Shopify Rollouts Now Handle Discounts. Test Your BFCM Offer on a Slice of Traffic Before You Commit the Margin.
Use this before Black Friday. On October 2, 2026, Shopify added discounts to Rollouts. You can now launch an offer to a portion of your online store traffic, compare conversion against the control, and expand only if it wins. If you plan to run 20% off sitewide because that is what you ran last year, you now have a way to find out whether 15% off with a gift threshold converts the same at a better margin. You have about four weeks to run that test.
What changed
Per Shopify’s changelog entry dated October 2, 2026, discounts can now be included in a Rollout. Three things matter:
- You can launch a new offer to a portion of online store traffic, then expand when ready.
- You can test different discount offers and compare conversion results before a full launch.
- You can coordinate discount changes with theme and checkout changes in one rollout. Existing customer eligibility rules are kept.
Per Shopify’s Rollouts help page, a rollout can cover your main theme, checkout and accounts pages, discounts, and product catalogs, alone or combined. There are three types: Launch (changes merge permanently), Event (changes publish temporarily, then revert automatically), and Experiment (treatment is compared to control before full deployment). You schedule by date and time, and you review analytics against your live store afterward. It lives under Markets > Rollouts in admin. Shopify notes plan availability and staff permission requirements apply, and the changelog says the feature is in the new admin. Check both on your store before you plan around it.
Shopify’s pages do not publish the traffic split options or minimum sample sizes. Check what the Experiment setup screen offers on your store rather than assuming.
Why this matters for BFCM
Most brands pick the BFCM discount by gut or by last year’s number, then find out the cost in December. A discount is a margin decision multiplied by every order. Two uses of this feature are worth setting up now.
Event rollouts for the promo window. Schedule the offer to publish on a date and time and revert on its own. No one has to be awake at midnight to flip a discount off, and no one forgets and leaves it running into Tuesday.
Experiment rollouts for the offer itself. Run the candidate offer against control and read conversion before you put it in front of all traffic.
The math
Illustrative. Replace with your own numbers.
Say a product sells at $60 with a $24 landed cost (COGS, freight, fulfillment), so $36 gross margin per unit before ad spend. Store traffic is 100,000 sessions during the promo window.
| Offer | Price | Margin/unit | Conversion | Orders | Gross margin |
|---|---|---|---|---|---|
| Control (no promo) | $60.00 | $36.00 | 2.0% | 2,000 | $72,000 |
| 20% off | $48.00 | $24.00 | 3.0% | 3,000 | $72,000 |
| 15% off + free gift (gift cost $4) | $51.00 less $4 | $23.00 | 3.0% | 3,000 | $69,000 |
| 10% off | $54.00 | $30.00 | 2.6% | 2,600 | $78,000 |
In this example, 20% off lifts orders 50% and adds zero gross margin. Ten percent off converts less but earns $6,000 more. You will not know which row your store looks like without testing, and the answer changes by category and by how much of your traffic is returning customers. Orders and revenue will look best on the 20% offer. Margin will not. Judge the test on gross margin per session, not conversion rate alone:
- Control: $72,000 / 100,000 = $0.72 per session
- 20% off: $0.72 per session
- 10% off: $0.78 per session
Sample size is the trap. At a 2% baseline, small traffic slices produce noisy conversion reads. A swing from 2.0% to 2.4% on a few hundred sessions is not a result. If you get under roughly 1,000 orders per variant during the test, treat the read as directional and do not overrule your margin math with it. Run the test for at least a full week so weekday and weekend behavior both show up.
What to do this week
- Confirm access. Open Markets > Rollouts. Check your plan, your staff permissions, and that you are in the new admin.
- Pick two or three offers, not ten. Your default plus one lower-depth variant plus one value-add (gift or threshold) is enough. Each extra variant splits traffic and slows the read.
- Set the success metric before you launch. Gross margin per session, with a floor on conversion. Write it down so the result cannot be argued after the fact.
- Start the test by mid-October so you have a clean week or two of data before Black Friday, and leave time to set the winner as an Event rollout.
- Exclude low-margin SKUs from any offer whose math only works on your best sellers.
- Schedule the BFCM window as an Event rollout with explicit start and end times in your store time zone, and have someone confirm the revert after Cyber Monday.
- Watch your paid traffic mix. If paid social drives most of your sessions, and your ads promise a specific offer, a test that shows a different offer on the site will hurt conversion for reasons unrelated to the discount. Align ad copy with the offer in each variant, or test only on traffic you can control.
One limit to plan for: this tests conversion on your online store. It will not tell you what a discount does to repeat purchase rate, return rate, or customer quality. A deep discount can bring in buyers who never come back or who return more. Check cohort return and reorder rates in January against the offer each cohort got.
The bottom line
A discount is the most expensive lever in your Q4 plan and the one least often tested. Shopify just made it cheap to test. If you run Shopify and plan any BFCM offer, spend an afternoon this week setting up one experiment. Even a directional read beats repeating last year’s number.
If you want help designing the offer test, setting up the rollout schedule, or modeling BFCM margin by SKU before the window opens, book a call.
More in Shopify & DTC
See all Shopify & DTC insights →