24 Sep 2014
A/X Testing? That’s Weird!
24 Sep 2014
A new term’s been surfacing up a lot on our blog, newsletter and website recently; A/X testing. But what exactly does it involve? Your immediate thought might be, “it sounds like A/B testing … but taken to infinity?” Good guess, but it’s not quite as scary!
To put it simply, A/X testing is a multivariate testing tool that allows you to test the performance of up to 10 different email campaigns. Through this tool, you’ll be able to use a mix of four variables; Subject Line, Sender Name, Reply to Name and Email Content, to create various email versions and run them through the test. This is a quicker process than your traditional A/B testing that tests one variable at a time.
To give you a better idea, let’s say you want to send an email campaign with an optimized subject line and sender name. With A/B testing, you’d have to run two tests – one for each variable, but with A/X testing, you can test both with just one test.
When first getting started with A/X testing, you’ll want to keep these four points in mind:
1) Pencil it in
There are two ways to implement your A/X test, either send now or schedule for a later date. Often marketers put off testing because they don’t have the resources or time to do so. Being able to schedule your A/X test for a later time allows you to create the campaign when your workload allows and plan ahead to analyze the results once the test is complete.
2) Sample Size
You’ll also want to send to a smaller, randomized portion of your contact list to test for the most optimized email version before sending to the rest of your contact list. Never send blindly again – support each email decision with hard data.
3) And the winner is…
From the very beginning of the testing phase, keep in mind what activity you’d ultimately like to track and it’s corresponding KPI. If you’re looking to find the winning email combination that will drive most website traffic, you might want to select “click rate” as your winning criteria. If you want to test the effectiveness of the sender name and subject line, you might want to select “open rate” as your winning criteria.
4) Length of test
Based on the average engagement of your contact list and the variable tested, you may want to adjust the length of your email test accordingly.
Happy testing! Stay tuned for more A/X testing tips in the coming weeks.