Often in life we need to make a choice between two specific things, and if you grew up in the 80s like I did, you will remember the Pepsi Challenge adverts where blind tests were administered to determine which tasted better: Coke or Pepsi.
Today is National Have a Coke Day. But I will be drinking Diet Pepsi. Why? Because I don’t drink sugary sodas and Diet Pepsi (unlike Diet Coke) is made without aspartame. Also (and this is my opinion) Pepsi just goes down cleaner and is more refreshing than Coke. I’ve always thought so, and I always will.
What is A/B/ Testing?
A/B testing is like the Pepsi challenge of any digital marketing strategy. Also known as bucket testing and split testing, A/B testing is a method used to optimize marketing strategies by comparing two web pages against one another to determine which performs better. Statistical analysis and heat maps are used to pinpoint which variations contribute to performance metrics. This also allows marketers to determine which questions to ask in regards to making optimization changes. This may be more complex than pouring Coke into one cup and Pepsi into another with a blindfolded test dummy at your table, but with the right plan and execution it is actually very straightforward.
How A/B Testing Works
Start by taking a web page that plays a pivotal role in your conversion process. The best digital marketing agencies offering custom Growth Stacks create strategic content that educated leads; nurturing them through the buyer’s journey until they can make an educated buying decision and convert into revenue. If you have a similar process, then it is good to prioritize A/B test pages that play huge rolls in your process. Once you have determined your web page, create a modified second version of it. The modification can be as simple as changing the placement of your CTA, moving the button, or re-wording your headline. Then show half of your traffic the original page (the control group) and show the other half the modified page (the variation).
Don’t Let A/B Testing Harm Your SEO
There have been times when I order a Diet Pepsi and the server gives me a Diet Coke thinking that “cola is cola”. But it isn’t, dammit. There I was, innocently anticipating the clean, crisp taste of a Diet Pepsi only to get a mouthful of aspartame and disappointment. When you perform A/B testing your goal is to ultimately find a way to have a better experience, but there are ways you can, to no fault of your own, crap on your own SEO efforts.
Google allows and even encourages A/B/ testing. The search engine has even said, on multiple occasions, that it poses zero risk to domain authority or search rank. BUT there is a way you can endanger your ranking by misusing an A/B testing practice, like cloaking. This is when Google is shown a different version of content than what the searcher is meant to see.
I have learned to ask restaurant server’s if they carry Diet Coke or Diet Pepsi because I don’t want a bad taste in my mouth. When it comes to avoiding naughty cloaking that can harm your SEO efforts, you need to make sure Google clearly understands what is meant for its eyes, and for your customers. You can do this by not abusing visitor segmentation to display opposing content to bots based on IP addresses or user-agents.
You should also use rel=”canonical” if you are split-testing with several URLs. This will point the variations to the original version and help prevent Google from getting confused by several versions of the same page.
Don’t use 301s; use 302 redirects because they tell Google that this environment is only temporary, as a 301 implies a permanent change. Also, make sure the original URL is indexed and by no means should you index the other.
Finally, only perform A/B testing as long as it is necessary to get enough data to determine a recourse. If the testing goes on for a lengthy time Google might regard it as an attempt to deceive them just like a silly waitress trying to pass off a Diet Coke for a Diet Pepsi (boo, hiss).
Why do A/B Testing When there are Other, Easier Tasks?
A/B testing is regarded by many businesses and even the agencies they hire as being too much work for minimal feedback. However, this task helps agency teams and companies make careful and necessary changes to improve their user experience by collecting data that isn’t available through any other method. When executed you are able to build hypotheses, and see which components interrupt the buying process. A/B testing can also be used to prove that opinions are wrong (or correct) and help managers and directors see who the real savvy members are on their team.
A/B/ testing should never be practiced as a “one off” thing. It should be used in an ongoing fashion to make continuous improvements to an agile strategy to help increase goal accomplishments.
Don’t Get Greedy
Just as you shouldn’t enjoy more than one or two Diet Pepsi’s a day for health’s sake, you also want to be conservative with your A/B testing in terms of what you are monitoring. Don’t try testing more than one thing at a single time. If you try testing more than a single change at a time then targeting in on the exact change that garnered the biggest shift in user behavior can prove to be very difficult. Go ahead and test multiple things, just do each at a time. Then, over time, you can combine the data from all of your A/B tests to discover an overall measurable way to make several changes to create a new user experience that draws better engagement and converts higher revenue (if that is the goal).
Did You Love these Tips on A/B Testing?
If you valued the A/B testing tips, we invite you to download our FREE guide on the WEBITMD customized Growth Stack. Our Growth Stack offers a set of tools and strategies that work together in unique ways, per client, to foster business growth. Then if you can imagine your business blowing up with our Growth Stack, give us a call and let’s discover the potential!