- 
November 2, 2006
Google Optimizer Test
Ok, so we've used our analytics software to pinpoint problem pages of our business website. But where do we go from there?
* Do we patiently perform hundreds of separate A/B landing page tests?
* Do we hire a third party to conduct expensive multivariate testing?
* How can we be sure that the ugly red button in the middle of our page isn't scaring away our customers?
Well, as it turns out, I just finished implementing a tool on our website which answers all of these questions and more. Quite simply, our visitors should be the ones that determine the content of our website using multivariate testing, and Google's new Website Optimizer seems primed to tackle any website testing problem.
This free tool (if you have an existing Google AdWords account) can let us find out which combination of headline, image, and paragraph text (to name the three sections we tested in our experiment) will give us that increased conversion rate.
Is the lack of an image holding our success rate back? The program will tell us that too, as we can test the success/failure of individual website components (We can, for example, test the existence of a section against its absence) to see whether or not a certain part of our website is actually hurting our conversion rate.
Which version of the landing page will have the higher conversion rate?
The implementation was easy and is well-documented by Google. The combinations we created only took a few minutes to set up.
For example, with just 3 variations (including the original) of a heading, an image, and an action button, we now have an experiment that can test up to 27 different versions of our landing page (a free Google Analytics newsletter offer)!
The results will depend on a few things: The amount of traffic our website generates, our current conversion rate, the relevance of the sections we are testing, and the effectiveness of our experiment's design. A tool is only as effective as the person wielding it.
After an appropriate amount of time has passed, and we've collected enough data, we will have more information on which combinations have the highest chance of increasing our conversion rate. Then, if the probability is high enough, we can go ahead and update our website code with the version that performed the best, and this can all be done automatically through the Optimizer.
I'm very excited about the potential power of Google Website Optimizer, and am anxiously waiting for the results to collect and give me a valid report. We'd love to share our experience implementing and using the tool with you, and you can sign up for our free 40-minute webinar here. If you want to see the page we are testing, check out our newsletter sign-up page and sign up for our free online newsletter. You can see Google's documentation and demos here.
