Sunday, March 16, 2014

Test Your Software Upgrade Announcements

If you're fortunate enough to have an installed base of, say, 1,000 software users, and you're about to announce an upgrade to your program, then don't send all 1,000 of your customers the same message. Divide the file into groups, and send each group a different message.

The common term for this type of testing is the "a-b test." If there are only two versions of your sales letter, then this name is accurate. With four variations, call your test an a-b-c-d test.

You'll probably find that one of the sales presentations will out-perform the others by a significant margin. And you can use that information in future emailings to improve your software sales.

If your installed base is larger, you could first run a test using a subset of your database, and then use the results of the test when you write to the remaining users. For example, if you have 10,000 users, you could create three different sales messages, and send each of these messages to 1,000 users. In a week or two, you could determine which sales letter produced the best results, and you could send that version to the remaining 7,000 people in your database.

The winning version becomes your control letter. And the variations you'll make to your control letter are your test versions. A year from now, when you have another software upgrade to announce, use your control letter as the measurement stick, and compare results with each new test version. Rinse and repeat each time you send an upgrade notification to your customers.

By fine-tuning your sales message year after year, you can generate significantly more income than you would if you were to simply guess which version of your sales message would be the most effective. A-b tests are good software marketing.

No comments:

Post a Comment