Test the basic elements first: subject lines, “from” names, and email copy. It does not take too much time or creative work to come up with some simple tests.
When running any A / B test, make sure to normalize your sent time. Even a 30-minute difference can change the results of your test. Do your best to send emails simultaneously to send the optimal time to your customers.
If your sample size is too small, you can actually say a winner without one. As a good rule of thumb, make sure you have at least 1,000 observations for any test. For example, if a subject line exam is underway, you need to make at least 1,000 openings per email to see statistical significance. If you did not find the significance, you should run the test again and collect the data.
It wants to test more than one variable at a time, because the more you test at once, the bigger the impact you can make and the faster you can improve. But the challenge is that you will not know the individual improvement or decline in each variable. Test one variable at a time, measure the results and roll out the winners in your new email over time.
You will see the test result which has no meaning. You do not want to believe your eyes. But email marketing is not a cupcake competition. The best, most engaging email may not be the most beautiful or your favorite. Trust the data and your email results will be starred! [: Gb]