A/B Testing Flows
One of Appcues' more advanced features is A/B testing. As you learn more about how your customers interact with your flows, you can use A/B testing to learn even more!
The best way to utilize this feature is to compare how one flow performs against another. Are you users more likely to engage with a slideout containing a gif or a static photo? Do you users like a tooltip tour, or would they prefer one modal containing all content? You can use A/B testing to learn exactly this!
How it works
Appcues will automatically assign your users into 2 different groups:
1 (Group A) or a
2 (Group B)
The property that Appcues will send over is called
_ABGroup, and the value will either be depending on which group the user was assigned. We generate this property, and whichever A/B group a user is assigned persists on his profile and will remain the same. If this sounds like jargon to you, don't fret. We geek out about analytics and are happy to help—reach out!
A/B Testing Your First Flow
You can enable A/B testing from the flow's Settings page. Scroll down until you find this section below:
Check the "Enable A/B testing" box to show the flow to a random 50% of your user base. This random bucketing is done for you, and you can also test two flows against each other by selecting Group A for your first flow and Group B for your second.
Analyzing the Results
To properly analyze the results of your A/B test, we suggest integrating Appcues with a third-party analytics system. Those systems will let you see your results in a larger context and provide reporting features designed for this kind of analysis. You can also view the results within Appcues, by comparing the Analytics page for each flow.
If you opt to view the results in your analytics tool of choice, you'll want to view all users who:
- Completed a "Flow Started" event for the flow in question.
- Were either in Group A or Group B, depending on which you want to analyze.
Most systems will let you view that segment in funnels and other types of useful business reports. Give it a try and let us know what you think!
To accurately measure the efficacy of your experiment, you'll first need to enable one of our third-party integrations and have the ability to segment users by their properties.