Multivariate and A/B Testing with Website Optimizer

So how do you improve your conversion rate? And how can Website Optimizer help?

Most of the time, we think we know what our customers want. But the truth is, we website owners are the worst judges. That's because we're so close to the product and the site. We understand them too well to know what the average visitor doesn't understand.

One excellent way to learn what converts best is to split the traffic between a few pages, each of which shows a separate version. This is known as A/B...N testing, or "split" testing. Alternatively, you can do multivariate testing. Multivariate testing (MVT) tests different combinations on the same page at the same time. With Google's Website Optimizer, even the smallest of companies can afford to do multivariate testing. (After all, the software is free.)

So, for example, you can test the area on your page that features a headline (with multiple headlines to choose from); at the same time, you might test a picture, a button, some text. Each time, the multivariate software serves up a different combination until your visitors have seen enough combinations for you to know "the winner." You get to decide how the winner is determined (usually, companies determine a winner based on how often a success, or "thank you very much" page, is served up.) You may learn that the winner is the long headline with the realistic picture and the red button. Or maybe it will be the short headline with the black and white picture and the red button.

(If the concept of multivariate testing is confusing, you might love this article about Mrs. Peacock in the Conservatory with the candlestick.)

Here are some of the cool things you can do with multivariate testing:

• Test multiple elements (think, places on your page -- like a title, or a picture), and for each element, test multiple versions.
• Pull more than one page into your testing scenario
• Test an offer, like "Free shipping", and have it included on every page, including the shopping cart
• Determine success based on multiple events. Lots of companies consider a success to be a thank you page (after all, that's when the customer made a purchase or signed up for a list), but success to you might just be progression to a second page.
• Run different tests at the same time, without the tests polluting each other
• Rerun the top two versions of a test to see which one is the absolute winner when all the visitors are split into just those two groups.

The hardest part of testing is knowing which pages to test, and what to test. You might already have done this analysis, and if you have, that's great. We're happy to help you set up your Website Optimizer test. (Or, go do it yourself! It's pretty easy, and who wants to pay a consultant if they don't have to?)

On the other hand, you might not be sure where to start. We usually like to see customers do an analytics assessment (we're also a Google Analytics Authorized Consultant) and user testing first, so that we understand what the likely issues are. Then, when we do user testing, we really know what to test...

So give us a call (412-343-3692), send us email, or fill out the form on the right. We can start to talk about what you need to test and how to improve your conversion rate.