AB- Split Testing The Ultimate Way To Rocket Your Sales

March 20, 2017 / Business

We all make changes to our websites in the hope of making them more effective but, for many of us, these changes are done blindly. What if you could make changes that you know will work? Well, you can, through a method known as A/B split testing. Split testing has enabled companies to make massive differences to their sales and in this post, we’ll explain what split testing is and how to use it on your website.

What is A/B split testing?

Split testing is when you create two versions of the same web page and show each version to half of your visitors. You then compare which one of the pages has performed the best and keep that one on your site.

The key to success is to make sure that the differences between the two pages are as small as possible. For example, if you want to test the effectiveness of the colour of your call to action buttons you would make sure the pages are otherwise identical. In this way, when you see one page perform better than another, you know it is down to the colour of the button and nothing else. According to Unbounce, one website improved its conversion rates by over 35% just by changing its Add To Cart button from blue to green.

Testing options

There are lots of things you can test with A/B split testing, such as:

  • Colours
  • Font style and size
  • Page layout
  • Images
  • Call to action placement
  • Call to action timing (for popups)
  • Call to action wording
  • Effectiveness of specific parts of the text: value propositions, product descriptions, headings, etc.
  • Pricing option design

By testing each of these and working on one thing at a time, you can fine tune your website into a precision-made conversion machine. Over time, you can improve the effectiveness of every area of your websit not through guessing what you think the customer will like but in response to what is proven to work.

Domain Name

How to carry out A/B split testing

To carry out split testing effectively, you need to take a methodological approach. Here, we’ll take you through the process, step by step.

  1.  Prioritise the improvements you want to achieve
    The first step is to take a look at your business goals and consider how you want your website to help you achieve them. You may want to boost sales, increase email subscriptions, get customers to choose more expensive membership packages or have other important actions take place.
    Your task is to discover the barriers that stop these things happening so that you can create a list of things to improve. For example, do you have an overcomplex signup process? Is your call to action uninspiring? Are your products too difficult to find?
  2. Choose which areas to test
    Once you know your priorities and the obstacles you need to overcome, you have to choose which improvements to implement and test. If your sales page isn’t selling, what changes can you make and test on that page? The list of options is a long one, so you will need to make an educated guess about what you think the weakest areas are before you can decide what areas to experiment with.
    Remember, you can only test one change at a time, otherwise, you will not know which of the changes has an impact.
  3. Get creative with your changes
    Once you have decided what needs changing, you need to come up with creative ideas about the specific changes you want to make. It is helpful, at this stage, to conduct some research on what other companies have done.
    There are lots of blog posts about how other companies have improved their websites thanks to split testing and these can be very useful resources. You should also examine the websites of your most successful competitors.
  4. Choose the right tool
    To do the actual testing, you will need to use a split testing software tool. The tools are designed to show your two versions to randomly selected visitors, making sure that 50% see your old version and 50% see your new one. The tools will collect the results of the test and will draw up conclusions so you can see how well the test has gone.
    To get the tool to work, you may need to add a little bit of code to your website’s files. Sometimes this can be done easily through the use of a plugin and API key, for other’s you may need to add it manually.
    There are a variety of different tools you can use to run your experiment. Most tools are paid for but considering the benefits they can bring, many companies see this as a worthwhile investment. Whilst you can carry out testing for free using Google’s Website Optimizer, the most well-rated tools are those provided by VisualWebsiteOptimizer, Optimizely and Unbounce.
  5. Decide the size of your sample
    To ensure that any test is valid, you need to test it on a reasonably sized sample of visitors. The more people you test it on, the more confidence you can have that any improvements in performance are a result of the changes you have made.
    Your choice of sample size will depend upon the amount of traffic your website gets. For a small website, you really need to test a couple of thousand visitors to get a good level of confidence. However, if you get several thousand visitors a day, this figure may not be high enough to provide the accuracy you need.
  6. Decide how long you will run the test
    Obviously, you want your results as quickly as possible so you can make your changes and move on to the next test. However, you need to give your test some time. For example, if you only have a small number of visitors, it may take you a while to generate the amount of traffic needed to have confidence in your results.
    Even for sites with heavier traffic, it may be worth waiting to see how time affects your results. It could be that your new changes have a delayed impact, causing visitors to revisit and buy a few days later.
  7. Making the changes
    One you have given your test sufficient time and have reached your target sample size, you should be able to demonstrate whether the changes you have implemented have improved your site’s effectiveness. If they have, you can keep the changes and move on to the next test.

If your site hasn’t improved, then you can always try something else and see if that works.


Testing changes to your website before fully implementing them is the most effective way to ensure that they will positively impact how well your site achieves your goals. From reading this article, you should now have a much better understanding of what a/b split testing is, how it is done and how it can benefit your website.

If you are thinking of setting up an online business, check out our range of hosting services. We offer secure hosting with guaranteed uptime and 24/7 technical support, all at affordable prices.

For more information check out our business plans and dedicated WordPress hosting packages.


  • Arjun Shinde

    I'm an experienced digital marketer with expertise in planning, SEO, SEM, and social media. I'm good at creating engaging content and optimising campaigns for a strong online presence.

    View all posts

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.