Creating a Split Test (A/B)
Split Testing (A/B) is available for each offer type within an upsell funnel, including the In-Checkout offer type. We recommend implementing a split test at every funnel position you're able to so that you're constantly optimizing your offers and their conversion rates.
With OCU Plus, you can also run multivariant A/B/C/D split tests. More details
For Pre-purchase and Post-purchase, there's even the ability to split test any/all offer types against one another.
For Pre-purchase for example, you can split test "add to order single" vs "add to order multi", "add to order single" vs "upgrade order" or any other combination of available offer types.
Similarly for Post-purchase, you can split test the "add to order single" vs "add to order multi", "add to order single" vs "best selling products (AI)", "recommended products (AI)" vs "best selling products (AI)" or any other combination of available offer types, by selecting the "Different Product" option when creating a new split test:
Example ideas for types of Split Tests you can run to find your best converting offers include:
One product vs another (i.e. offering more of the same product vs a complementary product).
Different discount levels and/or discount types ($ vs %)
Different text/copy for the pages (Product description, Content 1, Content 2)
Different Images (Content 1, Content 2)
Short-form content vs Long-form content
To create a split test for an offer, follow the steps below:
1. On an offer card, hover over the "Create split test" button:
2. You'll see 2 options provided; Same Product or Different Products:
Choosing "Same Product", OCU creates an exact copy of the offer page so you can split test the same product against itself, but with an offer page design or price variation. This way you can see which offer page design for the offer product is performing best in that specific upsell funnel position.
Choosing "Different Product", you can split test a different product in that upsell funnel position to see which one performs better there I.e., has a higher $ / Visit or conversion rate.
For Pre-Purchase and Post-purchase offers, if you select to choose a different offer you will first be prompted to select the type of offer. You can choose the same type of offer or a different type of offer for your split-test.
3. Once you select Product B (Same Product or Different Product), the Split Test will be created! The page will update to show which offer product is Page A (the control) and which is Page B (the variation) for the test.
In any split-test, if you select the Hide offer setting you may notice more views for one offer over the other. These because, even in a split test, the offer wouldn't be shown if the product is already in the cart. The data for the split test would still be accurate (the Conversion and $ / View stats provide the most useful information) but there wouldn't be an exact 50/50 split of page Views in this situation.
Starting a Split Test
Creating the split test will not cause it to work until you actually start the test. Until the test starts, 100% of traffic will continue to go to Page A.
To start the test, click the "Start split test" button.
Then click the "Start Test" confirmation button in the pop-up shown.
Pages will be split test evenly at a weight of 50/50.
Important Notes:
📝Starting a split test will reset the statistics (number of views, conversion rate, and revenue counter) to 0 for the particular offer position where the test is started. However, this will not affect the overall funnel stats.
📝 If Upsell Item A is out of stock, then the offer won't be shown even if Item B is in stock. If Upsell Item B is out of stock, but Item A is in stock then the offer will be shown 50% of the time.
Pausing or Stopping a Split Test and Picking a Winner
Statistics for your running offer page split tests update as soon as the action is tracked. So if someone is accepting an offer, you can check it almost in real-time by refreshing the Funnel Builder page.
Once the split test has run for some time and you're satisfied with the amount of data gathered for it (TBD by the merchant), it's time to either pause or stop the test and pick a winner.
Click the "Pause test" button to Pause the test. Once the test is paused, it will direct all traffic to Page A until you either choose a winner or resume the test. *To change the offer's status, the split test needs to be paused first.
In the Paused state, you can make edits to either of the offer pages involved in the split test if desired.
Once the test is Paused, you have 3 options available:
Resume Test
Choose Winner
Reset Statistics (resets all statistics to 0 for the split test)
Click on one of the three buttons depending on which action you'd like to take.
If you click the "Choose winner" button, you will be prompted to pick either Offer A or B as the winner. Once you make your choice, the split test will be completed and all traffic will be directed to the winning offer variant:
I added a split test to my funnels, is there anything else I should know?
Whether or not you choose to split-test your offers, it's time to track how they are doing! Learn how to track their performance with by checking out the Funnel and Offer Statistics! Click here to learn more.
OCU also offers a Statistics Suite for overall data about app and funnel performance, click here to find out more about this feature!
Best Practices:
Running split tests is going to allow you to tailor and personalize upsell to your audience.
One product vs another (i.e. offering more of the same product vs a complimentary product).
Different discount levels and/or discount types ($ vs %)
Different text/copy for the pages (Product description, Content 1, Content 2)
Different Images (Content 1, Content 2)
Short-form content vs Longer form content