A/B Testing for Beginners

A/B testing is when two versions of the same thing are compared and used to see which yields the best results. For example, you may have a display advertisement about to launch and this will mean that, if you are doing split testing, you will have two different versions of display ads that will go to two separate groups of people. This allows you to compare both advertisements and move forward with the better performing one. A/B testing goes for any digital marketing technique from website pages, to email marketing, display ads and more. It is effective with PPC campaigns in particular. 

How to conduct A/B testing

First, you need to know what your goals with the marketing campaign are. Is it to drive more traffic to your website? Are you looking to strictly gain more conversions? Trying out a new campaign? Understanding what goals you want to achieve should be essential in any digital marketing campaign, but we just thought we’d give you that gentle reminder.

Let’s start with an example of split testing using a website landing page; however, these principles can be applied to any campaign and we will provide you with another brief example after. First, you should already be aware of the current statistics of your website landing page and how it is performing. 

You may even discover through your research that it isn’t the landing page of your website that is falling behind in performance and decide to pivot over to another page.

Next, you should see if your bounce rate is high or page session duration low and evaluate how your page looks. Could it be the content? Could it be the site responsiveness? Could it be the colours? You want to make a list of elements on the page that could be variables to test. We don’t recommend completely changing the full layout of the page as that can be too drastic just yet. 

When you make changes to your site, you may either choose to keep A as the control and leave it as it is for your changed B to compare against. If you are putting something new out there with split testing, of course, A and B will both be new, different designs and elements. Both designs should be responsive and fast when loading because, if not, then it is clear the better responsive one will top the other. A good UI and UX is also a must!

After you have made your changes and know what you are now comparing, you can set the site pages live. A good idea would be to use two separate identifiable URLs to differentiate the pages. However, Google Analytics offer a feature allowing webpages to be compared which will provide you with a script code to add to your page. Both testing pages should run simultaneously and for a duration that is long enough to collate a significant amount of data.

Once your testing duration has finished, using Google Analytics or another digital marketing analytics tool, you should be able to compare the statistics of both pages and analyse which one did better overall. As a matter of fact, you may even find that for bounce rate B did better, but A outperformed it in session duration. This freedom of information allows you to further conduct more A/B tests and mix elements that will work the best together. Usually, split testing isn’t just one test and then you have results. As a digital marketing agency, we know that digital marketing is never finished and to stay ahead of the competition you need to constantly look for improvements.

As we are a kind bunch of marketers, we figured it would be also quite useful to incorporate a brief example of a different digital marketing campaign that is A/B tested. For this one, we will go with email marketing. If you have an email marketing campaign, you don’t want to change your whole template straight away. Begin by changing the smaller things such as CTAs or imagery. Usually, it can be quite hard to send two versions of an email out so sending them out to the whole list would usually be okay. Depending on what email software you are using, the likes of MailChimp will usually accommodate built-in testing tools. Once you have sent out your edited email, statistics you should aim to look for include click-through rate, open rate, and if they convert on your website through clicking from email. You can then compare your results to previous email marketing results and continue testing different elements of your email marketing campaign.

Why you should perform A/B testing

If you’re wondering whether all that work for testing is worth it. Yes, it is! You can see significant changes in your digital marketing when you test out elements of campaigns to see which performs better and further continue to test whilst implementing better-working elements.

Increased overall website traffic is one of the main benefits of split testing and this goes for any form of digital marketing activity. You will see that when you are testing and then applying the results after, more traffic will drive to your site from using what worked well in the initial testing phase. To then go further and test more will allow you further increase website traffic as it becomes a cycle of trying out what works best. 

Businesses who run an eCommerce or are aiming for increased conversion rate will benefit from A/B testing because they can give two variations of sales funnel to customers and the one that works best could be tweaked in the future and implemented. It is a lot better than leaving your site as it is and not trying out new things to drive more custom. If you don’t do the work, you won’t see the benefits.

One really important advantage of A/B testing is that you will see your website’s bounce rate decrease as you test out different things to engage users more and keep them on your website. Testing different content and layouts could help to decrease the bounce rate.

A/B testing can be very complicated and diverse. It enables multiple changes to be tested and further testing to continue after that to truly allow your site to become optimised with the best quality of traffic.

Still not sure about A/B testing? Get in contact with us today.