How to A/B Test Order Bumps in CartFlows

CartFlows Pro lets you A/B test (also called split test) an order bump, so you can compare two or more versions of the same offer on your checkout page and keep the one that converts best. You can test a different headline, product, discount or design without creating a new checkout step.

This guide shows you how to set up an order bump split test, read the results and declare a winner.

How Order Bump A/B Testing Works

An order bump split test has one Control, which is your original order bump, and one or more variations. CartFlows shows each checkout visitor one version, based on the traffic share you set for each version.

• Each shopper sees only one version of the order bump, in its usual place on the checkout page.
• A returning shopper keeps seeing the same version for 30 days, so their experience stays consistent.
• CartFlows tracks impressions, adds to cart and orders for each version, so you can compare them side by side.

When you’re happy with the results, you declare a winner. The winning version becomes your regular order bump, and the others are archived.

How to Set Up an Order Bump A/B Test

Step 1. From the WordPress admin, go to CartFlows → Funnels and open the funnel that has your checkout step.

Step 2. Click the settings icon on the checkout step, then click Order Bumps in the left menu.

Step 3. Find the order bump you want to test and click the A/B icon in its Actions column.

CartFlows creates a split test group. Your original order bump is marked Control, and a copy named “Variant” is added as Variation-1. Traffic is split 50/50, and the test starts as PAUSED so you have time to set it up.

Step 4. Click the edit icon on the variation to open its settings, and change the part you want to test. For example, write a new headline in Highlight Text on the Content tab, or choose a different product or discount on the Product tab.

Step 5. To test more than one variation, click Add Variation. Each new variation starts turned off with a 0% traffic share, so switch it to Active and give it a share in the next step.

Step 6. Click the settings (gear) icon in the split test header to open Split Test Settings. Enter the share of traffic for each version, then click Save Settings. The shares need to add up to 100%.

Step 7. Click Start Split Test. The status changes to RUNNING, and shoppers start seeing the different versions.

Your order bump A/B test is now live on that checkout step.

What Shoppers See

Each shopper sees one version of the order bump. In this example, some shoppers see the original “One Time Offer” headline, while others see the variation’s “Add a Glow Boost” headline.

Reading the Results

Every version shows its own numbers under its name:

• Impressions – visitors who saw this order bump.
• Initiated – visitors who added this order bump to their cart.
• Conversions – orders placed that include this order bump.
• Conversion Rate – the percentage of impressions that became orders.

Once the versions collect enough impressions, CartFlows marks the version with the best conversion rate as the winner next to its rate. Give your test enough visitors before you decide, so the results reflect real buying habits.

How to Declare a Winner

Step 1. To pick the Control, click the trophy icon on its row. To pick a variation, click the three-dot menu on its row and choose Declare as Winner.

Step 2. The Declare Winner window compares the versions side by side and shows CartFlows’ recommendation. Review the numbers, then click the Declare as Winner button, which shows the name of the version you picked, to confirm.

The winning version becomes a regular order bump on your checkout step, and the test ends. The other versions are archived and stop showing to shoppers.

Pausing or Changing a Test

Click Stop Split Test to pause the test at any time. While the test is paused, every shopper sees the Control, and you can click Start Split Test again to continue.

To remove a variation from a running test, open its three-dot menu and choose Archive Variation. The test pauses so you can check the traffic shares in Split Test Settings before you start it again.

Frequently Asked Questions