Google Website Optimizer

You signed up for Google’s free multivariate or A/B testing platform, and I’d like to welcome you to the world of conversion rate optimization on steroids. For the longest time a quality testing platform was too expensive for many, but now it is available to all of us. Getting started is really easy. If you are willing to invest 10 minutes, I’ll walk you through the process of setting up and performing your first test.

Google Website Optimizer
Google Website Optimizer – Image Copyright Google Inc.

The Non-Technical Technical Overview
Even if you are uncomfortable with technical stuff, you should understand setting up. That way, if you have to talk to someone technical, you will be able to communicate effectively. The process is really simple.
We’ll need a name for this experiment; we’ll call it “First Test”.
1. The first thing you must do is identify which page you want to test (Test Page). For the sake of getting your first test started, let’s say you are going to test your homepage.
2. Identify the page your visitor must reach to qualify as a successful conversion (Conversion Page). Is it your order thank you page, your registration success page, your contact us confirmation page, or a begin download page? In other words, the Conversion Page is the page your visitor gets to once they’ve taken the action you want them to take on your Test Page.
For this example we’ll use our Contact Us Thank You page.
3. The Google Website Optimizer will generate little bits of code, called “tags,” for the Test Page and the Conversion Page. All you or your technical person needs to do is a little copying and pasting. Google will even check (validate) that you’ve placed the tags correctly before you can begin. On the Test Page, put the “Control Script” code at the top part of the page (called the “head”) and the “Tracker Script” at the bottom of the page (before the </body> tag). You only need to tag the Conversion Page at the bottom with the “Tracker Script.”
Please note the tracking script specifies in its code which is the test page and which is the goal page.

Test Page: “Control Script” at top + “Tracker Script” at bottom

Conversion Page: “Tracker Script” at bottom

4. Now, look at the Test Page and decide what pieces or elements (Sections) of the page you want to test. You can test up to 8 sections on any one page.
For now, we’ll set up test of only one section. This is commonly referred to as an A/B test. Let’s choose a section we know has a high impact on any site.
a. On your Test Page there is one section almost every visitor sees: the first headline. Headlines are the easiest section to test, because you don’t need anyone else to come up with any copy or images – all you need to change is a few words (remember it is only a test). Your headline should start with an <h1> (could be h2, h3, etc.) html tag.
b. Once you have identified the section you are testing – in this case the first headline – you need to place a “section script” (a tiny piece of code) right before and after the html that renders your headline.
Google recommends excluding as much HTML as possible when defining sections. In this section script, you to change the name of the section in the code so you can identify it. For example:
Your Current headline
Change the “Description” to “Headline Test”.

Testing Headlines
For the purpose of this run-through, we didn’t spend time crafting a knock-out
headline. But you will definitely want to do this. Learning to create good headlines takes skill and practice. As you come to learn what works best for your audience, your ability to create truly persuasive headlines will improve. Here are some ideas for testing your headlines:

  • Test fractions (3/4) or percentages (75%) to prove your claim.
  • Test using numbers in your headlines.
  • Test asking questions in the headline (make sure you directly answer the
  • question after the headline).
  • Test using news in headlines to appeal to visitors.
  • Test using emotional-laden words.
  • Test different types of bolding, fonts, colors, capitalizations, sizes, and formatting.
  • Test background colors for headlines.
  • Test the number of words used in the headline.
  • Test using exclamation points.
  • Test using text to convey the benefits versus the features of your products or services.
  • Test self-focused text (we/I) in the headline.
  • Test customer-focused text (you) in the headline.
  • Test using quotations in the headline (consider the length of the headline).
  • Test the reading level of the headline.

<h1>utmx_section(“Headline Test”)
Your Current headline
Make sure the </noscript> tag follows your original content in a section.
c. Try to come up with 2 different variations of your headline other than the one you currently have (you’ll have a total of three headlines).
Once you tag your pages (see Step 3), you’ll be able to add these variations into Google’s Website Optimizer yourself and get your experiment running.
Hopefully you were able to tag these pages yourself or had someone else do it for you by now. When the Optimizer has validated these, you can begin to enter your section variations.
You’re ready to begin your first experiment!

Create and Enter Variations
Google’s Website Optimizer interface will show you a box (Variation Entry Form) labeled “Headline Test”, and it will show the html code you put the section script around (<h1>Your Current headline</h1>) grayed out.
1. On the left you will see a hyperlink that says “+ Add another variation”.
Click that.
2. A small box will appear in front of the Variation Entry Form that asks you to create a new variation. You just need to supply a short, descriptive name.
3. It will then show you the code to edit in the Variation Entry Form with the new variation name you just came up with. All you need to do is this: between the <h1> and the </h1>, where “Your Current Headline” is, type one of the 2 variations you came up with over the words ‘Your Current Headline”.
4. Repeat steps 1 through 3 to add your second variation.
5. When you are done adding variations click on “Save and continue.”

Review Test Setting and Launch
The Google Website Optimizer presents a summary identifying the Test Page address, the Conversion Page address, the name and number of variations you created and the total number of combinations you are testing. It also allows you to decide what percent of your total traffic should be exposed to the test. In our example, we’ll leave it at 100%.

Click the “Launch Now” button. The fun of testing now begins! Next step, wait for the results. You have successfully launched your first test.

How can you tell if your results are valid? Certainly it’s important to be aware of “freshness” factors that can influence your conversion rates, things such as seasonality or even time of day. It’s important to run your tests concurrently, ideally for a couple of weeks, making certain your sample size is adequate and the results between tests are statistically meaningful.
Remember, you do not want to marginally beat out what you have; you are looking for significant increases. Make sure you define what will constitute success when you start testing. You can base your criteria for success based on a certain number of conversions or sales, or you can measure results over course of time.
To be sure your test is showing a statistically meaningful impact on the variables, you have to know if you’ve demonstrated enough of a difference between the tests to declare a clear winner. As a rule of thumb, you should have at least a three times larger result (e.g., if A is 5, B should be 15).
The beauty of Google’s Website Optimizer is that it takes care of all the math and shows you which elements or combinations of elements will work best for you.
When you are ready to take the next step to multivariate testing (as in our earlier example), just add an additional section script before your element and go through the same process of creating and entering variations.

It’s that easy!

If you need assistance with your SEO or need setting up the Google Website Optimizer, why not give us a buzz on email? 


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s