Split Testing in Google Analytics Tutorial

Posted by & filed under , .

Let me answer the first question that pops into your mind straight away!

What is Split testing?

Split testing is a way to test 2 different version of a website/squeeze page/Sales page to see which one of them performs better in terms of CTR(Click Through Rate).

You can do minor or major changes to their web pages when split testing different version, you do this by either changing images, colors, placement and the body copy(your marketing text-mumbo jumbo) and see which one increases the rate of the action you want performed( More people clicked the green colored button version than on the blue colored button version)

Here is an example of a what 2 versions of the same website could look like:


(Image is of some split testing that was done to HighriseHQ)


How can I do split testing for free?

Lets be totally honest, there are a bunch of marketers out there trying to sell “premium split testing products” that do pretty much everything free solutions already do for well… free :)

What we are going to do is to make 2 versions of a single product eCommerce website(Simple HTML with a PayPal Buy Now button), for this particular test i will use a website that have been up and running for a few weeks already.

The website we are going to be split testing sells an Electronic cigarette kit called a “Lavatube”, don’t panic if you have no clue what any of that means, it is a new niche that I happen to have a bit of knowledge on and is trying to capitalize on.

The Website looks like this right now

(Click to enlarge)

split-test-original     split-test-variation1

Left: Original – Right: Variation

Since I am just doing this for demonstration, All i will do in the 2nd version of the website is to rearrange the “Kit Includes” and “Shipping” Section and move them over the image of the whole kit.
This version of the webpage i save as index2.html and upload to my web server.

After the variation have been uploaded to the web server, I go into Google Analytics and enter the account for the website i want to split test.

I then follow these step.


  1. Click Behavior.
  2. Click Experiments.
  3. Enter the URL of the page you want to split test(in my case it is index.html).
  4. Click the blue “Start Experimenting” button.
  5. Name your experiment, I put “Split Test”.
  6. select your “Objective for this experiment”, I am going to create a new objective.
    1. On the Goal Setup page click on “Custom”, then “Next Step”
    2. Give your goal a description, I want to track the sales count for every page variation, therefore i give it the name “Purchase”.
    3. The way my website works is when the user has ordered the product through the simple Paypal buy now button it will redirect the user to thankyou.html, so this is the page we want to check if anyone have visited. so we choose Type: “Destination”, then click “Next Step”
    4. Enter the path to your thank you page, my page is on http://mywebsite.com/thankyou.html so I should only enter thankyou.html Into the text box.
    5. I ignore all the other options and click “Create Goal”
    6. Now I press the “back” button in my browser 2 times(because it’s faster).
    7. I select the objective I just created in the dropdown menu.
  7. The percentage i want to split test is 50%
  8. I ignore the advanced settings and click “Next Step”
  9. In the “Variation 1” area i enter the URL to my variation, which in my chase is http://mywebsite.com/index2.html, Click Next Step
  10. I want to manually insert the code, so I click on the “Manually insert the code” button.
  11. Copy paste the code into the <HEAD> of your web pages that you want to test(the original and any variation that you’ve made)
  12. Now click on “Start Experiment”.
  13. Your experiment will now start and half of your visitors will see the original, and half will see the variation, by default Google Analytics will run this experiment for 2 weeks.


Thanks for reading, if you found it useful be sure to share it on facebook/twitter/linkedin/etc.

Leave a Reply

Your email address will not be published. Required fields are marked *