Google Analytics Content tool

Google Analytics Experimental Tool

Content experiment is a tool designed to create A/B test from inside Google Analytics. This tool has various advantages over old Google Website Optimizer, specifically when you are just starting the website testing journey. It provides a quick to test your main pages and requires few code implementations.

Some of the features of this tool

  1. Only the original pages is necessary to run tests, the standard Google Analytics tracking code is used to measure goals and variations.
  2. Website goal that are defined on Google Analytics can be used as the experiment goal, including Ad sense revenue.
  3. Google Analytics segment builder can be utilized to segment results based on any segmentations criterion.
  4. Multi armed Bandit approach yields faster results than classical testing at less cost and with just a statistical validity
  5. Tests will automatically expire after 3months to prevent leaving test running if they are unlikely to have a statistically significant winner.

How to use content experiment to create A/B tests.
For creating new experiments go to the Behavior section and click on Experiment link on the sidebar. You will observe a page which shows all your existing experiments. Above this you will have a button Create experiment. Then you will be asked for the following information.

  1. Name of this experiment
  2. Objective of the experiment– here you can choose an existing goal or create a new goal.
  3. Percentage of traffic for experiment– the higher the percentage the quicker you will get significant results.
  4. Email notification for important changes
  5. Distribute traffic evenly across all variations: if you turn it on, then you will get the benefits of Multi-armed Bandit approach which is mentioned above.
  6. Set a minimum time the experiment will run– it defines the minimum period upto which Google Analytics will not declare a winner. IF your website has behavior patterns on weekends and weekdays this must be highly recommended otherwise you could end up with a page optimized for only one of these segments.
  7. Set a confidence threshold– the higher the threshold value the more you can be in the result but on the same side it also means the experiment will take more period to finish After defining the information above click on it and you will get the following page.In this page you can add the URLs of your original page and the variations you would like to test. You will observe thumbnails of the page which help to make sure the URLs are correct.1
    Setting up the Content Experiment Code
    In this you have to choose either implement by yourself the necessary code to run your test or you will be given the option to send an email to whomever is implementing the code.Click Next

    Validating and confirming the content Experiment code
    You will need to implement one code in order to apply this tool. In this step your pages will be verified. If the code is not found you will observe an error message.
    It is to be noted that you will be able to skip validation if you want; you just click on Start Experiment. If you do so you will find a popup with the message “Experiment validation had errors or did not complete. Are you sure you want to start the experiment? If you are sure that your experiment is proper, you may continue.” But It is recommended that you check the code to understand why you are getting an error and try validating again.


    Content Experiment results
    Once your experiment is live you will see the following options

    • Conversion Rate- gives you the option to check the test result using alternative metrics.
    • Stop Experiment
    • Re validate
    • Disable Variation
    • Segmentation this is a valuable feature it allows you to understand better how each variation performs for each segment of visitors on your website.

    And below we see the result page of a test with winning version, the blue sign up with a lift of 52% in conversions as compared to the original Reviewing all experiments
    If you want to review your experiments visit

Leave a Reply

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