Through Unstack, you're able to run full A/B tests for the components that make up your webpages; in fact the process is very simple:
1. Setting up Your Test
The first thing you need to do before you can set up the test, is to identify which component you want to to test. For this example, we're going to set up a test to identify whether the quote on my webpage performs better than another customer's quote.
Therefore to get this test started, I need to click on the beaker icon as shown below to toggle on A/B testing for this component.
2. Configuring your Control & Variant
Once you've toggled on testing for this component, you'll see a couple new options under that menu option.
In particular, there are now options for both Control and Variant. In fact, it's here where you make the changes to the component that you're looking to test.
In the below example, I changed up the quote entirely. For your test, these changes can be anything from a completely different quote, to a change in the person speaking, to a different picture for the quoted person; it all depends on the component.
Once satisfied with the differences between your Control and your Variant, you're able to move onto the final steps; adjusting the final options for your test, and starting it!
3. Final Settings & Starting Your Test
Now that you've settled on the design for the Control and Variant, it's time to choose a goal, decide on a percentage balance, define a Result Certainty, and finally; give the test a name and start the test.
Choosing a Goal
When setting up an A/B test for a component, you must also define the desired outcome for the test; commonly referred to as the Conversion Goal. This is a crucial step, because without defining a goal, there's no way to measure the success of a test.
With Unstack, there are three different types of goals you can choose from:
- Views of a specific page - this needs to be a page hosted on your Unstack site.
- Submitted a form - either a HubSpot, Klaviyo, or Unstack form.
- Clicked button in section - whether the user clicked on the action in this component.
This setting allows you to define a balance between the Control and Variant. If you want to weigh the Control more heavily, you can just set the percentage for Control higher than that of the Variant. The same is true for those who want to weigh the Variant more heavily than the Control.
In other words, the higher the percentage, the higher chance that version has of showing up for a user.
By default, this setting is set to 50% for both options. For most people, this is going to be the recommended setting.
Don't just take it from us, Thanos feels the same way:
Next, you'll need to define how Unstack chooses the test's winner. This is done by choosing a percentage where you feel confident that if either the Control or the Variant were to reach that level of confidence, then they'd be the winner of the test.
This confidence level is increased as more users interact with the component being tested, and complete the goal defined above. Additionally, the length of the test plays a role in helping Unstack define that confidence level.
4. Conclusion + Next Steps
At this point, all that's left to do is to give this test a name and hit "Publish changes and start test" to start the test.
When giving it a name remember that you're going to be looking to this test for insights at a later date. Therefore, naming it something that's understandable to you at a glance will be paying it forward to your future self. For the above example I'd name the test something along the lines of "Homepage Quote - Tyler Redner/Brian Arnold" so that down the line I'll remember exactly what's being tested.
All in all, we hope you found this guide helpful. If you're looking for the answer to a question that we didn't have in this guide, give us a shout as we're always happy to assist!