Testing is one of the most important aspects of creating new web design. If you have a current website and think you can improve upon it, how shall you go about it? You should test new designs against your old design and see which one performs better.
There are two types of website testing that are popular and recognized by most industry experts: A/B testing and multivariate testing. Which one is right for you?
In general, it depends. However, I think A/B testing is appropriate for most circumstances.
A/B testing is where you take one component of your web page and you test it against a new version of it. For instance, you take your page headline and tweak it slightly then present an A/B test to see which one users like the most. The A version is usually the current web page published. The B version is the one with the new headline. You don’t test any other components on the page.
Multivariate testing allows you to test multiple components at the same time. You can add a new headline, move your lead photo from the right to the left side of the page, and make your sidebar wide and narrow.
Testing multiple variables at the same time can give you an idea about different versions of your page, but it won’t necessarily tell you which variables users are attracted to. In the above scenario, for instance, users might like version B of your web page more than version A, but do they like it because of the headline, the image or the sidebar? All you can is they like the combination of elements better.
This is why I recommend A/B testing over the multivariate testing most of the time. Still, it’s OK to use multivariate testing because the end result is a better website that visitors love.