Friday, 15 March 2013

A/B Testing: What it is and Why you Need it

There’s a lot more to the world of web optimization than SEO and one of the more important areas that’s often overlooked and left in the wake optimization of Google’s search results, is conversion optimization. 

Conversion optimization is essentially the structure and design of a site and how well it caters for and motivates users to convert, sign up or purchase a product. So, how can you improve conversion rates – well A/B testing plays a large part.

What is A/B testing?

A/B testing is a lot simpler than it sounds. Essentially, testing involves showing users two versions of the site – one the original and one a variant with a tweak. These variations are each displayed to a set number of people and the responses in terms of conversions, bounce rates and other on-site actions will be logged and evaluated. These actions or metrics will be set pre-test, with specific ones used and compared to allow a solid conclusion to be made. 

If one conclusion is deemed as offering a more positive outcome than the other and seems to perform to a better level, generally a change is made and an implementation takes place. 

Customer Needs

A/B testing allows you to quickly get to grips with the preferences of your niche and allow you to offer the best line up possible of products or services in the most attractive way conceivable. Though using specific metrics it’s easier for you to discern what you customer wants, limiting the incidence of trial and error.

Bounce Rates

If bounce rates are high your conversions are generally in trouble. If users come to your page and leave just as quickly, it’s generally a sign that something is wrong and you are not giving the information leads want. This will of course make for fewer conversions and sales and so a less successful website. Good A/B testing can lower bounce rates and encourage people to spend more time on site and convert to leads more often.


One of the most important things with A/B testing is to give it enough time. Like any sort of experiment you need a large enough test to get an accurate result. Testing overnight and garnering information from only a few results may not provide accurate results. The more time you run the test the lower the chance of a skewed result. 

With A/B testing it’s also possible to test more than one thing at once. For instance, if you have more than two variations, then you can perform three separate tests at once – so, A v B, B vs C and C vs A testing. This may take a slightly longer period of time and is a little more complex, but can offer better results If you wish to test a number of variables – this is known as multivariate testing as there is more than an A and B factor involved.

A/B testing can really make a significant difference to your bottom line and be of significant benefit to your site and online work.

No comments:

Post a Comment