Direct Mail Statistical Analysis
When measuring the success or failure of your A/B test campaigns, its fundamental statistics will help you gain insight into which call to action, which copy change has derived the most responses. In direct mail you understand your results through the use of codes, PURLs or telephone numbers placed on every item you mail. Every response you receive you monitor day by day so that after a short time you can forecast results learn the lessons and implement without waiting until the offer ends. By having a control and a test group, you will be able to divide the number of responses by the number of recipients to understand which works best, meaning you can make modifications for repeated mailings.
The test quantities can vary but a minimum of at least 1,000 but preferably 5,000 is required so that you can have some trust in the statistical probability of your test being accurate. Experiment with the distribution of 90/10 percent when trying a new method that might not require the even split of your database, or if you’re cautious about its potential effect. It’s a trial with reduced error, which should put your ‘mind at ease’. It’s a small step towards great change; one of the continued benefits of employing direct mail in your marketing.