You could be the best marketer in the world and have hundreds of thousands of visitors to your website, but what really matters is your conversion rate!
Let’s do a quick calculation: a conversion rate of 0.5% for 100,000 equates to 500 sales. A 5% conversion rate for 10,000 visitors amounts to exactly the same.
Which of these do you think is easiest to achieve?
If you apply the advice given in this article and implement an A/B testing strategy, there’s a very good chance you’ll see your conversion rate increase in this way.
Why do A/B testing?
You might think you’ve got a brilliant marketing department and that with such a good team around you, there’s no chance you can go wrong when it comes to your website or newsletter design.
However, even the best team in the world still makes mistakes, believe us. And the key strength of A/B testing, also known as Split testing, is that it helps you to change your perspective. You may have designed or approved the site, but you’re not the one who’s going to notice the details that work or that don’t; you’re simply too invested in it.
A/B tests will help you to understand which elements on a website attract customers’ attention, which calls-to-action work, and how far users scroll down the page, compared to what you might have imagined.
So what exactly are these A/B tests?
An A/B test is a way to test the design of your website or newsletters and see what works vest. To do this, there are some programs that can initially help you to adapt certain elements of your website and show them to half of your audience, while the other half continues to see the site as it was before the modifications were made, constituting the “test” group.
There are some software tools available that will help you to achieve this, and we’ll discuss them below. What do you want to test? All the elements that allow you to make sales on a landing page or in an email. The aim is to drastically increase your conversion rate by focusing on what works!
In a newsletter
Let’s start here, because the number of elements in a newsletter is necessarily lower. You can modify 3 main things: the subject of the email, the call-to-action, and the offer itself. Of course, you can also test different types of copy, that is the text itself, but the 3 elements mentioned above are likely to create more of an impact and do so with greater ease. It’s the 80/20 rule.
Let’s talk about the offer. For us, this could well be the most important point. You could offer a business gift for any purchase made in the first group, and a coupon for a 10% reduction in the second case. At this point, it’s essential to build two landing pages, both of which correspond to the different CTAs (calls-to-action) of your emails. By doing this, you will get clear and distinct results about what works and what doesn’t, and can make the necessary decisions.
Here’s another example for CTA buttons: here’s a graph taken from the website Quicksprout, presenting the keywords which work best. As you can see, not all the words are equal in terms of their conversion rate, with “free” accompanied by a verb proving to be the best source of clicks.
On your website
Your website and landing pages are, of course, the key places to conduct A/B tests. The main things you can test here are the titles, CTA buttons, sales texts or product descriptions, and the graphs you use to highlight the benefits offered by your products.
If you can compare the performance of your emails and adjoining landing pages in an artisanal manner, it’s best to use a dedicated app to manage the traffic towards your website and guide it towards the different versions.
Some of the best known competitors are of course Crazy Egg, founded by Neil Patel, Hubspot or Google Analytics. Crazy Egg, for example, allows you to see how far users have scrolled, and offers you “heat maps” which reveal the points on your website that generate more or fewer clicks.
In all cases, you can choose the size of the test samples and make modifications easily, by integrating them into your site.
In your advertising campaigns on Google
Finally, Google has several functions which allow you to have different copies of your Google Ads campaigns, so you can see which formula performs best.
How do you rate your A/B testing ?
Testing is a good idea. But all it takes is a few mistakes and the means you’ve invested will bring little in the way of results. Here’s some useful advice to help you avoid this situation.
First, remember to test regularly. Many people come to us asking why their tests don’t work or why they don’t see the expected increase in their conversion rate. This may simply be because they haven’t continued to test. Indeed, the effectiveness of certain elements changes with time and context.
Therefore, remember to test regularly.
What’s more, testing samples that are too small also won’t bring meaningful results. If you want your tests to teach you anything, you need volume.
The second thing that could bring different results is wanting to test excessively. As always, refocus your efforts on what counts the most (and it’s not always the shade of blue on your CTA button: it might be its colour, but not the shade). It could also be the signup procedure on your website, or the way to confirm the shopping cart. To know which the key variables are to analyse, take a look at the problems on your website, such as where people stop and why they don’t go all the way through to make a purchase. You could even ask them, perhaps via an email reminding those who have put products in their basket but not completed a purchase.
Finally, the third mistake concerns received ideas: what works for other businesses won’t necessarily work for you. The site webmarketingROI gives the example of two landing pages, one of which has a video, the prime interactive medium, but does not have an optimal conversion rate. Overcrowding a landing page or making it illegible isn’t necessarily a wise strategy either.
Source : WebmarketingRoi – La page de gauche convertit davantage, même sans la vidéo.
Conclusion
As A/B tests are multiplying in number, it would be a shame not to implement them. However, to avoid wasting time and valuable resources, it’s vital to conduct them in the right way, paying close attention to the classic mistakes.
For our own Google Ads campaigns, we also conduct split testing with version A and version B of our sales page.
Which do you think converts best?
References
Quicksprout guide on AB testing
AB testing: here’s what you could be doing wrong
CrazyEgg
Neil Patel: 9 tests to double your conversions
Guillaume Duckerts is a Belgian writer who specialises in information and communication technologies.
