Main Menu

 

How to Start your First Google Website Optimizer Experiment

Here's our top 10 tips on starting your first successful Google Website Optimizer experiment:

  1. The Test Page - decide on the “Test Page” you want to optimise. Good places to start would be your homepage, registration page or payment processing page.
  2. The Conversion Page - Keep it Simple Stupid (KISS) Choose your “Conversion Page” carefully and keep it simple. Choose an easy multivariate test with a straight forward conversion such as successful registrations or payment completions. Choose a high volume conversion goal.
  3. Choose Page Sections - look at your test page and identify all the sections on this page that might drive conversions. Things like page titles, buttons (position, shape, colour, call to action, size) images and marketing copy. You should consider links to other persuasive content such as web pages with information on delivery details, returns policies, security and privacy notices. Also include marketing strap lines and other calls to action. You will produce a long list of potential drivers to conversion.
  4. Priortise Page Sections - to keep things manageable chose the 3 page sections that you believe to be most important for conversion.
  5. Create Section Variations - create variations for each of these page sections – 3 page sections with 5 variations will give you 125 different combinations to test.
  6. Calculate Experiment Duration - you can use the Google Website Optimiser Duration Calculator which will help you estimate the potential duration of your experiment.
  7. Insert the Google Website Optimizer Control, Tracking, Page Section and Conversion Scripts and Validate as outlined in Stage 1 to 3 in the Google Website Optimizer Tool as set out
  8. Launch - launch your experiment remebering to set the percentage of traffic you wish to send through the experiment in Settings
  9. Collected data may take up to a day to become available in reports.
  10. Reports - Google Website Optimizer reports will give you information on the performance of each of your combinations and individual page sections in terms of Estimated Conversion Rate, Chance to Beat the Original, Observed Improvement. When enough combinations have been served Google Website Optimizer will identify the high confidence (p<0.05) winning combination. The tool will also indicate which of your page sections are most significant in driving conversion as well.

 

That’s you first experiment completed and you will probably have plenty of ideas about your next. If you found that the call to action text on the “Sign Up Now” button is your most important driver of registrations, you’ll probably want to experiment with colour, shape and size. You'll probably have numerous other page sections you want to experiment with too - the process of web site optimisation is continuous.