29 Don't buy the frequency line. It's the myth that you need to run at least three times before the ad will start to pay out. The best results almost always are with the first exposure of the ad in a new medium. Thereafter, let the extent to which the publication's audience changes and how strongly the medium pulled on the first test be your guide to determining how often to run. Experiment and test. (Ask for a frequency discount while you're at it.)
30 Unless an offer depends heavily on color visual, it's best to start out with a black-and-white ad. Color can add 20 percent to 40 percent to your cost.
31 Split test. Many advertisers get tired of their own ads, or make what they believe to be improvements to the ad, only to have poor results convince them they made a mistake. To avoid risking heavy losses, test any change—whether large or small—by way of a split test.
On the other hand, any ad eventually will get tired. It pays to test new creative, changes to the offer, bonuses, pricing, illustrations, color or even different products.
32 An A/B split is the only way to eliminate outside influences and get a true reading on the performance of one ad against another. With a perfect A/B split, you provide the media with artwork (coded separately) for your control ad and for the test ad. Every other copy of the publication comes off the press with the same ad. There's no geographic bias. Both ads appear in the same issue, in the same position, surrounded by the same editorial, so any significant difference in response reasonably can be attributed to the ad itself.
33 Newspapers that offer an A/B split usually will charge less for the split fee than a magazine. However, it's best to conduct the test in your best media, regardless of cost. If you run in different media, try to schedule the same test in at least two or three places to see if results are consistent.