Last week Google made one of the most awesome upgrades to Google Website Optimizer! If you’ve used GWO much at all, you’ll agree that this upgrade will make your life much easier. I’ve outlined the primary changes below. If you want to read more about the upgrade directly from Google, visit the Official Google Website Optimizer Blog site.

There are 3 major changes in the upgrade that have me jumping up and down with joy:

1. You can now validate an A/B experiment by uploading a local file (this was needed badly), enough said;
2. You can disable a combination in an active experiment; and
3. Better color representation of a combination’s performance confidence.

The image in Figure 1 below is from the new GWO report. Click on the image to view a larger version. The biggest thing to note about the changes is that now you can disable a variation that is performing badly and concentrate on the variations that are doing very well. Just click on the checkbox (item 1) then on the disable button (item 2). This option will allow GWO to produce more definitive results much quicker and without having to stop, copy, and re-start the experiment. That is so awesome that I’m almost plum giddy!

The final major improvement to the report is the color representation of how well a combination is performing (item 3). Although Combination 2 is performing much better than 1 or the original, the graph will not turn green until it is truly a high confidence winner.

Go ahead and give it a try for yourself!

The number one question that seems to be asked is “What should I test on my site”. Ultimately, that depends on the type of site you have (i.e. eCommerce, brochure, etc.) and the goals of your site. When I’m analyzing a site for website optimization strategies, the first thing that I ask is “What is the primary goal of the site?” (i.e. lead generation, purchase a product, sign up for a newsletter, etc.) I believe that this is the most important question to know the answer to because it dictates all subsequent actions that you should take. Let’s look at one example answer to this question and how it affects the testing that should be done: “I want to generate qualified leads that I can use in email marketing campaigns.” From this statement, you can deduce that the best place to generate these leads is through some type of sign-up form.

A Sample Contact Form Analysis

Let’s assume that Google Analytics (GA) is already installed on the site. Let’s also assume that a “Contact Form” goal has been established. The image on the right is a sample funnel visualization for the contact form. Click on the image to enlarge it and see the detail. Notice that 201 visitors landed on the contact page and only 13% completed the first step in the funnel and moved on to the contact us form. This generally means that you have lost 77% of the potential new leads. Scary isn’t it? Now notice that of the 13% that completed the first step, only 40% actually completed the form. Overall, only 7.9% of all visitors that entered the funnel actually completed it. Now what?

Since 77% of visitors are abandoning the funnel from the Contact Page, that’s the first place to look to perform testing. The image below shows our fictitious contact page. Notice that the contact page also includes a phone number and mailing address. It’s very difficult (if not impossible) to identify how many of the 174 visitors that didn’t continue into the funnel actually called or send a snail mail in lieu of filling out the form. Or if any of them contacted the company at all.

There are many tests that could be performed on this page such as moving the form here with and without the phone numbers and addresses below the form; putting the form on this page and providing a link to the additional information, especially if the primary goal is to get visitors to fill out the form; or this page could contain 2 links, one to the contact form and the other to the additional information.

Use your imagination and see how many variations you can come up with. Share them here as an online exercise. Happy Testing!