If you are a small or very targeted marketer, you may sometimes struggle with putting direct marketing theory into practice. In my direct marketing career spanning more than 20 years, I have worked with sophisticated larger organizations as well as with smaller ones who had a hard time meeting direct marketing's so-called "requirements" for testing volumes.

Beyond these issues, marketers of all sizes often do not take advantage of learning from other marketers. They may even pay DM agencies to tell them things they could have copied and done for themselves. Or they could use the information to become better clients and contribute more to their company's programs.

Large, medium and small tests
When you test a new direct mail program, how many packages or catalogs do you send out? When I worked for Allstate, we tested millions of packages; at one business cataloger, we tested 25,000 to 100,000; while at one of my smaller clients, we sometimes test fewer than 5,000. All of these programs can yield valid test information. Although "statistical rules" remain the same for small as for large marketers, techniques exist that can help smaller marketers create and read small tests to their advantage.

Intimidated by Statistics?
Marketers who are new to DM testing can be put off by statistical rules of thumb — such as that you must always test at least 5,000 names or maybe that each cell needs to include 5,000 names. Those numbers are hard to reach if you are a small marketer testing four concepts; you would have to mail at least 20,000 names!

Fortunately most of these rules are over-simplifications of what you need to do to make sure that your tests help predict future results. But statistics do count and if you test without understanding this, you may end up with useless results. Testing, after all, is primarily designed to help improve results and limit risk for future mailings. And with a smaller budget, you need to limit risk all the more.

Predictive testing
Whether your test is predictive or not is based more upon the number of responses you get than on the size of the test mailing itself. There is an easy-to-understand statistical spreadsheet that will tell you how much you can count on your test results — and even give you a minimum size for your test. It is too detailed to show here, but here are the kinds of information it will give you:

Test Size	Response Rate	Predicted Response Rate*
1,500	1.00%	0.50% to 1.50%
5,000	3.50%	0.60% to 1.10%
5,000	3.50%	3.00% to 4.00%
6,000	1.00%	0.75 to 1.25%
20,000	3.50%	3.25% to 3.75%

* Statistics say that 95% of the time this will be the range of responses when you remail. Call me and I will send you a copy of Murphy Marketing's Effective Direct Mail Testing: How to Determine the Predictability of Your Tests.

Use Cumulative Results
Even if your tests do not appear to be predictive based on statistics, you can add up several mailings to be more certain of your results. So there is no problem if you are mailing only 1,500 names from a certain list and your responses are in the fractions of a percentage point (and the program is profitable). If you mail several times per year, you can add up the results and compare this list to the cumulative results of other lists to decide whether to continue to use it.

Of course, you can also watch the trend of your responses to make your decisions. You can use the same approach for making decisions on testing offers such as premiums and discounts by back-testing against an offer that won in a small test mailing.

Keep in mind that you need to reserve your larger test panels for testing critical elements such as offer or full package. You do not want to make strategic decisions based on results based on extremely small tests!

Follow consistency of results
You also can rely more securely on your test results if you pay attention to consistency. If all your tests with premiums beat the non-premium packages, you can use fewer names to test premium packages. Similarly, if you become knowledgeable of your industry and find that other similar marketers are having good luck with premiums, you can feel more confident about rolling out your premium offer based on a small test.

Seasonality and timing
Since you have to make the most of smaller mailings and budgets, be sure that you test and mail the most in your strongest season. The beginning of the calendar year is the strongest season for many consumer as well as business-to-business marketers. Know your seasons and take advantage of them.

 

For further information and a copy of Effective Direct Mail Testing: How to Determine Test Sizes, an easy-to use test analysis spreadsheet, call Maureen D. Murphy at (310) 395-9722.

Maureen D. Murphy is owner of Murphy Marketing of Santa Monica, California, serving the direct marketing services needs of southern California businesses. Her background includes working with business-to-business marketers such as 20th Century Plastics, Dell Computer, United Ad Label Co., and the Merritt Company, as well as consumer marketers such as Santa Anita Park, Century National Insurance, Allstate, Transamerica and Nissan.