The following blog has been written to better educate users about the current Google testing offerings, as well as a technical walk-through of their Google Content Experiments tool.
Split tests are the first steps toward optimization. They represent a simple use case wherein you are asking your users “which is better?” and they reply through their conversions. Google Analytics (GA) offers this functionality (pending some setup) it is called Google Content Experiments (GCE) and its interface lives inside of your GA account under the “Behavior” reporting section in the sidebar.
Google Content Experiments have been available inside of GA since 2012, it replaced the old and now deprecated Google Website Optimizer. It is a free offering that allows you to freely split test content without having to pay any usage fees that are part of other testing company’s business models.
The primary (and simplest) use case for GCE is to divide traffic evenly between two or more pages using a redirect. Below we will discuss and illustrate the roadmap of testing as well as the user flow for GCE experiment creation.
Testing Success Roadmap
Before you run to test the one element that you have questions about, take a step back and consider how you want to run tests on your site. Also, consider what you will do if the grand slam variable you created does not outperform the baseline.
Step 0: Log into GA and travel to the Behavior > Experiments section and click create experiment.
Step 1: Name the Experiment, select a goal and choose what percentage of traffic should enter the experiment and whether you would like to enable email notifications. For advanced users you can:
Step 2: Build the pages that you want to redirect, label one as the original (baseline) and the other using a descriptive name of the change you are making.
Step 3: Add the GCE script to your original and variant pages ONLY.
Step 4: Review and start the experiment. Keep in mind if you have a non-standard implementation of GA, or you haven’t placed the GCE tracking code at the top of the head, then you will see errors on the “Experiment Code Validation” step. All that is left to do now is start the experiment!
Now when you travel to your experiments section you will see a test running in your “All Experiments” list. Click into that list to collect reporting. Keep in mind that you can use “variant” as a secondary dimension in some of your standard reports as well. This is useful for diving further into devices, landing pages, and referrer type in regard to your experiment.
When tracking your experiments’ success be sure to look into the status that Google updates while it runs, below are the possible statuses.
Setup (Step number)
Setup (Ready to run)
Ended (Stopped manually)
Ended (Time limit reached)
Ended (Winner found)
Ended (No winner)
Ended (Processing)
It is possible to create an experiment without having to redirect the page; however it is much more complicated. It requires that you perform all of the changes on your page via JavaScript. The technical details are extensively documented on Google’s developer site, but they are not for the faint of heart.
No-redirect experiment documentation:
https://developers.google.com/analytics/solutions/experiments-client-side
If all of this seems like too much hassle, or you want a more powerful platform with targeting and personalization features, then you may want to consider Google Optimize 360. This is Google’s enterprise optimization tool and it is likely to be released sometime in the next year. If you want to learn more about it then please see our Google 360 Suite blog.