETE Broadcast Organization in Central New York Headley-Reed, National Representatives time periods against the same time periods— unless type of air advertising to be used is what you're testing. And if the most effective way to use a medium is being tested, then all other factors must be the same in all test cities. 3. Running the test. If planning and market selection are careful 1\ thought out by expert market researchers, the test itself should come off with comparative ease. How long does a test normally take? Between three and six months is the usual run — a trial period which agency and independent researchers universally consider far too short. Nine months to a year's test would suit them better. Main reason for wanting a longer test is to allow time for re-purchase sales. Initial sales are apt to be onetime flashes which give a false glow. What testers want to know is what the long-term picture will look like — will the product hold up? They can't always wait for a completely definitive answer, since each extra month adds expense, gives competitors more time to work up counter-moves. One device which some researchers use to speed up tests is to "over-spend' in the test markets. This means telescoping the advertising impact into a shorter space of time by running bigger schedules than the budget could afford on a national scale. The practice isn't universal, since it does some violence to the principle of not doing anything in a test market that can't be duplicated nationally. A safer way to save money is by working out cooperative arrangements with local dealers who kick in part of the media costs in return for the extra advertising their market is getting. One important thing to remember is that market tests must be kept top secret. The reason is obvious. For one thing, there's no point in spending a small fortune to find out something, only to tip off a competitor for free. Besides which it's not unknown for a well-heeled rival to throw an intensive couponing drive right into a competitor's test market. Such "coincidences" occur often enough by accident without encouraging such a deliberate upset. One of the classic examples of what can happen when a competitor learns of your test is the story of "Surf" I Lever Bros, i and "Tide" (Procter & Gamble i. In L949, Lever planned to buck the head-tart which the P & < 8-4 SPONSOR