Using Google Optimizer to improve your conversion rate

By Nathan Donaldson in Design on April 10, 2009

Image 0408 full 2x

It seems that week on week Google introduces new tools without fanfare and it’s easy for these to slip by unnoticed. A year or so ago Google Web Optimizer appeared in the Adwords toolbar. There hadn’t been much of a buzz about it appearing but who could ignore a link that promised to ‘optimize’ your website, and from Google no less.

Taking a closer look at Google Web Optimizer it was clear this is a very useful tool in the web designers toolkit. In a nutshell it allows developers and designers to test variations of design, copy or even whole pages.

Google Web Optimizer enables you to set up tests to determine which copy, design or combination of both is best going to meet your conversion goals.

Setting up and running a test is easy, with Google Web Optimizer taking you through it step by step. You will need a Google Adwords account to access Google Web Optimizer, this is free and there is no need to be running an Adwords campaign. When you login to Google Adwords you will see the Web Optimizer link under Campaign Management in the toolbar.

Website Optomizer is under Campaign Management in your Google Adwords account.
Website Optimizer is under Campaign Management in your Google Adwords account.

There are two choices when setting up tests, AB testing which enables you to test two seperate versions of a page, and multivariate testing which enables testing of changes to multiple areas of a page with multiple variations. Both of these are extremely useful, though the amount of traffic a site receives determines how useful the multivariate testing can be.

The choice of two types of test Google Web Optomizer can run are nicely explained when you start the experiment
The two options are nicely explained when you are starting your experiment.

Once your test has started you can login to Google Web Optimizer at any time and review progress. It is important to note that with the multivariate testing the number of areas and number of variations for each increases the number of variations and consequently the number of visits needed to get a conclusive result. Google suggest only using multivariate tests if your page gets more than 1000 visits per week. Our experience suggests that it is useful to have a great deal more than that!

Results can be viewed while the experiment is running
Results can be viewed while the experiment is running.

Google Web Optimizer is designed for testing copy and design in order to improve conversion rates. We have used it to fine tune the copy on our homepage to increase the number of visitors visiting our portfolio pages.

When redesigning an interactive tool for a client it was not clear the exact text we should use for the call to action. We suggested that once we had the site live we could run a Google Web Optimizer test to see if small changes to the text could improve the conversion rate. We tested two new variations of the call to action in two seperate places. This provides nine combinations (including the originals).

We ran the test for two weeks and then analysed the results. It was clear that a number of combinations were doing worse than the original while a couple of combinations were doing better. At this stage we did not have enough data to conclusively decide which was working the best. We turned off the under performing combinations so that we could more quickly see which combination was working the best.

Variations that aren't performing can be stopped
Variations that aren't performing can be stopped sending more traffic to the remaining variations.

With the best of the new combinations currently showing a 10% increase in conversion performance it won’t be long before we have a clear winner. The total time to setup the test was around 4 hours.

Google Web Optimizer is quick and easy to setup, delivers useful data with a minimum of fuss. Optimizer is an essential tool in improving conversion for any website and who doesn’t need that?