The best way to improve your direct mail program is through constant testing. Save money and time by following these eight rules.

Rule #1: Test things that are significant

Don’t test a blue signature versus a red signature. Test a six-page letter versus a two-page letter. Test a full-color package versus a black and white package. These would be significant tests.

Rule #2: Test things you can control

You can control your printing, length of the letter and photography. Don’t test factors you have no control over-for example, mailing during a recession. You could have no control over that in the future.

Rule #3: Make your test large enough to be significant

Mail at least 5,000 pieces to get 50 responses at a one percent response rate. Fifty responses is considered the minimum number you must generate to achieve a trustworthy test.

Rule #4: Test one variable at a time

When testing, only to change one variable at a time.  If multiple variables change within one mailing, you will never be sure which variable truly drove the higher response rate. Try to test one variable at a time to best calculate results.

Rule #5: Don’t let large gifts skew your results

If your gifts are typically $25, $35 and $45 and you have one donation of $15,000, remove this amount from your average gift amount.  That number can skew your results and future projections. Your spreadsheet for this campaign may tell you that the average gift was $92, when the reduction of this $15,000 gift will significantly lower your average gift to $36.

Rule #6: Test acquisition mailings for people, not profit

Concentrate on acquiring loyal donors in the greatest numbers at the lowest cost.  The results will follow.

Rule #7: If you are starting out in direct mail, aim to acquire volume in donors

If you’re just starting a direct mail program, aim to acquire as many donors as possible versus collecting the highest profit. Success in the mail comes after the first gift.

Rule #8: Test every time you mail

If you’re mailing to a large number of people, always test something. Test two asks, test the ask string on your reply device, test the teaser copy, test the length of the letter, test the signatory, test something of significance every time you mail. You’ll learn a lot, and you’ll improve your packages and your approach over time.