Skip to content
Elooo SystemHelp Center Go to website

Funnels and websites

How to run an A/B test between two versions of a funnel page

Go to Sites › Funnels, open the step and click Create Variation. Edit the variation in the builder, use the slider to split traffic between the original page and the variation, and click Start Split Test. Each version's results are on the Stats tab.

Updated on

Before you start

  • The funnel must have a domain connected. Without a domain, the create variation option doesn't appear.
  • Split testing only exists in funnels. To test a website page, import that page into a funnel and run the test there.
  • Prepare a difference people will notice (another headline, image or button). If both versions look almost the same, the result won't tell you anything.
Where to find it

Sites › Funnels › (open the funnel) › (step) › Create Variation

Step by step

Create the variation and start the test

  1. In the left menu, open Sites, click Funnels and open the funnel.
  2. Select the step you want to test and click Create Variation.
  3. Choose how to create the variation from the options shown, then edit it in the builder with the differences you want to test. The original page is now called the Control.
  4. Move the percentage slider to decide how much traffic goes to the Control and how much to the Variation (for example, 50/50).
  5. Click Start Split Test, then Apply to confirm.
  6. Optionally, click the gear icon under each version to change that version's dedicated URL path.

See the results

  1. Wait for traffic to come in. The split starts showing numbers for each version.
  2. Click the Stats tab above the page.
  3. Click the page under test to expand the Control and Variation numbers (visits, conversions, rate).
  4. Once you have enough data, keep the version that converts better.

Tips

  • Don't touch the Control while the test is running; if you change both versions midway, you'll no longer know what made the difference.
  • Visitors arrive through the step's normal address and are sent to one of the versions according to the percentage. The choice is stored in a cookie, so they always come back to the same version.
  • To see the other version on your own computer, open the address in an incognito window or a different browser.
  • Ideas to test: headline, hero image, button text and color, number of form fields, single-page versus multi-step checkout.

Common problems

  • All traffic goes to the Control → The address in your ads isn't the same domain as the step under test. Use the funnel's domain in your ads or set the step under test as that domain's default page.
  • The numbers don't match the percentage I set → One of the dedicated addresses (Control or Variation) is being shared directly, outside the split. Only share the step's main address.
  • The Create Variation button doesn't appear → The funnel has no domain yet. Connect a domain to the funnel and try again.
  • I want to test a website page → Create a new step in a funnel, import the website page into that step and run the test in the funnel.

Frequently asked questions

Can I have more than one variation?

You can compare the Control against one or more variations, splitting traffic between them with the percentage slider.

Was this helpful?