The BFCM 2026 readiness plan.

The dates, and what they do to your calendar

Black Friday 2026 is 27 November, Cyber Monday is 30 November, and in practice the trading window now opens in early November and runs to mid December. For Indian brands the festive peak lands earlier, which means the two seasons overlap: your Diwali and BFCM preparation is one project, not two.

Working backwards from 27 November gives you roughly twelve weeks from the start of September. That is enough time to fix infrastructure and build a campaign properly. It is not enough time to also replatform, so if a rebuild is on your roadmap, it either shipped by August or it happens in January.

September: the fix window

Everything in this month is about removing existing weaknesses, while there is still time for changes to settle and for field data to confirm they worked.

• Fix performance, on mobile, using field data. The vast majority of your peak traffic will be mobile. Pull your Core Web Vitals from Search Console and fix what fails, prioritising interaction latency on product and cart pages, because that is where a slow store loses money at peak.
• Audit the app stack. Peak is the worst time to discover that four widgets are fighting over the main thread. An app audit in September buys you speed in November and removes vendors from your risk surface.
• Verify analytics and pixels end to end. Place real test orders and confirm that purchases, values and currencies match across Shopify, GA4 and your ad platforms. Attribution you discover is broken on 28 November is attribution you cannot fix for the season.
• Clean the product data. Titles, images, variant availability, inventory counts and product feeds. Feed errors surface as disapproved ads at exactly the wrong moment.
• Check accessibility. Keyboard operability, contrast and form labelling affect real conversion at volume, and carry compliance exposure if you sell into the EU.
• Confirm your rollback path. Know how you revert a theme change under load, who can do it, and how long it takes. Test it once, in September, on a live store, so it is not theoretical.

October: the build window

• Build campaign pages early and leave them unpublished. Landing pages, collection pages and merchandising all built and reviewed in October, published on a schedule rather than assembled in November.
• Test discount logic at realistic scale. Stacked discounts, bundle rules, tiered thresholds and gift-with-purchase logic all behave differently against a full catalogue and a busy cart than against three test products. If you are on Plus and still have anything running on the retired Scripts framework, that logic needs to be on Checkout Extensibility already.
• Write and test the email and SMS flows. Abandoned cart, browse abandonment, back in stock, post-purchase. Check send limits and deliverability with your provider before you triple your volume.
• Plan inventory buffers and the out-of-stock experience. Decide now what a sold-out hero product does: back in stock capture, recommended alternatives, or a clean message. Deciding in the moment produces a dead end.
• Prepare support. Peak generates several times normal ticket volume. Macros, an updated FAQ covering shipping cutoffs and returns windows, and a clear escalation path to someone technical.

Early November: rehearsal

Two weeks of proving that what you built works, on the real store, under conditions as close to peak as you can arrange.

• Complete a real purchase through every payment method you offer. Every one, on mobile and desktop, including the ones you assume nobody uses. For Indian brands that means UPI, cards, wallets, netbanking and COD, and it means checking that the COD flow, verification and any prepaid incentive behave correctly.
• Test the full discount path end to end. Apply the code, hit the threshold, add the free gift, remove an item, change quantity, and confirm the totals are still right. Discount edge cases are the most common peak-week emergency we get called into.
• Rehearse the launch. Schedule the publish, watch it happen on staging or a preview theme, and time it. Anything that takes a human twenty minutes on the day will take forty under pressure.
• Brief everyone. Who watches the dashboard, who can approve a fix, who can publish, what constitutes an emergency, and how they reach each other at 2am.

The code freeze

Set a freeze date and hold it. Mid November is the usual point, and by then the store should be in the state it will trade in.

Freeze means no theme changes, no new apps, no app updates, no new tracking scripts and no experiments. It does not mean nobody can touch anything: it means changes require a named approver, a tested rollback and a reason that outweighs the risk. Content and merchandising updates through the admin, publishing a scheduled page, adjusting inventory, are all still fine. Structural changes are not.

The reason for the rule is simple. In peak week the cost of a mistake is not the mistake, it is the mistake multiplied by your busiest hour, and almost nothing you could ship in that window earns more than a broken checkout costs.

Peak week: what to watch

• Checkout completion and payment errors. The first place a problem shows up, and the most expensive to miss. Watch gateway failure rates, not just order count.
• Real-time orders against your own forecast. A dip against expectation is a signal to investigate, not to wait and see.
• Site responsiveness on mobile. Check it yourself, on a phone, on mobile data, several times a day.
• Inventory on hero products. Oversells during peak turn into cancellations, refunds and support load in December.
• Ad spend pacing and feed health. A disapproved feed on Black Friday morning is a real scenario and it is silent unless someone is looking.

Afterwards: the part that gets skipped

The week after is worth more than most brands give it. Segment and keep the customers you just bought, because acquisition cost during peak is the highest of your year and a one-time buyer at a discount is not yet a customer. Set up a post-BFCM retention flow before the weekend, not after.

For Indian brands, watch RTO closely through December, since festive and BFCM volume brings a spike in refused deliveries. Then write down what broke, while it is fresh. That document is next September's plan.

Common mistakes

• Launching a redesign in November. The most expensive mistake available. Ship it in January or ship it in August, not into your highest-traffic week.
• Installing new apps in peak week. Every install is untested third-party code on your busiest day, with a support queue that is also at peak.
• Testing discounts on a quiet cart only. Stacking and threshold logic fails on the complicated carts, which are the ones that appear at peak.
• Out-discounting instead of preparing. The brands that post their best numbers generally fixed speed, inventory, feeds and flows months earlier, then discounted the same as everyone else.
• No named owner overnight. Peak traffic does not observe office hours, and a broken checkout at 3am costs the same as one at 3pm.

FAQ