AND NOW FOR SOMETHING DIFFERENT

	Package insert programs

These inserts can usually be larger in size, as they are inserted directly into the packages of outgoing orders. The inserts ride along with the merchandise, usually in an envelope marked something like “Special offers for you.” Since orders are being shipped continuously throughout the year, you can buy insertion at any time, based on quantity or time of year, etc.

Here, you pay for printing your own inserts, plus the cost of insertion, which can run in the range of $25/M-$45/M; expect printing to be somewhere in the $8/M-$15/M range. Lead times are in the two- to three-month range. Responses usually “lag” as they are shipped out with orders over a period of time.

	Affiliate marketing

Online, the evolution of CPM to CPC is now continuing to CPA, cost per action. Affiliate marketing — if done properly and with the right people — can be a powerful way to leverage your online dollars. As the advertiser, you pay only when a specific action you define — such as a purchase or catalog request — is completed.

Aside from finding the right affiliate program, the key is constantly managing your performance across each publisher Website, optimizing strong sites — you can often set your CPA differently among publishers.

A test program will likely take two to three months to get off the ground, and a minimum test usually lasts six months. Costs will vary, but if you can afford roughly 7.5%-15% of an order total to get new customers, then you will do well with affiliate marketing.

	WHERE TO BEGIN

There are three essential keys to a successful insert media test:

1. Test a wide range of different files

   Design a test, on your own or with a trusted broker. For example, say a gift marketer wanted to run an alt media test. A good recommendation would be to test 12 very different consumer catalog insert programs, three credit card statement insert programs, and three targeted business-to-business card decks. When the results are in, not all will be successful — but several will.

   Next, zero in on the successful files, and find all the others in their categories. Launch a second round with increased circulation in the repeat, successful programs, and test circulation in a dozen or so of the new, test files.

2. Test as little circulation as possible within each file to spread your spend more among many files

   But remember, the results need to be projectable. Be sure to give yourself enough test circulation to get statistically valid results.

3. You must test multiple offers in any new media

   How many offers depends on your business, products and number of controls. As a rule of thumb, if you can, try to initially test at least two controls or one control plus a variation (free shipping, multiple payments, a new premium, etc.). This will help give you a read on the first round as to whether it's the media or the offer that is responding, or not responding.

---

Jeff Marsh and Karen Mayo are principals of Mayo and Marsh, a direct marketing consulting firm based in Waterford, VT.

	Five tips for alternative media success

1. Study the data cards carefully

   A program may seem like a perfect match for your demographic, but a closer look might reveal otherwise. Consider average unit of sale, continuous usage, and how the catalog's mailing list was developed.

2. Never test a prospecting or reactivation file

   You'll want to make sure your insert is riding along in the catalogers' drop to their house lists, which will serve as a good control. Try to be in the first drop to house list as often as possible.

3. Test as many files at the same time with your initial test

   Have at least two offers in a statistically significant A/B split test.

4. For larger catalog insert programs, be sure to test at least 100,000 inserts to get a good read

   As for package inserts, you can test a much lower volume.

5. Tap your broker for advice on offers that have worked well in each file, as well as any publicly known information about repeat advertisers

   He or she should be able to provide a wealth of knowledge that may not be reflected on the data card.
   —JM/KM