How to do Split Testing Inside of Google Analytics

When any visitor is browsing your site, the factors that impact their willingness to buy a product/service include your website design, the layout and its content. As an online marketer, you have to give a thought to how confident you are about these aspects on your site.

Creating a Positive Experience

Even as you make certain decisions and update your website, how can you ensure that these changes are going to have a positive impact on your visitor’s experience? A/B Split Testing is the one way you can increase confidence in these decisions that you make about your site. Using these tools you can scientifically test & validate the hypotheses around your marketing goals.

The Comprehensive Guide

The “Experiments” tool in Google Analytics makes it easy to set-up these A/B Split Tests on a website with the use of Google Analytics. This is a guide to how you can do Split-Testing within Google Analytics

Define your Goal

Some visitors’ actions are tied directly to your company’s website goals. It is important to set them as Goals within Google Analytics. Business websites should be tracking leads & sales of their services/products. Non-profit sites should track fund-raising & supporter engagement. Political sites should track volunteer registrations & donations

Forming the A/B Test Hypotheses

Once you have established your goal of improving donations, increasing leads or engagement, you will have to form a hypotheses. You will have to think about the messaging, design, layout or content decisions that will help you achieve this particular goal in a more effective manner.

When you use Google Analytics, there is no restriction on the number of hypotheses you set-up. The algorithm that Google uses is called the “multi-armed bandit experiment”. Though the statistics behind this are quite complex, the idea is fairly straightforward. Over time, the hypotheses will begin showing promise and Google will automatically send a larger number of visitors to the leading sites.

The advantage this has over conventional A/B testing is that the traffic is distributed evenly till the point of completion. Every hypothesis will have to be set-up as a single page and must have a unique URL on your site. The manner in which this is done will be dependent on the CMS your site uses.

Setup the A/B Split-Test Experiment

Once the goal & hypotheses are in place, set-up the experiment within Google Analytics. Access “Experiments” via the reporting bar in Google Analytics> Choose “Behavior”> Select “Create Experiment”.

Experiments only need a Name & the Goal you are measuring success with. There are some other options such as the amount of traffic that should be experimented with and e-mail notifications etc. When you are starting out, simply use the default experiments but do enable e-mail notifications

Drop in Original URL & Hypothesis Variation URLs

Now go ahead and add all the hypotheses to the “Experiment”. All you have to do is copy & paste the URL’s that you set-up in. For tests on existing pages, use it as an Original. For new content, choose the URL that you will finally use as the winner

Insert Experiment Testing Code

Once the hypotheses URLs have been setup> Choose “Manually insert the code” & install the experiment code-snippet inside thetag of the Original Page> Save the template> Click “Next Step”. Google Analytics will then verify that the experiment code has been placed properly

Now, all you have to do is “Start Experiment” & Google Analytics will then take charge of routing visitors to the hypotheses & recording their effectiveness.

Who is the Winner?

The winning hypothesis gets decided algorithmically over time- generally over 2-4 weeks. This period will depend on the volume of traffic your website sees & the relative-performance of the hypotheses as well as the rate at which the leading-hypotheses converts etc. For certain experiments, you might be able to decide to move forth with a hypothesis even before Google’s algorithm statistically is able to.

Since Split Testing inside of Google Analytics is a free tool that is easy to set up and use, it helps you make much more informed decisions about all the factors that matter in a website and will help you achieve your organization’s goals in a much better way.

Thanks for reading,
Bizow Online