Splitting your email list in half and sending two different emails does not an A/B test make. Going into your test, make sure you have a set of clear cut variables. Instead of testing three random subject lines, try testing one that includes an offer against one that doesn’t include an offer, and one with some info about the product, for example.
Try to identify different ways to promote something with a replicable variable. This will give you a pattern to replicate in subsequent emails. This is your chance to try several different subject lines, like something cheeky or off the wall and see how it would work. And, of course, don’t forget to do something with your results.
Limit the Number of Your Variables
While it may be tempting to test a bunch of different things, hold off. With each additional variable you introduce, you’re also introducing more uncertainty. Was this the best performing test because of the puppy, the blue box, or word choice? Instead, try something like testing a product image versus a lifestyle image, or a male versus female image, for example. Remember that what works for retailers might not work for B2B or restaurants, so be sure to test and test often.
Remember to Plan Ahead
All of these tips I’ve just mentioned will be null and void if you don’t get ahead of your email program. Give yourself time to test. Make sure you have at least a full day for your test to run and collect results. Some people may not check their email until later in the day.
On a bigger scale, try to get a few months ahead of your programs so you’ll have time to plan what kind of testing to do, and be able to put your results to work. If you’re testing something like a monthly newsletter, it could take up to a year to get through just the first round of your testing variables.