A/B testing campaigns for promotions
Last updated: 29-Feb-2024
Rate this article:
Overview
Set up A/B testing campaigns for your promotions to identify what type of discount is more efficient.
Requirements & limitations
- You can test up to five promotions in one experiment.
- You can only select pairs of promotions that have at least one common product.
- Promotions applied through the COUPON parameter directly into the URL are not taken into account in the A/B test. "Auto-apply" promotions, however, are taken into account. Nonetheless, promotions with the “Auto-apply” function will not appear in the A/B testing available promotions drop-down.
- Only regular promotions are available for use in A/B testing campaigns.
Important
When the shopper places the order, the test is considered successful regardless of the order or payment status.
Setup
- Go to the Promotions page and define the promotions that you want to test. Choose between percentage-based discounts and fixed discounts.
- Navigate to the A/B Testing page and click the Campaigns tab.
- Scroll until you find the Promotions section and click Edit Campaign on an inactive campaign.
- Fill in the details of the new campaign, such as the campaign name, maximum number of tests and the date when the campaign should end.
- Create a new scenario by selecting a promotion from the drop-down menu and assigning the percentage of traffic you want to go through it. You can choose between existing promotions and assign them as A/B testing campaign variants. You can also test specific promotions against using no promotion at all by checking the No promotion option.
- Click Add Scenario.
- Repeat steps 5 and 6 until you have used the entire available traffic.
- Click Save campaign.
- Go back to the Campaigns tab and click the Start Campaign button when you want the campaign to begin.
Where can I see what promotions are used in my A/B tests?
When you open a promotion's configuration page, a notification informs you if the promotion is being used in an A/B test at that moment.
Rate this article: