Now that you’ve set up your tracking system, let’s look at why and what you should track.
Testing is the “’life blood” of direct response marketing. Testing is critical for all DR Lead Generation customer acquisition channels. If you don’t test, how do you know which are your best offer(s)? Your best list? Or your best creative package?

If you continuously use the same images, the same offers in the same vehicles – consumers will eventually see you in the same light as the furniture store that’s been going out of business for the past 7 years. There will be no urgency to “act now” in your message. “Sally, we don’t have to call for

“Nth” select testing is the best methodology. By “Nth,” I mean every other piece is the test, every 3rd, every 4th, every 5th, etc. You maintain your testing selection throughout the universe you’re using. If the “Nth” select is not practical in the medium you’re using, while not totally scientifically perfect, you can also use an even/odd zip-code test, different radio/TV stations within the same market can yield valid testing results.

What do I mean by “element?” 1) Graphic design, 2) Offer, 3) Target audience, list, magazine, broadcast station – these make up the Marketing “Triangle.” If you’re testing the graphic design – the offer and target audience have to remain the same. If you’re testing different offers – the graphic design and target audience have to stay constant. You get the idea.

Next, you have to have statistically significant response numbers to be valid. Example: you tested and have twice as many responses from effort “B” as you do effort “A.” Great, you found a winner. Or did you? You got 2 responses from “B” and only 1 from “A.” Theirs is statistically insignificant responses to make any decisions.

The following are annualized results on various tests conducted over a period of a couple of months each. Of course this was over an entire region of the country encompassing 5-6 states. But I believe you can reduce the revenue gains down to a local level and still clearly see the potential benefits of conducting tests for your company.

NOTE: The lowest CPL or best Response % doesn’t always reflect the best ROI

NOTE: For those “math-letes” who are 110% scientifically inclined — if you have a local university handy, you might ask a professor in the math department to write you out a simple spreadsheet with the Chi-squared formulas built in. If you enter your known or “control” response percent and the “confidence level” you want for your test, the macros will tell you how many pieces you have to mail to attain the test results. The higher the confidence level the larger the test and control quantities will be. Depending on how radical your test, this could be a big and expensive gamble. For an initial test, I’d generally recommend a confidence level of 65 – 70%. If the test “beats” your control, then next test should be larger, maybe 75-80%. If the test still wins, I’d consider you either have a new “control,” or you have a test that equals your current control so you can alternate these packages and neither will “fatigue” as quickly.

It’s 2009, you cannot expect different results if you continue doing the same things. Test and see if you can beat your current “control” creative, offer or find a new audience for your products and services.