Testing discounts on a share of traffic - a new Shopify Rollouts option
Since October 2, 2026 a Shopify rollout can include not only theme and checkout changes, but discounts as well. That means a new offer can go live to part of your store traffic only, letting you compare conversion against your existing offer before deciding which discount everyone gets.
The practical value is simple: ahead of Black Friday you no longer have to guess whether 15% off earns more than free shipping. You can check it with real shoppers.
What do discounts in Shopify Rollouts mean?
Discounts in Shopify Rollouts are an option, available since October 2, 2026, to include a discount in a rollout alongside theme, checkout and catalog changes. Rollouts are managed in the Shopify admin under Markets > Rollouts.
"Schedule discount changes: Launch a new offer to a portion of online store traffic, then expand when you're ready." - Shopify changelog, 2026-10-02
Until now Rollouts was mostly about design and checkout - we covered that in our article on Shopify Rollouts A/B tests. The same mechanism can now test the commercial offer itself, not just a button colour.
How do you test a discount on a share of traffic?
A discount test in Shopify Rollouts is set up as an Experiment rollout: one group of shoppers sees the new discount, the other sees your store's standard behaviour. In a Shopify experiment the control group gets no discount, so you are comparing the offer against its absence.
Rollouts has three types, and all three are useful for discounts:
- Launch - merges the change into your store permanently. Use it once an offer is approved.
- Event - publishes the change for a set period, then rolls it back. Right for a Black Friday or Christmas promotion that has to end at an exact time.
- Experiment - compares a new variant against a control group so you see the impact before launching to everyone.
A rollout can publish to less than 100% of visitors, and you can review the results in analytics against your live store.
Which Shopify plan do discount rollouts require?
Rollouts work from the Basic plan, while the Experiment type - discount testing against a control group - requires Grow or higher. Discount changes in a rollout can be managed by staff with Markets > Rollouts and Discounts permissions.
That is an important distinction: scheduling a promotion through Rollouts works on Basic, but comparing two offers against each other does not.
What should you check before Black Friday?
Before the season, four limits stated in Shopify's documentation shape how you plan promotions:
- You can add up to 250 discounts to a single rollout. Large promotions built from many individual discounts have to fit that number.
- A discount can be in only one unended rollout at a time. The same discount cannot run in two tests at once.
- A discount keeps its own dates independently. It is active if its own dates apply or a rollout activates it, so the dates and the rollout schedule have to agree.
- Combination settings still apply. Treatment and control visitors may qualify for different discounts, so a combination can work for one group and not the other.
You can edit a discount that is already in a rollout, but Shopify shows a warning when the edit affects the schedule or an experiment's results. Mid-season that means one thing: either let the test run, or stop it and start over.
What a discount test looks like in practice
Say you are weighing two seasonal offers: 10% off the whole cart, or free shipping over 50 EUR. Until now you had to pick one and live with the guess.
- Create both discounts as usual, in the Discounts section.
- Add one of them to an Experiment rollout - it goes to a share of traffic, while the rest of your shoppers see the store without it.
- Run the test long enough to collect a real number of orders. Weekend and weekday behaviour differ, so anything shorter than a week usually shows noise rather than a result.
- Review the conversion results and roll the winning discount out to everyone with a Launch rollout.
One nuance worth planning around: the control group in an experiment gets no discount, so it answers "is this promotion worth it" rather than "which of these two promotions is better". Comparing two offers means two separate tests, run one after the other.
How this differs from simply setting discount dates
Discount start and end dates have not gone anywhere - you can still schedule a promotion without any rollout. The difference is that a rollout does three things dates cannot: launch an offer to part of your traffic, compare it against a control group, and put a discount on the same schedule as your theme and checkout changes.
That last point often matters most in practice. If a seasonal promotion has to start together with a new homepage banner and changed checkout copy, that used to be three separate manual steps in the middle of the night.
Frequently asked questions
Do you need a separate A/B testing app for discounts?
No. An Experiment rollout includes the control group and the result comparison, so a discount test no longer needs an additional app.
How many discounts fit into one rollout?
Up to 250. The same discount can be in only one unended rollout at a time, so it is worth grouping discounts in advance for wide promotions.
Will the promotion end by itself, or does it need to be switched off?
An Event rollout publishes changes for a set period and then restores the previous state on its own. A Launch rollout keeps the change permanently.
Can a discount already in a rollout be edited?
It can, but Shopify warns you when the edit affects the schedule or the experiment's results. Data from a running test will no longer be consistent after an edit.
If you want to start the season with a tested offer rather than a guess - we will review your discount setup and help you put a rollout plan together. Get in touch.
