What to Consider Before Finalizing Your 2011 Catalog Circulation Plan

What should the sample size be? One consideration is the cost of running a test. Most mailers can’t afford to bet a Volvo to test a single mailing list. Figure out the smallest test possible that will yield good data on whether a list will respond profitably. If a catalog costs $1.00 for print and postage, mailing 30,000 may be a great sample size, but it costs $30,000. Would a test of 5,000 be adequate?

You might be able to test a mailing list using a sample mailing of 2,500 names, knowing that you'll be able to read if the list has any potential, but also knowing the test is too small to gauge how responsive a list is going to be.

Catalogers use “stage prove methodology” to test larger and larger quantities. The rule of thumb is if you have a successful test of 5,000 names, then you can increase your test quantity to 15,000. If the test of 15,000 proves profitable, increase the test size again by a factor of three to, say, 45,000 or 50,000. What you should never do is take a small successful test of 5,000 names and roll it out to 100,000 names or 250,000 names without an intermediate test. The increase in quantity is too risky. It's always preferable to have several data points when you want to increase the size of a successful mailing list.

Why a list is only as good as its last mailing. First, lists get fatigued as they're mailed over and over because you've successfully harvested buyers from the lists. Second, economic conditions were much different in the past. Mailers have had to either heavily discount or throw out results from old tests conducted during the downturn in response that was caused by the recession. The economic downturn rendered many old tests almost meaningless.

Why future tests get worse. With initial and small tests, the list owners providing the names may give you the best names possible so that you’ll continue to use the names. You may not get the very best names when you recycle a list. Lists rarely, if ever, do better than they have responded in the past.

If a list has performed poorly in the past, it will probably perform poorly in the future. The list industry is designed to sell mailing lists. List owners and brokers will often encourage mailers to retest a list because “it should work.” Be very conservative in testing lists that have responded poorly in past tests. There's a bias toward mailing more rather than mailing less because all suppliers (e.g., printers, list brokers, company management) want growth.

When a promotion ends, expect your sales from that catalog to also end. Catalogs deliver sales over a long period of time, and this order curve is usually a very stable number. Today's catalogs, however, rely increasingly on promotions to drive sales. If a promotion expires before the end of a catalog's natural order curve, sales will also be cut off and the normal order curve won’t apply.