How to A/B test a landing page

on August 26, 2015 at 18:47


If you're new to e-commerce, or even if you're not, you might be mildly mystified by the concept of A/B testing. Wonder no more.

A/B testing sounds scientific and advanced, but really it's a simple matter of changing something on a page and then seeing if it has a positive or negative result. That's it.

Of course, it's more advanced than that in practice, but that is the nub of A/B testing. Its value, though, is immense and it can revolutionise your business.

Simply changing the copy on a button has been found by some companies to increase conversions by 200%. So, if you apply changes in an organised fashion, test them, discard what doesn't work and keep what does, then you can send your website conversion rate through the roof.

Always run one test at a time, you can't rush these things and changing several variables at once will simply give you nonsensical results. Thorough website conversion rate optimisation is methodical and scientific.

You can and should test entire elements, design a totally different call-to-action button with different copy, different colours and a different position. It's good to start like this as this kind of test will produce the biggest results, then you can tweak the fine details like just changing the colour.

Measure the changes the whole way down the funnel, too. A change on the landing page might boost the conversion rate in terms of people clicking through, but does it affect the actual sales? Every site and every situation is different and it's the sales that count, you might find those additional click-throughs do nothing for you, so it's time to look down the funnel or simply discard the change.

Regularly revert to the control. There are so many variables that might give you false readings that the only way you can truly be sure that your conversion rate is increasing due to the changes you have made is to go back to the control regularly.

Also test your changes at the same time, adopt a scientific approach and keep the variables to a minimum. It's the best way to ensure that your A/B test is effective.

 

Last modified: Tuesday, 01-Dec-2015 09:57:02 GMT