Do you want to run AB tests? But are not sure how to get started or what it costs to run them? Then this article is for you, as you’ll learn what exactly AB testing is, why it’s important, and how you can easily run AB experiments for your websites without massive marketing budgets.
What is AB testing exactly?
AB testing is often called split testing because you are splitting half of your visitors to one version of your site, and the other half to another version of your site. By running this experiment, you can test if a specific change is improving the conversion rate of a specific goal on your website.
For example, lets say you want to test if adding a refund policy within the buy button on your ecommerce store would increase the number of conversions of adding items to a shopping cart. With an AB test, you could create an alternative version where the buy button says “Buy now (30 day money back guarantee)”. Instead of trying out this change and applying to all your traffic, you would only have 50% of your traffic seeing this.
When should I AB test?
You should run AB test experiments on your website when you have steady traffic, or are about to run an advertising campaign which will direct a consistent amount of traffic for a period of time. As well, your website should have a clearly established sales funnel that takes the visitor from point A to point B, where point B is the sales conversion.
What does a sales funnel on a website look like?
For example, a sales funnel for a website which sells charter fishing trips could look like the following.
Charter Fishing Sales Funnel
Your funnel could be broken out like the one above into 5 steps but the actual clicks and steps on your website may be less. For example, you may build a single page which takes a visitor through a funnel as they scroll down the page. This type of long form landing page is a very common design trend as it makes it easier for the visitor. The less things they have to click and wait for, the less likely they will leave or get confused. The rule is don’t make them think and focus on clarity!
Keep in mind that a conversion does not have to be an actual sale, it can be signing up for a newsletter, inquiring about a product, or just talking to chat support. It is best practice to sprinkle points of conversion throughout your funnel to support the various stages a visitor may be in.
How do I get started with Google Optimize?
Before you get going with Google Optimize, you need to have Google Analytics installed with specific goals defined. If you are not familiar with Google Goals need help, I recommend you read their official page about setting up goals. Alternatively, if you need help you can contact us and we can get you setup.
Setup Google Optimize Account
After Google Analytics is setup, visit Google Optimize at https://optimize.google.com/
Next, create an account and start the process.
Assuming Google Optimize has not changed too much, you should see something like the above.
Create your Experiment or Simple AB test
Click “Let’s go” and then select the type of experiment you want to run.
In this case, pick “A/B test”, fill out name for your experiment and the URL of the page where it will run. Ignore the other options for now, we’ll go into those in future tutorials.
Add a variant to the experiment
The variant is the “B” path in the AB test, or the new change you want to test. Give it a name, and then save the changes.
You will now see something like below:
Look at the picture above and observe how changes is marked as 0 and there is an Edit button to click. Click the Edit button to get started with creating your new variation. If this is your first time doing this, you will be prompted to install a Google Chrome plugin first.
Visually edit your variant
Once it loads up you should see the page with an editor options like below.
Great, you are one step away from making your first variation!
Next, click an element on your webpage that you would like to change. In the scenario above, we will change the copy on the book button to say “Book Your Free 30min Consultation” instead of “Book a Free 30min Consultation. Also, we will change the colour of the button to be red.
Finish your changes by clicking Save, then click Finished to close the editor
Notice now that the variation indicates 2 changes, you can hover over this to get an idea of what you have changed.
Add Targeting Rules to the experiment
Next take a look at the breadcrumb trail at the top.
The next step is to select “Add Targeting Rules”.
Keep this to the default for now. In the future, if you wanted to target specific visitors using specific rules you could modify this.
In the above, you could run your AB experiments based on ads or various types of visitors.
Link it Optimize to Analytics
Next, click the breadcrumb option to Link to Analytics.
Select the Google Analytics account that you have for your website, click the “Add experiment objective” button and then select “Choose from List”.
This list of objectives will show you the Google Analytic Goals that you have setup already. The objective is the conversion goal you will be measuring. In the case for my website, I will be selecting the “Contact us” goal as this will count as a conversion in my experiment.
Install Google Optimize Code
Lastly, setup the Google Optimize code by visiting the section below.
This step is similar to installing Google Analytics. You will need to copy and paste a small code snippet onto your website. This piece of code will “inject” your variation on to the website so no custom web design or development is needed! You can control your entire experiment with the visual editor that you saw at the beginning of this tutorial.
That’s it! You can now run the experiment by clicking the Start button at the top right of the breadcrumb.
How can I monitor the results?
Once your experiment has started, you will see a table and graph of the results to date like below.
Look at the image above and notice how the variation has a 95% Probability to be the Best solution. You may have to wait a while for Google to tell you what the result is, this depends on the amount of traffic and conversions your website has. For most small businesses, you should expect at a minimum 2 weeks before you can make a decision.
Wrapping it up
In this tutorial, you have learned what AB testing is, why you should do it, when your website is ready for it, and how to do it for free using the Google Optimize tool.
This is one of the most powerful and important tools for running a high performing website. Keep in mind there is a lot more to AB testing than what we have covered here. To run AB tests successfully you should adopt a program, like a fitness program, that you apply to your website periodically.
In addition, understanding what a visitor is doing and creating a hypothesis on changes to test requires experience. Visit our Conversion Rate Optimization page and checklist to get an idea of what’s involved in a full program. We offer affordable solutions for businesses of all sizes and can find the right tools for your budget and needs.