Website SEO Monitoring

Website Optimization

While it's easy to set up experiments using Website Optimizer, each one will benefit greatly from a bit of preparation and planning before you begin. You'll learn what works best for you after you've run a few experiments, but it's always valuable to learn from others as well. We've provided a list of best practices and things to avoid when setting up your experiments.

• Test a page that gets a lot of traffic
  It's sometimes a good idea to set up test experiments on a fake or low-traffic page to get comfortable with the tool. But when you're ready to do a real test, make sure you pick a page that gets a lot of traffic. Pages with lots of traffic are generally faster to optimize than low traffic pages.
  
  
• Test a few things
  Yes, the power of Website Optimizer is that you can test a bunch of things at the same time and see how they impact each other. However, the more things you test (such as 12 titles, 20 images, and 12 promotional areas) the longer it will take to see results. For your first test, we recommend testing 2 titles and 2 images. You can get fancier once you've completed your first test.
   

  Keep in mind the cardinal rule: as you add sections to your experiment, you'll need either more traffic or more time to get get useful results.

   

• Pick a high volume conversion goal
  Eventually, you may want to test things like order completion if you're an e-commerce site but for your first test, try something that will be more common (e.g., adding an item to a shopping cart, clicking product details page, etc.). If you use a conversion goal that doesn't happen that often, it will take much longer to find the most effective content.
  
  
• Be bold
  If you're promoting an email newsletter sign up and want to test a new headline, don't try a headline like "Gardening Advice Newsletter" if your original title is "Gardening Tips Newsletter." Instead, try testing an alternative like "Stop Killing Your Plants Newsletter" or "Green Thumb e-Newsletter" Once you've completed your first experiment and have gotten results, you can always tweak and try more nuanced alternatives afterwards.
  
  
• Pay attention to your combinations
  Consider designing a series of different experiments if your ideas are likely to clash. As a simple example: a combination that displays red text on a red background. Of course you can expect some combinations in every experiment will be a bad fit. Ideally, though, you want to make sure that every combination is readable.
  
  
• Consider your traffic percentages
  Sites with low traffic and/or a high number of sections and variations will need to include a high percentage of their traffic. If your experiments must minimize customer impact, you will probably need to limit the scope of your experiment; consider trying a phased approach using a sequence of small experiments.
  
  
• Finally, you are the best judge of what will work for your site - all we do is give you the tools to test out your theories.