Waterfalls in CUX allow you to see how users move through your website step by step. They are especially valuable in A/B testing, as they show not just whether users convert better in one variant, but also how their journey differs between variants when additional steps are included. By setting up separate Waterfalls for each variant, you can identify:
- Where users drop off.
- Conversion rate of each variant.
- Whether one version creates a smoother progression than the other.
How to compare A/B test variants using Waterfalls:
- Create a Waterfall for your first variant
- Name your Waterfall accordingly. Set the first step as your Variant A by choosing the corresponding type of event:
- Page Load for URL or its parameter,
- Click for element ID,
- Data Layer for data layer values.
- Add the test goal
- This step represents the key conversion of your experiment (e.g., purchase, registration, form submission, adding a product to the cart).
- Place this step as the last step in the Waterfall.
- Add additional intermediate steps (optional)
- Include steps between your variant event and the goal to better understand differences in user journeys.
- Be cautious: optional actions not directly related to the test (e.g., clicking optional consent checkboxes or adding a discount code) may reduce sample size and narrow results, so we do recommend creating an additional set of Waterfalls if you want to check the impact of your test on those actions.
- Save the Waterfall
- At this point CUX will begin filtering data for this variant.
- Create a Waterfall for the second variant
- Click Manage → Copy on your first Waterfall.
- Change the first step to Variant B while keeping all other steps identical. This ensures the Waterfalls are comparable.
- Compare results
- Wait for both Waterfalls to filter data.
- Ensure the same time range is selected for both variants.
- If some steps show zero visits, verify that events are correctly configured and that enough data has been collected.
Example: Imagine you are testing two registration page designs:
- Variant A is the current page with multiple form fields.
- Variant B is a simplified version with fewer fields and highlighted benefits of signing up.
You create two Waterfalls for each variant: a basic Waterfall containing only the variant step and the goal step (successful registration), and a detailed Waterfall that adds optional intermediate steps, such as tracking form input changes or clicks on optional consent checkboxes.
- Variant B converts more users overall than Variant A.
- Users of Variant B start filling out the form more often, showing that the simplified design encourages additional interactions.
- Optional elements, like consent checkboxes, are clicked more often in Variant B, demonstrating that highlighted benefits are being noticed by users.
- Drop-offs are higher in Variant A at each step, confirming that the simpler form and highlighted benefits of Version B worked as intended.
This approach demonstrates how comparing Waterfalls of different lengths—basic and with optional steps—provides both a high-level conversion view and deeper insight into user behavior differences between variants.