The move from mass marketing to targeted closed-loop marketing is usually accomplished through testing. Best practices for testing include:
■ Test one thing at a time.
■ Determine a goal/objective for your test. It is important to know that you are doing what you set out to do.
■ Understand what makes for a statistically valid result. The simple rule of thumb is to include enough records in your test to ensure that you receive at least 100 responses.
Market testing within your own database requires that you use random segmentation selections. This means that you establish a group that you would expect to perform in a particular way and then randomly select the number of records you want to test against.
Statistical validity is measured not just by having test groups of the right size, but also by the use of control groups. Test groups must be measured against a control group. The control group will receive the same package as usual or be the same group as usual. This group will receive the usual expected result unless an outside phenomenon changes the outcome. The test group will perform differently than the control group.
If the objective is to “recruit more donors”, then you can not say that the test failed because the return on investment is lower than the control group. Return on Investment is a different metric than percentage return. The percentage return better measures if this objective was met.
To accurately determine whether a test is successful, it should be repeated. Future campaigns should repeat the test on new groups of randomly selected records. After the test has produced similar results three or four times, consider increasing the size of your test groups and introducing your new tactic on a permanent basis. Again, vendors can work with you in developing tests, measuring their effectiveness, and deciding when the test has finished successfully.