Corner Pocket

New, Total Online Web Management

Free UK domain name

Free email addresses

Introduction to A/B testing

A/B testing

A/B testing (or as it is sometimes known, split testing) is comparing two different versions of the same web page to determine the one that performs the best.

Every business website needs to make the most of their visitors "converting" them from just a a visitor to a customer. This is known as the "conversion rate" and the higher the better. In the world of website marketing many small, almost imperceivable changes can have a huge impact on conversion rates and it is very difficult to determine which changes can improve these rates.

This is where A/B testing comes into play, by trying out different things you can compare the differences using the quantifiable results. At is simplest level A/B testing shows the first version of the page for a set period of time or for a set number of views and records how many sales or conversions take place during this period using website analytics. The second version then undergoes the exact same period or views and then a direct comparison can be made - the one with the most conversions is the better performer.

Many different items can be considered for A/B testing including:

  • Colours or text of call to action buttons
  • Headings
  • Images
  • Testimonials and social proof
  • Awards
  • Position of calls to action or links

Don't let this list limit you, almost anything can be changed and compared using this kind of testing but you should follow a set scientific methodology. Examples of these would be:

  • The test should provide a result for some hypothesis: for example, changing call to action button to green will increase the number of clicks
  • The result should be quantifiable, use website statistics such as Google Analytics to record the visitor interaction
  • Work out a reasonable amount of time to test your two designs
  • Act on the results, even if the results were not as expected make sure you act on them

A/B testing is only a starting point, these test can be repeated on all of your pages to make the the best converting pages but you will only be able to make a change to a single page element. You can even go further with making multiple changes using multivariate testing.


Get in touch

Contact us with a question or to get a no obligation quote.