Split Testing in Funnels

What is split testing?

It’s a marketing tactic used to refine funnels by determining what messaging, design, layout, etc. gets the best conversion rate and optimizes your lead and sales generation.

By utilizing Split Tests, you can have one link to a funnel page that has multiple versions. You can determine the ‘split’ or percentage chance that a lead visits page version A or B.

Then, after a while, you can determine which page version had the better outcome or conversion rate and use that information to enhance your funnel.

Benefits:

Improved Conversion Rates: Identify the most effective variations to boost conversions, whether for sign-ups, purchases, or other actions.

Enhanced User Experience: Tailor your website and funnels to better meet your audience's preferences, creating a more engaging experience.

Informed Decision-Making: Use data-driven insights to make strategic decisions, avoiding assumptions and improving marketing effectiveness.

Please Note

Split testing will only be possible once you have added a domain to your funnel.

Step 1: Creating a split test

Navigate to Sites > Funnels and create a new funnel or access an existing funnel.

Choose the step you want to create a split test for.

Click “Create Variation” and two options will appear. You can create a duplicate version (which will clone the default/control page), or you can create from blank.
- Most commonly, you’ll create a duplicate version as you will be making edits to the control page (perhaps adjusting the layout, color, or other stylistic features.)

Step 2: Running a split test

Once you’ve created your variation page, you’ll be able to view a preview of both the control and variation pages. (In this example we made some stylistic changes to the color, fonts, etc.)
Choose the percentage for your split test, to determine how much traffic you want to be directed to the control vs. variation page. Slide the button to adjust the percentage and click “Apply Changes” to confirm.

Use the updated link that is generated wherever you use your funnel link, to perform the split test.
This link will ‘stick’ for a person once their computer is cookied, so if you test the link yourself and get the variation page, you’ll always get the variation page in that browser. (You can test the link in a new browser, on another device, or within incognito mode to see if you get the other version.)
Once you’ve had traffic to the site, you will be able to view statistics for the pages to see which one has better results. These statistics can also be found within the Stats tab in the selected funnel.

You can hover over the page and “Declare this as Winner” to continue with the better version of the page.
Doing this will promote the ‘winning’ page as the control page moving forward. It will also delete the other page. (A popup window will confirm this before you make the change.)
Click OK to accept and continue.

NOTE: You may restore an old split page that you intentionally or accidentally deleted. Simply click “restore” from archived pages at the bottom to immediately restore an old version.

Use Cases:

Online Course Sales Funnel: Test different headlines and images on the landing page to boost enrollments. For example, compare "Learn Spanish in Just 30 Days" with "Unlock the Secrets to Mastering Spanish Fast" to see which attracts more students.

Webinar Registration Funnel: Optimize registration rates by testing different headlines and form designs. For instance, compare "Discover the Top SEO Strategies for 2023" with "Boost Your Website Traffic with Proven SEO Techniques" and different form layouts.

E-commerce Checkout Funnel: Improve conversions by testing page layouts, button colors, or trust badges. For example, compare a single-page checkout with a multi-step process, or test red vs. blue buttons to see which reduces cart abandonment.

Lead Magnet Funnel: Enhance lead generation by testing different designs and opt-in forms. For example, compare a visually appealing lead magnet cover with a compelling headline, and test form designs with varying numbers of fields.

FAQ

Q1: Why is all my traffic going to the control page instead of splitting between the control and variation pages?

Ensure that the domain used in your ads matches the domain of the funnel step where the split test is being conducted.

If traffic is directed to your root domain or a different subdomain, the split test may not work correctly. To resolve this, set the default page of your root domain as the page on which the split test is running.

Q2: Why don’t the stats match the split ratio I configured for the test?

This issue can occur if the direct links to either the control or variation pages are being used elsewhere, such as in another funnel or ad that bypasses the split-testing logic. Ensure that all traffic goes through the correct split-testing link to maintain accurate data.

Q3: How do I manage split testing with Google Ads or Facebook Ads if their policies don’t allow path URL redirection?

Google Ads and Facebook Ads often restrict path URL redirection, which is common in split testing. To work around this, conduct your split test on the second step of your funnel. Direct traffic from your ads to an initial step containing engaging content, then guide users to the split test in the subsequent step. This approach adheres to ad policies while still gathering valuable split test data.

Q4: Can I recover a variation page that I deleted during split testing?

Yes, you can restore a deleted variation page. Simply navigate to the archived pages section at the bottom of your funnel steps, find the page you want to restore, and click "Restore." The page will immediately reappear in your funnel.

Q5: How can I be sure which page version is winning?

You can monitor the performance of both the control and variation pages through the stats tab of the funnel. Once you have enough data, you can declare the better-performing page as the winner by hovering over it and selecting “Declare this as Winner.” This will promote the winning page as the new control and remove the other version, although you will be asked to confirm this action before it takes place.