What is it and why is it important? A/B testing is exactly what it sounds like – testing multiple versions of a website element with a metric (a measurable goal) to see which one wins the bout. It’s also known as Conversion Rate Optimization (CRO), or UFC (Ultimate Fight for Conversion).
There are three phases of a successful, lead-generating website, and CRO is part of the third phase. The three phases are Attract, Inform and Engage.
Attract – Your website should be attracting the right audience. This begins with a solid SEO strategy.
Inform – Once you’ve attracted a visitor, your website should inform them of your expertise through content strategy.
Engage – Your SEO brought them there; your content intrigued them; now you need to convert them. This is where CRO comes in.
The first step in the CRO process is to identify something you want to accomplish – a goal. Some of the questions you need to ask are: What is your website trying to achieve? What’s the best strategy to achieve it and where do you start? What pages and design/content elements do you want to send into the website octagon first?
Those elements could be anything from images and headlines, to copy, layout, forms or call to action buttons. You can test practically anything on the page. An important thing to remember is that CRO is not just for e-commerce. Your goals might include blog subscriptions or comments, page views or (a biggie) form submittals.
The next step is to choose the tool you’re going to use for the test. A free tool that you probably already have access to, but may not know, is Google Website Optimizer (GWO). If your site is hooked up to a Google Analytics account, you’ll find it under the Content tab, where each test is referred to as an Experiment.
Besides GWO, we at R+M also like Visual Website Optimizer. It costs to use this tool, but they have some pretty cool features – like heatmaps and geo targeting – built right in.
Another tool we LOVE is Ninja Button. It’s super simple (it tests buttons only), it’s fairly cheap, fun to use and any time you can invite a ninja to a cage match that’s a win.
Basically, all of these tools have the same approach – segmentation, variation, measurement and goal. It’s a lot like the science experiments you did in middle school. You take two plants (your audience), separate them (segmentation) and expose each of them to different type of martial arts (the variation). One plant learns the art of Brazilian Jiu Jitsu and the other learns Greco Roman Wrestling. In the end we conclude (what we already knew was going to happen) the Jiu Jitsu plant forced the Greco Roman plant to tapout (the measurement).
Your tests will be very similar. Your audience will be segmented and driven to a variation of whatever element you’re testing (the green button versus the red button). Your chosen tool will track which button receives the most clicks. When your test has achieved statistical significance (more on that later), your goal is reached, your test concludes and a winner is chosen. Then you can implement that change into your site immediately and move on to test bigger and better things.
And trust me, you will want to. It’s fun to watch and it can be addicting. Will the red button win with an Anaconda Choke? Or will the green button achieve an Armbar and win with a tapout? We won’t know until you invite them into the website octagon. We love organizing these sanctioned matches. If you need help picking a fight let us know.
Props: we’d like to give a shout-out to our friends at Newfangled for their awesome insight into attracting, informing and engaging.