What to test before Black Friday and Cyber Monday on Shopify

Work back from the freeze. A test needs at least 7 full days, and usually two to four weeks on a typical Shopify store (see how long a test should run), plus a few days to read the result and ship the winner before the freeze.

Pick tests whose answer you will use during the sale. Each of these changes something you'll show to every BFCM shopper, so knowing which version works is worth more in that week than in any other.

• The offer format. 20% off against $20 off on a typical basket, a flat discount against spend tiers (spend $100 save 10%, spend $150 save 15%), or buy X get Y. Judge it on profit per visitor, not orders: a deeper offer usually sells more and can still earn less.
• The free shipping threshold. A threshold slightly above your typical basket can grow orders during the sale. Work out the break-even first with the free shipping threshold calculator, then test it (the guide).
• The landing pages your ads and emails will send people to. A BFCM collection page, a gift guide, or the product pages of your best sellers: layout, the order of products, how the offer is explained.
• How the offer is announced. A site-wide banner, a message on the product page or a note in the cart. Make sure every version promises exactly what checkout gives.
• Delivery and trust messages. A delivery cut-off date for the holidays, returns policy and reviews near the add to cart button.

Keep each test to one idea and one primary goal, and use fewer variants than usual: there's no time to run a second round if four weak variants split the traffic.

Some tests are fine to keep running through Cyber Monday. Others put the busiest week of the year at risk for an answer you can get in January.

Two Shopify rules are worth checking before the sale. Discounts don't combine unless you set them to: each discount lists which classes (product, order, shipping) it combines with. And a store can have at most 25 automatic discounts active at once, app discounts included. If your BFCM offer, a loyalty perk and a test all use automatic discounts, count them before the freeze.

If you're on Shopify Plus and last year's BFCM promotions ran on Shopify Scripts, they won't run this year: Shopify stopped running Scripts on 30 June 2026. The same logic now needs Shopify Functions, through a discount app or a custom app. Check it well before the freeze.

Traffic is high over BFCM, but the window is short. What decides whether a test can give an answer is how many visitors each version gets, and four days of even a busy store's traffic is often less than a normal three-week test.

• A small store can only detect very large effects in four days. At 2,000 visitors a day, a test from Black Friday to Cyber Monday only reliably detects a lift of about 33%. Know that before you rely on a BFCM test to settle anything.
• Benchmyrk gives no verdict before 7 full days, so a test that runs only Black Friday to Cyber Monday ends without one. That's on purpose: a weekend of sale traffic is not a normal week.
• Sale-week shoppers can be different. Expect more first-time visitors and more people who came for a deal, buying with a discount in the cart. A version that wins with them may not win with December's shoppers. Treat a BFCM result as true for BFCM.
• Novelty fades. A bold banner or a new layout can lift clicks for a few days simply because it is new. Over a four-day window you can't tell that from a lasting effect.
• If the goal is to earn, not to learn, use auto-allocation. It moves traffic toward the version that's ahead as the test runs, so most shoppers get the better one. You give up a precise answer in exchange.

For BFCM the lever is the offer: how deep, in what form, and for which baskets. Base prices are a different question, better answered in a normal month.

• Test the format before the sale, at the same cost. 20% off and $20 off cost the same on a $100 basket; which one shoppers respond to is a cheap thing to learn in October.
• Decide the depth with the margin maths, then test the shortlist. A 20% discount on a 50% margin needs two-thirds more orders just to break even. The prices or discounts post has the formula and a table.
• Judge offers on profit per visitor, after the discount. Revenue per visitor rewards the deepest discount; profit per visitor tells you whether the extra orders paid for it.
• Leave price tests for January. Shoppers compare BFCM prices across stores and with what they paid before. A price test needs a quiet month, consistent ads and a fair way to treat shoppers who notice; see price testing on Shopify.

Two outside rules make prices around a sale delicate. Google Merchant Center asks that the price in your product data match the price on your landing page and at checkout, so different prices for different shoppers can break your Shopping ads. And in the EU, an announced price reduction must show the lowest price you applied in at least the 30 days before it (Article 6a of the Price Indication Directive, added by Directive (EU) 2019/2161). How a price test in the weeks before a sale affects that figure is a question for whoever advises you on consumer law.

1. 1.List every test that will be live over BFCM, who owns it and when it ends. Pause everything else.
2. 2.Open each version with a preview link on a phone and a desktop, including the cart and checkout.
3. 3.Place a test order in each version of every offer or shipping test, and check the discount and shipping charged.
4. 4.Check how test discounts combine with your BFCM discounts, codes in emails and loyalty perks.
5. 5.Set guardrails (orders, errors, missing page elements) so a broken version pauses itself.
6. 6.Keep tests on the same pages mutually exclusive, so no shopper is in two tests that change the same thing.
7. 7.Schedule end dates, so nothing runs on into December by accident.
8. 8.Tell your support team what shoppers in each version see, so they can answer questions about different offers.

• Wait for refunds. Orders placed over BFCM keep coming back for weeks. Compare revenue after refunds before you call an offer the winner.
• Write it down. For each test: the question, the versions, the dates, the result with its interval, and the decision. Next October starts from these notes.
• Don't carry sale winners into January unchanged. A banner that won during a 30% off sale is a hypothesis for a normal month, not a result. Test it again.
• Look at BFCM buyers later. Whether shoppers who came for a deeper offer buy again is a question for February, not for the week after.
• Restart regular testing once traffic settles. Early to mid-December traffic is still unusual; many stores get a cleaner read from January.