Firms must be aware of the common drawbacks associated with split testing if they are to improve the success rate of their email marketing campaigns. Split testing can provide firms with valuable information that can improve email marketing campaigns, but only if it is correctly carried out.
A split test is a simple way of determining what kind of email will produce the best results from a firm’s audience. It involves dividing an email audience into two separate groups (A and B), and sending two different versions of an email to the two randomised groups.
Below are three common mistakes associated with split testing that firms should understand before creating their own split test.
- Making sure your list is random
One of the most common mistakes of split testing is dividing a contact list in half, and using those two halves as the A group and B group. The problem with this is that one half of the list will be made up of older subscribers, and will perform in a mediocre way compared to the newer segment of the list. A better option is putting every second name in a contact list into one group.
Once a firm’s contact list is split in two, firms should then check each half of the list. If one half has a majority of something, like Hotmail addresses, then a firm may experience email deliverability issues with that half of the list. This will then impact on the end results of the email test.
- Only test ONE variable at a time
Testing more than one variable at a time prevents users from determining the impact each element has on a campaign’s performance. The only way to determine if a specific variable is the reason behind a change in email results, is to test that one variable by itself.
- Don’t overanalyse
If the split test was set up correctly, then the end results are the results. The value of a true random split test is that the results are typically quite conclusive.