.png?fit=max&auto=format&n=EQK5eX9kTD8NzWwA&q=85&s=878008bf159fcc4964d0c0d508b6e400)
How A/B testing works in funnels
Each funnel step can have multiple page variants. When A/B testing is active, HoopAI splits incoming visitors between the variants based on the traffic percentages you configure. Each variant gets its own design, content, and URL path. Stats are tracked separately per variant so you can compare performance directly.Setting up an A/B test
Add a new page variant
Click the gear icon on the step you want to test. Look for an option to Add Variant or Split Test. Enter a name for the variant and assign it a unique URL path.
Design the variant
Click Edit Page next to the new variant to open it in the builder. Make the changes you want to test — headline, layout, copy, button color, or any other element. Changing one element at a time makes it easier to attribute performance differences to a specific change.
Set traffic split
Define what percentage of traffic goes to each variant. A 50/50 split reaches statistical significance fastest. You can weight the split differently if you want to protect revenue while still testing (for example, 80% to the control, 20% to the variant).
What to test
Focus your A/B tests on elements that have the biggest impact on conversion:| Element | Examples to test |
|---|---|
| Headline | Benefit-focused vs. curiosity-driven vs. pain-point framing |
| Sub-headline | With vs. without; long vs. short |
| Call-to-action button | Text (Get Instant Access vs. Yes, I Want This), color, size |
| Hero image or video | Image vs. video; person-facing vs. product-focused |
| Form length | Full form vs. email only |
| Page length | Short vs. long copy |
| Offer framing | Price presentation, bonuses, guarantee wording |
| Social proof | With testimonials vs. without; different testimonial formats |
Reading A/B test results
Monitor test performance in the Stats tab. Each variant appears as a separate row, showing:- Unique page views per variant
- Opt-ins and opt-in rate per variant
- Revenue per variant (if the step has an order form)
- Run the test until each variant has at least 100–200 unique visitors.
- Allow at least 7 days to account for day-of-week traffic variation.
- Look for a difference of at least 10–15% in conversion rate to be practically significant.
Declaring a winner
Once you have enough data and a clear winner:- Open the step settings and disable the losing variant.
- Set the winning variant to receive 100% of traffic.
- The losing variant’s page can be kept as a reference or deleted.
Best practices
- Test one change at a time to isolate what caused the performance difference.
- Run tests on your highest-traffic steps first — low-traffic steps take too long to reach significance.
- Keep track of your test results and what you learned in a simple log. Past test data helps you make faster decisions on future funnels.
- Never run a test without a clear hypothesis: “I believe changing [element] from [A] to [B] will increase [metric] because [reason].”
A/B testing traffic splitting is random — the same visitor may see different variants on different visits. For the most accurate results, let the test run until you reach your target visitor count and do not pause or modify the test midway through.