Let’s imagine for a moment you run a web site, and that you make money from this web site somehow (for example selling time tracking software) – how can you improve the web site to increase your profits?
One very common method is to blindly make changes to the site and hope for the best, sometimes it even work, but there got to be a better way.
And that better way is A/B testing or split testing – in this method you show half the site visitors your old design while the other half get the new design – then you compare the actions of visitors in the first group with the actions of visitors in the second group and see which design is better.
This is similar to how new medication and medical procedures are tested before they get approved, it gives you scientific, evidence based and clear results.
I’ve installed split testing software on this site and I’m going to document the tests I run and what I’m learning from them here.
In the next post I’m going to talk a little about some important concepts and considerations in A/B testing.
You can also split test with more than two options or run multivariate tests where you test multiple parts of the page instead of two complete designs – but I’m not going to talk about those here any time soon.
posted @ Wednesday, February 24, 2010 5:36 PM