Overview
Split testing lets you run two Slide Cart Drawer designs at the same time and compare how they actually perform, instead of guessing. When you start a test, OCU splits your storefront traffic evenly between your current Live cart and a second variant, then tracks each cart's conversion rate, AOV, and revenue so you can pick a data-backed winner.
Before this, the only way to compare two cart designs was to switch between them manually and eyeball the analytics across different weeks. That is unreliable, since traffic and seasonality change from one period to the next. A split test removes that guesswork by running both carts side by side under the same conditions.
Before you start
A few things to know before you set up your first test:
You need an active/live OCU Slide Cart Drawer. The test always compares your Live cart against a second variant, so a split test can't be created without one.
You can run one split test at a time. Once a test exists in any status, the option to create another is unavailable until you finish or delete the current one.
Split tests are managed from the Cart Drawer List page, using the ⋯ action menu on your Live cart.
📝 If you use a duplicate as your second variant, it counts toward your 15 Cart Drawer limit. If you are already at 15 carts, the duplicate option won't be available (you can still use an existing Draft cart instead).
How to create a Split Test
You create a split test from your Live cart, then choose a second cart to test against it. Follow the quick and easy steps below:
Go to the Cart Drawer List page.
Open the ⋯ menu on your Live cart and select "Create Split Test".
In the popup modal, choose your Variant B:
Duplicate of the Live cart creates a copy of your current Live cart to test against. You can then edit the copy to try a different design.
An existing Draft cart lets you test against a cart you've already built. This option only appears if you have at least one Draft cart.
Confirm your choice by clicking the "Create Split Test" button. The split test is created with "Not started" status.
While a test is in Not started status, all of your traffic stays on your original Live cart. Nothing changes for your shoppers until you start the test.
📝 The "Create Split Test" option is disabled if you have a Draft cart scheduled, or if a split test already exists. Hover over the option to see why. It also won't appear at all in the ⋯ menu of a Draft cart, since a test always starts from your Live cart.
Starting the Test
When you're ready to send live traffic to both carts, start the test.
On the split test, click the Start button.
A warning appears letting you know that statistics for both carts will reset to 0. This is expected. During the test, each cart's stats reflect only the traffic gathered while the test is running, so you're comparing the two designs on equal footing.
Confirm to begin the test.
Once started, the test moves to Running status:
Traffic splits 50/50 between Cart A and Cart B.
Both cart toggles show ON but are locked, so neither cart's status can be changed while the test runs.
A day counter shows how long the test has been running, for example 3 Days. If less than a full day has passed, it shows < 1 Day.
📝 Your other Draft carts stay available while the test runs, and you can still edit them. Just note that if you publish a Draft cart while a test is running, it ends the test. You can't schedule a Draft cart during an active test.
Split Test Actions
Once a test exists and is running, you manage it from the action icons on the split test row: pause or resume, select a winner, reset statistics, and delete. Here's what each one does:
Pausing and Resuming
You can pause a running test at any time, then resume it later:
Pause sends all traffic back to your original Live cart (100% Cart A, 0% Cart B). The day counter keeps counting while paused, since it tracks total time since the test first started. While paused, you can edit either cart's page.
Resume returns traffic to the 50/50 split and continues the test.
Selecting a Winner
When you've gathered enough data, choose the cart that performed best. You can select a winner from either Running or Paused status.
Click the Select Winner button.
In the "Choose Winner" popup, select either Cart A or Cart B as the winner.
Confirm your selection by clicking the "Publish winner & end test" button.
The winning cart becomes your Live cart and stays active. The losing cart moves to Draft so you don't lose the design. All split test labels, traffic indicators, and status badges are removed from the list.
📝 What happens to your statistics: Ending a test does not permanently erase the numbers you saw before it. Each cart's final stats are its pre-test statistics plus the stats gathered during the test, merged back together.
Resetting Statistics
Reset Statistics (available while Running or Paused) clears the aggregated stats for both carts back to 0 so you can start measuring fresh. A confirmation modal appears before anything is reset.
📝 Your stats are safe. This only resets the split test data. Your original cart statistics are restored once the split test ends.
Deleting a Split Test
Delete (available in any status) ends the split test without picking a winner. A confirmation modal explains what happens: Cart 2 reverts to a Draft, and Cart 1 stays Live. All statistics collected during the test will be added back to each cart's total stats:
👉 Important Notes
A few interactions and limits to keep in mind while a split test is running:
Scheduling is paused for the duration. You can't schedule a Draft cart while a split test exists. If you try, you'll see the tooltip "Finish the split test first." End or finish the test first.
Carts in a test can't be deleted. Neither Cart A nor Cart B can be deleted while they're part of an active test.
Publishing another Draft cart ends the test. Your Draft carts stay available during a test, but making one of them Live will end the split test. When you flip a Draft cart's Status toggle to Live, a Publish Cart confirmation appears ("Publishing this cart will make it live and cancel the pending split test"). Confirm with Publish & End Test, and that cart becomes your new Live cart while the split test ends.
Shoppers get a consistent experience. Once a shopper is assigned a cart variant, they keep that same variant for their entire session, including through an active checkout. A returning shopper may be assigned a different variant in a later session while the test is still running.
FAQs
Does starting a test reset my cart statistics?
Does starting a test reset my cart statistics?
Your stats reset to 0 visually when the test starts, so you compare both carts only on traffic gathered during the test. When you select a winner, each cart's earlier stats are restored and merged with the stats from the test, so nothing is permanently lost.
Can I run more than one split test at a time?
Can I run more than one split test at a time?
No. You can have one split test at a time. The option to create a new one becomes available again after you select a winner or delete the current test.
Will a shopper see both cart designs?
Will a shopper see both cart designs?
No. Each shopper is assigned one variant and keeps it for their whole session, including through checkout. This keeps their experience consistent and your data clean.
What happens to the cart that loses?
What happens to the cart that loses?
It moves to Draft, so the design is saved. You can keep editing it, use it in a future test, or publish it later.
Can I edit a cart while the test is running?
Can I edit a cart while the test is running?
While a test is Running, both carts are locked. If you pause the test, you can edit either cart's page, then resume when you're ready.
Do I have to pick a winner?
Do I have to pick a winner?
No. You can delete the test at any time instead. When you delete, Cart B returns to Draft and Cart A stays Live.