The secret to success in business-to-business direct mail lead generation is testing. I have hunches. You have hunches. But testing settles the debate. When you test your package against my package and we measure the results, we know for certain which package performed better.

Assuming your list is good (in other words, assuming you are mailing to people who are likely to buy), the first thing you should test is your offer. Your offer, after all, is the second most important ingredient of direct mail success. And as Axel Andersson says, If you want to dramatically increase your results, dramatically improve your offer.

How do you dramatically improve your offer? By creating an offer that is different from the one you are using now, and testing both at the same time. Basically, you mail two packages, one with Offer A and one with Offer B. You test nothing else. Then you measure which offer outpulls the other. Here is how you do it.

Test a hard offer against a soft offer
A hard offer asks for an appointment. A soft offer invites prospects to request more information (such as a white paper).

Test a product literature offer against an educational offer
Product literature includes brochures, catalogs, sales sheets and technical specification sheets. Educational offers include white papers, special reports, books and article reprints.

Test one educational format against another
Some prospects prefer their information on paper. Others prefer going online to a special page on your website. Still others prefer watching a video, or listening to a CD.

Test offer descriptions
Should you say Buy two for the price of one or Buy one and get one free? Test and youll know.

Test a deadline against no deadline
Giving your potential customer a deadline for responding might boost response. And it might not.

Test one premium against another
Offer an Apple iPod in one mailer and a Blackberry in another. Offer a gift certificate to Amazon.com in one letter and a CD player in another. See what happens.

In all your tests, remember to test one thing at a time and to make your tests big. Keep all the other variables (timing, design, list and so on) the same while you test just the offer. When I say make your tests big I mean that you should test two very different offers. If the offers you test are too much alike, you will not be able to trust your test results. And thats not just a hunch.