3 Ways to Use Google Analytics to Run Web Tests
Google Analytics is a great way to keep your site relevant and up to date with insight such as traffic stats. You can use it for a variety of analytics and find out what you need to work on with your website to stay competitive. It is one of the many tools available for marketeers online and one that should not be overlooked or forgotten. It is also a free service and you can link websites to it no matter where they are hosted or how they were created, as long as you can embed tracking code onto the site – although on Wordpress this can be cumbersome without a plugin plan.
You will want to learn about your audience and see how they fit within your website’s business model to drive further business intelligence and sales. You will also want to increase traffic and find out what helps your site in this regard. Google Analytics is probably the best way to do it.
Some of the things Google Analytics does is it helps you answer the questions:
- Where the majority of your audience or customers are located geographically? → useful to see if you need to tailor your website to that geographical region.
- Which browsers did they mostly rely on? → useful for compatibility and designing your site with certain browsers in mind.
- What screen resolutions does the majority of my audience use? → very useful to see if mobile users affect traffic.
How to Start Using Google Analytics
Signing up for Google Analytics is simple. You can go to Google Analytics’ website sign up page and click on the sign up icon to the right to start the process.
First, you will provide some info on the site, or even a mobile app, you want Google to monitor through analytics then you will receive a tracking code you will need to insert into your html. That is all as afterwards you just wait until Google implements analytics onto your website to start checking for results.
An example what you will find on Google Analytics’ dashboard:
First off, you will need to provide some info about the site you want Google to monitor through analytics then you will receive a tracking code you will need to insert into your html. That is all as afterwards you just wait until Google implements analytics onto your website to start checking for results.
An example what you will find on Google Analytics’ dashboard:
Keep in mind that some Web hosts, such as Wordpress, will have their specific implementations on adding Google Analytics if you use their services. You may need to consult with them or find the instructions from your host’s website if you used a third-party host to set your website up. Here is a guide explaining how to do it using Wordpress.
Wordpress and some hosting solutions will allow you to use a plugin to start using Google Analytics right away, but you can also insert the code or script yourself (via FTP transfer), keep in mind if you switch themes you will have to reinsert it every time at least with Wordpress. The code looks something like this:
window.dataLayer = window.dataLayer || ;
gtag(‘js’, new Date());
You can also set up a User ID that will allow you to access Google Analytics from a wide range of devices ranging from mobile to desktop on demand. To do this, you will need to agree to Google’s rules and regulations and go through the set-up process for Analytics above.
Hopefully, you now realize the importance of having data about your visitors or customer base and being able to analyze this data for business intelligence. To sum it up, it will help to decrease your bounce rate or how fast Web viewers leave your page and increase page views and traffic to your website and thus your brand.
As you see, Google Analytics will show you a lot of data including bounce rate about your user or customer base:
Now that we got that out of the way, lets talk about specific ways you can use Google Analytics to your advantage. This is done by testing various pages against one another to see which one will produce the best results for you so you can use it on your website.
There are 3 Main Ways to use Google Analytics for on-site tests. These include:
- A/B Split Testing — This is testing multiple versions of a page to see, which one retains the least bounce rate and greatest page clicks.
- Multivariate Testing — This method tests several elements of a website at once and the goal is to find out which of these elements work together the best for best results.
- A/B/N Method — This method is based on the idea of using multiple versions of a webpage and not just two to compare variables or metrics.
Now that we got the basic three ways you can use Google Analytics to test website efficiency in a nutshell, lets dive into each method so you can take advantage of Google Analytics right away. But before we do that, here is what you will need to set up Experiments via your Google Analytics account:
Setting Up Your Own Experiments
Here are ways to get started using the experimental methods above with Google Analytics:
- Log into Google Analytics and you can see the ‘Experiments’ section under Behavior.
2. You will have to configure your code and make it available on the Web:
- The original landing page you’re testing has to be configured with the Analytics tracking code, and be available on the web.
- Your variation pages have to be fully developed, configured with the Analytics tracking code, and be available on the web.
3. Click on ‘create experiment’ to get started.
4. There will be a wide range of options for you to pick from.
5. After you click next, you will see various variants to your experiment you can configure:
- The percent of traffic you will include in your tests is important as the greater the number the larger pool of visitors you will be testing and should get greater variations in results.
- There will also be three different metrics for you to test:
- Bounces: Refers to the bounce rate or how quickly users left your webpage vs those who stayed on and clicked on multiple elements or pages within the site.
- Pageviews: Refers to the amount of pageviews you received during the given time frame the test was running by all viewers.
- Session Duration: Refers to the length of time users stay on your site. This is useful, for instance, if you run a news or entertainment website, where you want to keep readers hooked and engaged.
You can also add your own goals or variables by click on the ‘create a new objective’ icon:
You will then be taken to this screen to find your own metric you want to measure:
You can also set up your own goals and use them for metrics in the Experiments in this section. You will have a choice from various metrics and here is how Google describes the categories of choices you will have:
- System Objectives: These deal with revenue, transactions, Adsense impressions and things related to monetization as a whole. Some other options in this category include Adsense impressions, pageviews, session duration (if you did not click on it from the drop down menu you can customize it here further), and more.
- Google Analytics Goals: These are goals you already set up that are imported from Optimize once you have Read & Analyze permission to the linked Analytics view. You need to enable permissions for this to function.
- Custom Objectives: This is where you can really customize what you are looking for exactly and create a goal or metric for you to measure. You can start out by choosing from event category, event label, event action, event value or URL.
However, the way you will see the sections split up in this section will look like this (If you have an Adwords account you can play with goals related to it under the Smart Goal section):
You will also have the original page and a variant page to compare it to. If you leave it as such, you will be using the A/B Split testing method below. However, you can add variant pages beyond just two (A/B/N method of testing) or other variables (Multivariante method of testing) to try the other two methods. The run-time for an experiment is at least two-weeks as required by Analytics.
Google Analytics uses a multi-armed bandit approach to manage traffic to the various pages you have running and declare a winning variation. You also have the added option to distribute traffic evenly among the variations.
Testing Multiple Pages in Analytics for Best Results
As websites get redesigned, changes come and go that may affect traffic or user engagement, sometimes not for the best. This is where running the three tests comes in.
We will outline the step-by-step process for the A/B split test method below (relies heavily on the Experiments section above), and there will be some minor variations between running the other two tests once you know how to set the A/B split test method up. The variations will be discussed below in the sections of the other tests.
All three tests will allow you to test different versions of a particular webpage, or multiple, and let you know, which of these versions give the best results with analytics. Here is a visual depiction of the process showing the basic A/B split test method.
To run A/B Split testing, first of all create different versions of a particular page you want to run the test on. You can create a copy page from your website in a variety of ways, for instance going into the Pages section of the Wordpress site editor will show you all the pages you have on your site. Then, click the ‘…’ icon and select copy.
The page will automatically be copied and you will be directly taken to the html/visual editor that way and can edit the page to your liking. Other website hosts and designers will have similar methods to copy sites, otherwise you can simply copy the html code and modify it later while placing it to your site via a FTP. Obviously, you will not want to run the new versions of the page live just yet.
More information about copying a Wordpress page can be found here.
A/B Split Test Method
Here are the basic steps to run A/B Split tests on your website
- Create two versions of a webpage by either changing one small thing like a header or font, or making the variation as complex as you like.
- Leave the initial webpage the way it was and find one element to change, such as the headline.
- Then run the second website under variant 1 as described in the ‘Setting Up Your Own Experiments’ section above.
- You can pick one of the predefined variables in the Experiments section or choose your own, as previously mentioned. Here are additional variables you can test or consider when testing different webpages against one another using the A/B split method (again under experiment in Analytics you will need to click on the ‘create new objective’ option).
You can find out what is important for you first and what drives your business and concentrate on these variables first. Then, if you see a huge difference between the two sites, pick the more successful version and use it as your page. However, if the numbers are close, look for other variables to test as well that may be significant in your revenue stream.
5. Finally, click on the ‘manually insert the code’ tab and input the code shown onto the opening head tag at the top of your original page:
You will then need to wait for a period of at least two weeks to see the analytics play out, but can check in periodically via the Analytics dashboard.
Multivariate Method of Testing
The A/B/N method is similar to the A/B method above and all the steps will be the same to start seeing its results. The big difference is you will be adding a wide range of variables or metrics to the two (or more) sites you are comparing besides just one variable at a time.
This way you can see what works with what on a site and this may show you better results than just comparing one variable. It is also effective when creating multiple variable sites under Experiments and each having various metrics to compare one another with.
Again, here is the section on the three basic metrics that you can add your own metric onto for multivariate method of testing (you do this by creating a new objective):
There are some interesting variables or metrics you can compare and check when adding your own metric as shown here:
You can check successful sign ups to your site, for instance, which are very useful in eCommerce.
A/B/N Method of testing
A/B/N method of testing is an extension of the two testing methods outlined previously. However, unlike both A/B method and A/B/N method, that test two versions of a webpage against one another, using either one variable (in the case of the A/B method) or multiple (in the case of the multivariate method), multiple webpages will be used simultaneously.
To get started in this process, simply follow the outlined process of the ‘A/B method and the ‘setting up your own experiments’ sections. Then, when you get to the section asking for the original webpage and a variable, click on ‘add variant.’
You will need to provide the URL for this webpage to add or as many as you wish to add from your website.
Content Experiments Framework is a very interesting and in-depth tool that anyone tapping into Google Analytics should try out. The three methods outlined here should give you great insight into your traffic and future success with your website or business as a whole.
Here is how Google Developers website explains Content Experiments Framework:
With all these tools at your disposal with Google Analytics, you have no excuse to give it a try. It is free after all and available for anyone who can insert some code into their site. It can really give you some insight into your audience and find out that edge you need to keep your online presence viable.
What other tactics do you use to optimize SEO, user engagement and drive page views?