Best Practices for AB Testing


Best Practices For AB TestingIt is time to get real about A/B testing.

What is A/B testing, I hear you ask? The problem is that this is a question many micro business owners have. If they have even heard of the practice at all, of course, which many of them haven’t. No one has been around to explain it to them.

You might be surprised to learn that A/B testing is both a marketing and a design practice. Normally those two elements would be separate without much overlap. Here, the rules change. The idea is to not only improve the functionality of your site, but to do so based on information that makes it more engaging for users.

By giving the average site visitor what most appeals to them, you are first improving the chances of them returning. While this isn’t the only application (we will also go over other ways of utilizing the tool), it is an important one to address as soon as possible. When your site meets the needs of the consumer, they will stick around to buy a product or service. In an age where alternatives are all over the place, you have to go that extra mile to come out on top.

What Is A/B Testing?

Using it in a design context, A/B testing is the process of taking two different versions of the same content and diversifying it. You then use a control base to see which one has the better outcome. From there, you establish which is better based on several variables, such as:

  • Amount of time a visitor spends on the site.
  • Amount of pages visited on the site.
  • Amount if sales made through certain designs.
  • Amount of social media shares or email subscriptions.

A good example would be using two different layouts with the same images and text. While what you are offering is the same, the way it is presented is not. You would take one group and have them browse one design, and use another to browse the second. Then see which one leads to greater results. This could be done all at once with two comparable designs and metrics tools, or one after another during a greater block of time.

In a marketing context, the idea is similar. Say that you send advertising emails to the people who have subscribed to your newsletter.

Email 1: “Get 20% off when you shop using PROMO CODE A!”

Email 2: “Hurry and save! Get 20% off now when you use PROMO CODE B on any order!”

Two different messages, two different codes, but the amount off is the same. You send each email off to a hundred customers each. Then you sit back and see whether the first or second code sells more, to determine which marketing tone works best.

By studying contrasts in both design and advertising, you can establish a more targeted overall campaign.

How To Get Started

The first step is deciding exactly what your goal is. Are you hoping to make a more functional, intuitive site that is easier for user navigation? To engage on a more personal level in order to build brand loyalty? Improve your marketing results? Get more people to use a sign up form?

All of these, and much more, can benefit from A/B testing. The general principle behind it is always going to be the same, just comparing one thing against another. That is what makes it such a valuable tool in both design and advertising.

Once you have established exactly what you hope to achieve, it is time to decide how long your experiment is going to run. Recommendations on this differ from person to person. Some experts swear on the short lived tests that go for a couple of days to a week. But remember that you can only work based on how much data you have.

You want to walk the line between enough time to get your metrics, but not so long that the data is hard to sort through. Look at how many visitors you tend to get to your company website per day, week and month. You want to have the chance to test at least a couple hundred users for each version. However long it will take to reach that number is the length you will want to dedicate to your test.

When you launch each variation, make sure to remain consistent and to only cater it to a single group. If you go back to test again, use a new group for the new content. Recycling users in tests is a quick way to muddle your results, and you will want to run A/B tests multiple times to double and triple check results.

Finally, once you have the results, use them! As the owner of a business, you are used to following your instincts on a great many matters. Discernment is a valuable asset in most things, but not here. The data is the only important element to A/B testing, and you should base all decisions on that data. Even if it doesn’t correlate with what you expected, or even what you would have liked.

The facts don’t lie.

Tools For A/B Testing

A/B testing is cheap, but time consuming. It is better to automate and simplify as much of the process as possible. There are plenty of tools aimed at micro businesses that will help you to do just that.

  • Visual Website Optimizer – Create two or more versions of your site, make variations multiple times, target user behaviors and then see the results. No coding necessary.
  • Optimizely – A much more extensive A/B testing tool that gives you everything you need for thorough tests. Allows you to track goals through established metrics.
  • Unbounce – Focusing on landing pages, you can use this tool to build, publish and then track and test the results.
  • Genetify – If you have experience in coding, this is a full A/B testing software that caters to advanced developers.


A/B testing is an often overlooked practice, despite it being such an important one. It is also an inexpensive way for micro business owners to better target their site and marketing to their demographic. The sooner you start, the sooner you will be able to improve your engagement, and by extension, your sales.