A/B testing is the method through which you optimize your website and your sales funnel. Wondering what the A/B is? It is probably obvious, to be completely honest. You test two forms of a given variable against each other. The first is the A and the second is the B. Once you have done the testing, you figure out which one is working for you and then you selection that one. Now you have a new A. Make a new B and repeat the process ad infinitum. These types of tests are also, sometimes, called split tests.
Data is the product. Maximizing data delivery is your main goal with your site. Data is your lifeblood. Here are some of the benefits of A/B testing the data on your site:
- Increased profitability – The way your pages are laid out. The way that the links are arranged. Even your copy. All of these things, and much more, can be modified through A/B testing to maximize your profitability.
- Content – Your content is generally going to be the same. The way that it is delivered, however, can always be tweaked. One word can be the difference between a sale and a pass. Using A/B testing will let you test your content against itself over and over until you get the maximum benefit from it.
- Risk alleviation – Your inherent risk is going to go down the more you choose to utilize A/B testing. The worst thing that you can possibly do, in terms of risk, is never change. All A/B testing can do is give you more options to use moving forward into the future and give you better versions of what you already have. You never have to worry about getting worse, because you can always revert to your current form after you review the stats.
- Understand visitors – You’re a/B testing is going to let you see what content does what. Which content affects the ways that your readers are going through your website. It lets you see how the CTA text might affect their next move. It lets you see how your headline might pull readers in (or push them away).
What to Look For in Your Tests
Your A/B testing is going to provide you with some very specific pieces of information, depending on what sort of statistical tracking you are doing. You need to look at that information very closely to determine which version of your test is the better version.
Here are some of the most important pieces of information to look for in terms of your A/B testing:
- # of visitors to your site
- # of unique visitors to your site (note: this is NOT the same thing as the above metric)
- Average length of stay
- Bounce rate
- Pages visited (traffic flow patterns)
There are others, of course. It depends on what you are trying to do. These are the most basic. Where are your visitors coming from? What are they doing? Are they showing up and leaving immediately? If so, maybe your headline is the issue. What are they doing once they go to your landing page? Are they going through with your CTA? All of this information can be found in your metrics. Pay attention and use that information.
Implementing an A/B Testing Program Today
The best part about A/B testing is how easy it is to start doing it. Presumably you are already tracking your site performance using tools like Google Analytics, right? (If you aren’t – stop reading right now and go implement that, it’s absolutely essential as well).
Pick some things which will have a large impact to begin with. For example:
- Sales copy
- Your headline
- Your subhead
- Your CTA copy
- The colors of your CTA
- Your hero image
One thing you will need to note, and keep in mind, before implementing any A/B test: You only change one variable at a time.
You don’t want to go changing your entire layout every single time you do something. You want to change one thing at a time. One headline. One link. One font. This takes longer, yes, but you will better be able to tell what information is doing what and how well it is working if you go small to large rather than the other way around.
Here are the steps to beginning an A/B testing program:
- Install analytics or another tool you can use for metric and information gathering onto your site.
- Pick a variable you believe could be improved.
- Create a second version of the variable.
- Run them at a 50/50 split between visitors to your site.
- Review the statistics and see which version is better.
- Repeat the process.
Sound simple? That’s because it is. Let each test run for a couple of weeks (or months) before picking one, however. You don’t want to be running a test for a very short time because the metrics that you get back are going to be biased. Allow the test plenty of time to run. Don’t rush things. Once you have a good sample size you can then move forward with the best option from the split and then repeat the test with a new variable if necessary.
Note: The exact steps will depend on your platform. For example, if you are on WordPress there are A/B testing plugins available. If you are on Leadpages, the same thing may apply. If you site is hard coded, this will be a bit more involved. The process can be pretty overwhelming for the uninitiated. Elaborating on the process of running pages against each other on normal websites is a bit beyond the scope of this article, but here are some resources that can help you:
- WordPress – With WordPress you will likely need to use a plugin to make things easy on yourself. Each plugin has its own method of action, so take a look at a few and see which one fits.
- Google Tag Manager
Finally, keep in mind that you don’t only have to test things on your landing page. Everything is up for grabs. Your emails, your newsletters, your traditional marketing, even your online ads. All of it can be tested. All of it should be tested.