Google Website Optimizer A, B or Multivariate?

So, you’ve decided to try out some tests with Google Website Optimizer. You log in to your Adwords account and go to the Optimizer section…

Lo and behold, the interface asks you if you would like to set up and A/B test or Multivariate. “Which one should I choose?” you think.

Well, it depends on what your goals are, what kind of site you are running, and what you want to find out. Lets go over the two types of test:

A/B Testing

This is the simpler of the two tests. It involves setting up two separate versions of a page and testing them against each other. You may want to change one or two elements, or look at testing two different layouts to see which performs the best.

A/B is also simpler to set up. You simply produce the two pages, set up the experiment in GWO, and add code to both of the pages. Website Optimizer does the rest, even distributing traffic evenly between the two experiment pages.

You can also set up experiments in this way with more than two variations, so you can have A/B/C, A/B/C/D experiments and so on.

Multivariate Testing

Multivariate testing is a little more complex. It allows you to test multiple versions of elements on a page. For example, you can set up a test which has two variations of a headline and two different buy now buttons. Website Optimizer will then tell you which combination of the four different elements worked together for the best result.

Multivariate tests are more tricky to set up. What you have to do is determine the page elements that you wish to test and load the actual code for the variations into GWO. Once you have done this, you need to insert the code that GWO provides into the correct portions of the page. It is very important to double check the variations you have set up and test to make sure the pages work properly.

So Which One Should I Use?

Well there are a number of factors to consider?

1. Technical know-how. If you are starting out with GWO for the first time, or you don’t build web pages yourself it would be a good idea to start with an A/B test as they are a lot easier to set up.
2. What are your goals? If you are looking to find out something simple like whether one background colour works better than another for conversions, you may be better off with an A/B test. However if you have more complex goals and are looking to test multiple theories on what will work best on your site, it may be a good idea to go with Multivariate.
3. How much traffic do you have? Multivariate tests typically need more traffic pushed through them to determine a winning combination than A/B tests do to determine a winning page, so if you have low traffic numbers start with an A/B test.

Both methods are very powerful for proving theories on how your web pages work, and both can be used to great effect to increase the performance of your site. You will need to be patient though. You will only start seeing a true idea of how your pages perform after a goodly amount of time. We usually look at two weeks’ data as a minimum.

You will end up learning a lot about your visitors and your market and hopefully boosting your revenues, but remember that Website Optimization is an ongoing process. It is wise to plan well and use a conversion methodology such as this one.