How to split test effectively
Which of these pieces of art is more beautiful? Now, this is, of course, subjective and a matter of opinion. Some among you may prefer da Vinci’s mysterious Mona Lisa, some of you may prefer the vibrancy and colour of Van Gogh’s Starry Night. You need to split test as there is not really a “right or wrong” answer.
As email marketers, we often have to make decisions on which of two subjective things is “better.” Which subject line will get you more clicks? Which design is more effective from a visual point of view? More than anything, what is going to get me the results I need?
“Testing, we will never do enough of it.” – Greg LeMond
The fact of the matter is, no matter how experienced you are, no matter how strongly you feel, you don’t know the answer until you’re looking at it in the stats. We all know this and we’ve all been there. We agonise over whether to go with title case or sentence case, whether to personalise a subject line, which of two banner images will perform better. But why sit there and worry about it when you can do both?
Using a split test
Everyone knows about split testing – send two, see which one does better. This can, of course, be done manually. Segment your audience into two halves, send them two separate campaigns, and then manually compare the results. Gator actually makes this a lot easier for you – all you need to do is pick a percentage of your audience to use as your test sample and the criteria which will decide which has done better (either open rate or click rate). The campaign will then pick a winner automatically and send it out to the rest of your audience. This saves you from having to segment your audience, analyse your results or re-send the rest of the campaign when your split test is done. Easy!
Getting started is simple. In fact, we have a guide which shows you how to do it.
What should I test?
EVERYTHING. There is nothing about your email designs which you should not be testing – if you’ve got the time, then do a test. Start off with what makes the biggest difference, the most fundamental things in your designs: the subject line, the wording of the main CTA links and the use of personalisation are all good high-impact places to start. After that, why not try expanding to images and the visual aspects of your design? The options are endless, and the more split-testing you do, the further you can refine what’s working and what’s not working in your campaigns. When it comes to split testing, the world is your oyster. (Or lobster. I haven’t looked at the stats yet…)