A/B Split Testing

The NZS.com A/B Split Testing article contains information and examples on what A/B Split Testing is and when to use it.

Summary

What is A/B Split Testing?


Find out what A/B Split Testing is really all about and how it can be useful to help increase your conversions and ultimately the ROI on your website.

New Zealand Website Resources New Zealand Website Resources New Zealand Website Resources

You may have heard of A/B Split Testing (also referred to as AB Testing, A-B Testing & Split Testing) in the past but wondered what it is really useful for and why it is so often talked about in the context of improving performance and ROI (return on investment) on the web.

Webmasters and website operators consistently find the need to test various things across their websites from advertisements to sales flows to simple word changes on their navigation.

A/B Split Testing is the process of having two or more different pieces of material in a single location on a website which are loaded the same number of times to different users. The results (often clicks) are then compared between the different options presented. Typically there is a baseline or control version with which the other versions are compared.

The classic use of this is different advertisements on a single page to gauge their CTR (click-through rate) however the idea of using A/B Split Testing on single variants within any specific image on a website is becoming more popular as businesses look to find the absolute optimum value from user interaction with their website.

As an example, imagine an image where the user sees the text "Discounted Sales Items" compared to "Clearance Products". Here you have two quite different terms which could easily have the same offer behind them. In testing both of these options in a position on your homepage you may find that will have a much higher CTR than the other to the clearance page.

Moving beyond just the CTR, with A/B Split Testing your main goal should be to prove better conversion to sale on the design your presenting to the customer. In the above example you may have found that 20% more people clicked through to the "Discounted Sales Items" however there was a much higher sales value per product from the customers who clicked on the "Clearance Products" option.

This form of testing is often used by larger companies who are dealing with significant level of traffic where a few percentage points in sales turnover can mean a solid revenue increase. For smaller businesses, this form of testing is useful on a change that impacts their major online revenue generator.

When attempting to increase conversion rates through specific landing page testing a business may offer two or three different core 'buy now' button designs to access their shopping card for a specific advertising campaign that is running to see which resonates best and converts the highest.

To read more about A/B Testing, see the Wikipedia page.

In the end the decision to move forward with this form of testing should be based on your ROI (return on investment). If you know that a 20% increase in sales is going to be worth spending time and resources on testing then you should be looking into A/B Split Testing as one way to bring your conversion up to the next level.


Usability : User Behaviour images from Flickr: Mobile Website, Clean Study Bag and Corporate.

Featured