You've built your website for a reason, and in most cases, that is to sell something. But how can you be sure the website you have designed is converting all those visitors into actual sales? Maybe it's time to do some testing.

Every website exists for a specific purpose. No, that purpose isn’t to get more visitors. Rather the purpose is directly related to business goals. These business goals -- selling products, collecting leads, signing up members or simply creating a brand -- ultimately determine if the website justifies its existence or not. If you think about it, how useful is a website if it does a shoddy job in converting visitors into customers.

Traditionally, 99% of the efforts online are applied to get more visitors to your website. Though nothing is wrong with this approach, it definitely leaves a lot on the table. Through marketing or SEO efforts, once you get visitors to your website, the job of your website is to sell products or services and that is what ultimately matters -- not the total number of visitors.

Essentially, conversion rate -- the percentage of visitors who end up completing website goals -- is the most important parameter determining success of online efforts.

What is A/B testing?

Thankfully, conversion rate is a metric that you can optimize.  And the best methodology for that is A/B split testing. The term may sound cryptic but it is pretty simple in reality.

In A/B split testing you design two or more variations (hence the name A/B) of your website and visitor traffic is equally split amongst them. Traffic is diverted to the variations in random fashion and then it is observed which variation resulted in maximum conversions.  

The variations can differ from the original design in either a subtle way (see this case study) or you can have completely different designs (see this case study).

An Example of A/B testing

Take for example a simple website that sells mobile phones. The company decides to test their shopping cart button text “Buy Now” against a variation “Purchase Today! Only 10 left”. If they get 1000 visitors a day, 500 will see the website with the “Buy Now” version and 500 will see the “Purchase Today” version.

Depending on how many visitors end up buying in each variation, you can conclude whether your idea of creating urgency really worked or not.

A/B testing case study

A/B testing is not a hyped up phrase. Because of their commitment to A/B testing, companies such as AquaSoft have seen as much as 20% increase in bottom-line sales. You can see screenshots below of the A/B variations they tested.


Version A - default


Version B - Redesigned

As compared to version A, version B has a modern design, more trust-building elements and a prominent ‘Download’ button. If they had simply changed the design without testing, they would have never quantified if the redesign increased sales or some other external factor (such as increased traffic).

In this A/B testing case, key points to note are that they did not change products they sell, they did not add new features and they did not devise any new policies. Only the design of the page was tested and it made an everlasting impact in their sales. No other methodology can claim to have a better ROI than A/B testing.

Why should you try A/B testing today

Unlike doing changes on website for improvement based on hunch or expert opinion, A/B testing uses your visitors’ clicks to determine what works and what doesn’t. This methodology is currently used by all kinds of marketers to optimize their landing pages, homepage, offerings, Adword campaigns, etc.

Some even use A/B testing to discover the right price point for their products. No matter what kind of website you have, if you haven’t tried A/B testing, it is a great time to catch up with most successful marketers and add this sophisticated tool to your arsenal.