How A/B Testing Can Help You Figure out What Your Customers Want


Everyone’s different and unique in their own way, we all know that. But when it comes to setting up a retail site or launching a new online advertising campaign, this doesn’t really make things any easier.

You can take a generic approach with regards to the whole process, but the point is that somewhere down the line, you’re going to have to test the waters and figure out what works best with your visitors.

Some marketers and website managers get too hung up on the issue of generating loads of traffic. Don’t get me wrong, site traffic is good. However, what’s also as important as traffic generation is to make sure that you’re making the most of this traffic by converting them into paying customers. That’s where A/B testing comes in.

AB Testing

Image credit: arztsamui

What is A/B Testing?

A/B testing is a relatively simple technique that will help you figure out which content and design will generate more conversions and what will entice more customers to purchase more.
One can view A/B testing as a systematic approach to determine the effects that manipulating a certain variable has against a control.

A/B Testing On Your Website

Let’s consider the technique as it’s applied to your web page. On a basic website, there are four main parts: the header, the navigation menus, the content, and the footer. You might have already followed measures to optimize your site in terms of SEO and site loading time—and this is what we meant with generic earlier on. Applying A/B testing will be like putting your finished and optimized website to the test and seeing what type of specific text, content, or images will hit it off with your user base.

What will catch the eye of your site visitors will differ. For some, it could be the images that you’re using. For others, it might be the text or the headlines on your page. Regardless of all the differences, there are also bound to be similarities. And the answer to what works lies in these similarities.

A/B Testing On Your Website


Image credit: Salvatore Vuono

How Do You Do A/B Testing?

First of all, you have to decide which element you’re going to be testing. You might have a lot of things that you may want to test out at a time, especially during your first few rounds of testing. However, for later tests, it’s better to modify or change only one element at a time so you will know the reason behind the differences in conversion or the results that you were able to gather.

For example, you could try serving two versions of a banner ad on your web page. One could feature images of the products being put on sale, while the other could have the goods in the background while the word “Sale!” takes center stage.

You can then use a script on your server to display the two versions of the banner ad randomly to your site visitors for a specific period of time. When the duration for your testing is over, you can review the results and see which campaign was able to garner more clicks.

A helpful tool that you can use to get started is Google’s Website Optimizer. It’s a free resource you can access online to conduct basic A/B testing and multivariate testing on your website.

What To Make Of The Results From A/B Testing

When you have gathered sufficient data from your tests, it’s time to interpret the results. Note that there are many ways you can interpret what you’ve gathered. In the case of the previous example, you could gauge which ad was better in terms of the number of clicks, the total sales each was able to lead to or generate, and so on. After you’ve collated the results, then you can use the ad format that was able to get you better results for your succeeding campaigns.

But the testing doesn’t end there. Although you’ve already got the banner ads down, there remains a whole lot of other elements on your web page that can be tested: the positioning of the ads, the text accompanying the actual ads, the type of color theme of your website, and so on and so forth. After you’re done testing these out, you can go deeper and modify elements within these elements. For example, for headers, you can test out what type of positioning, size, or color scheme would work best for your site visitors.

With an AB Test, you will be able to change certain parts of your site in order to improve conversions and increase sales, one element at a time.

About the Author:

Founder and President Maxymiser Inc. Mark Simpson is Founder and President of Maxymiser, the global expert in multariate testing and behavioral targeting solutions. Mark introduced Maxymiser’s revolutionary Customer Experience Optimization platform to the market in 2006, which improves website and mobile conversion rates, revenue and overall brand experience for all commercial industries. Headquartered in New York and London, Maxymiser works with brands such as Finish Line, Hertz, Teleflora and Time Warner Cable, to instill cultures of conversion best practice, increase brand loyalty and improve revenues.
  • Dedicated Server Hosting

    Your suggestion for banner ad on website impressed me specifically. To figure out conversions usually I use Google Analytics to check the flow of visitors on my website, however, next time I will consider A/B testing technique as well. Thanks for sharing this new methodology.

    Paul Lopez

  • Cathy Dutton

     Hi Mark

    There are some useful tips in this
    tutorial, I agree that it makes sense to only test one small aspect
    at a time, this can make for a very long process though particularly
    if an appropriate amount of time is allowed for each test. I think as
    mentioned by in the previous comment using a mixture of a/b testing
    and Google analytics would be beneficial.


  • web development

    A/B testing should be included in the web development timeline. What is the use of website if this is not patronized by the intended users? They can just  click the exit button. On the other hand, I also recommend to everyone Google’s Website Optimizer.

  • 99designs vs Crowdspring

    Nice article. Great to learn about customer needs. Thanks for providing such a useful tips.

  • hawaii web design

    A/B testing is very important and needs to be done every time, whats the point of having a great looking full of features website if they end users cant figure it out?

  • Web Designers Sydney

    I di not know about A/B testing system as i consider myself a student of Web Design World. I am very happy to read this new technique. Your site is awsome and i would love to subscribe to your RSS feed.

  • Facebook Statistics

    Brilliant post. Good to read some useful things about marketing options which really in favor of customers. Thanks

  • natural crystal x

    Yups. Great to learn about customer wants. nobody have same perspective

  • Day 2 Day Online Printing

    The place I am working at right now is currently doing A/B testing. It takes some time and we are currently gathering as much data. Its not as much as what they customers needs but plays on more of their psyche and what they want to see. 

  • Agence Solution

    A/B testing is a very hard work….i am not sure that its a good idea for all project…

  • Imran Haider

    Yups. i know that what my customers want thanks Simpson

  • Web Design

    Interesting and helpful article for me. Nice strategic tips shared. Thanks

  • Website Ideas

    Great articles. Thanks for sharing.

  • Nika@Buy SoundCloud Followers

    Excellent post! Thanks so much for sharing your ideas. Just a quick question, will it be fine if do A/B testing daily? I mean, are there any effects if you constantly change something on your site, like daily?

  • pnr status

    Nice and cool post. Your idea to use A/B testing to a website is really awesome. Thanks.

  • wall_wallpaper