How Important Is A/B Testing?


How Important Is A/B Testing?

In a word – very.

End of blog……….


No, of course not. I’m merely being flippant. But the importance of properly conducted A/B testing should not be questioned.

Designers may think that they have all the answers, but the reality is that we don’t, and in fact we need meaningful input from test subjects as much as anyone else.

In our heads, we think we know everything there is to know about design, and that what we produce will work out there on the web. But, just because we’ve got an artistic eye, doesn’t always mean that we have an objective one. After all, art, naturally, is a subjective endeavour.

Thankfully, though, we do not have to solely rely on our best judgement – for there are tests that we can do to try and make our designs as profitable as possible for our clients. I’m talking of course about A/B testing – sometimes referred to as split testing – and I’d like to highlight the importance of it in this blog.

Firstly – What Is A/B Testing?

A good question to get us going. In web design, A/B testing refers to the creation of two variations of a single design element. When a site goes live, these two variations are substituted for one another for certain lengths of time, and then we use analytics to measure the performance results.

For example – a website might have a CTA (call to action) button with the words “BUY NOW” written on it. We may be interested to know whether a green button or a red button will entice more people to click. So we run an A/B test – on week 1 we go with the green button, on week 2 we go with the red button. At the end of the 2 weeks we will have results that we can compare which will inform our ultimate decision.

Without running an A/B test in this manner and simply opting for, say, green, as we believe this connotes “go” or “proceed” or what have you, then we would never actually be certain that this was our best choice.

What Should Be Put Through An A/B Test?

Anything and everything.

There isn’t a single element of a web page that wouldn’t benefit from A/B testing – from headlines to images to backgrounds, to CTAs to copy. The whole idea of split testing is to optimise your website so that it produces the most profitable and convertible results. Just because you or your designer likes pink as the background colour for your site, doesn’t mean that your visitors will. Indeed, your visitors might not even be aware of their unconscious biases towards the colour pink, but, by running a few A/B tests, you can very quickly find out that in fact a grey background improves your bounce rate and conversions.

A Word Of Warning

As great and revealing as A/B testing may seem on the surface, you do still need to exercise some caution and control when embarking on the process if you don’t want to end up with skewed results.

Indeed, there is quite a lot of material that can be found online that warns of the dangers of improper split testing.

In order to avoid any such bad results, then you need to make sure of a couple of things.

Firstly, you need to make sure that you run a good sized sample for the test – i.e. that a significant number of users actually participate (10 people isn’t enough, nor is 100 – 1000 should be your minimum). Secondly, you need to make sure that your tests run for adequate lengths of time. And thirdly you need to make sure that you’re using some reliable split testing tools.

3 Great A/B Testing Tools

#1. Unbounce

Unbounce is the split testing tool to use for building and testing landing pages – arguably the most important pages for creating conversions. It’s a very simple to use tool, and lets you change any element of your design and integrates brilliantly with Awber, Hubspot, Mailchimp, CampaignMonitor, ConstantContact, Zoho, Salesforce, and many more besides.

#2. Google Analytics Experiments

Google Analytics Experiments is a very useful tool for A/B testing. Google, as you might expect, has great power over what people are returned with when they use the search engine. And with Google Analytics Experiments you can A/B test as many as 10 full versions of a single page, each of which will be delivered to users from a separate URL. You can even define what percentage of users are to be included in each of your experiments.

#3. A/Bingo

A/Bingo is put to work as a plugin on the backend of your site. It’s a Ruby on Rails split testing framework, and uses just a single line of code to detect behavioural differences in users. Ideally, it is an extremely fast tool, meaning that it barely interferes with page load times and is generally a very unobtrusive means of testing some of the finer details of your website.

What tools do you use for A/B testing? Let us know in the comments below.


No Comments

Post A Comment