If your Shopify Black Friday checkout slows down mid-sale, the switch-off order is: session recording and heatmaps, then A/B testing scripts, then cart-page upsell and social-proof widgets, then any checkout UI extension you added in the last quarter, then carrier-calculated shipping rates. Stop there. Do not touch tax, payments, fraud review or inventory sync while orders are flowing. That sequence takes about twelve minutes if someone has admin access and a second screen, and it reverses just as fast.
BFCM is about a week out. Every prep checklist you'll read this month tells you to pre-test, pre-bake collection pages and freeze the theme. Fine. But the theme is frozen, that's the whole point, and at 9:04am Eastern on Friday when p95 time-to-interactive on the cart page is 6 seconds and your Slack is on fire, a prep checklist is useless. What you need is a triage order and a short list of things that are safe to kill while the register is open.
Before you disable anything, find out whose problem it is
Three checks, in this order, and they take four minutes total.
- Shopify's status page. If checkout is degraded platform-wide, nothing you turn off helps, and turning things off while Shopify recovers means you spend Saturday figuring out which toggle broke your attribution.
- Your payment provider's status. Shopify Payments, PayPal, and the BNPL providers all publish their own. A hung Affirm or Klarna widget on the checkout payment step looks exactly like "Shopify is slow" to everyone in the war room.
- Load the store yourself from a phone on cellular, not office Wi-Fi, in a fresh incognito session, with a real product in the cart. Go all the way to the shipping step. Time each step out loud.
That last one is the useful one. It tells you which step is slow, and each step has a different owner. Storefront slow but checkout fine means theme and app-script weight. Cart-to-checkout transition slow means a cart script or a redirect chain. Checkout loads fast but hangs after the address form means shipping rates. Hangs on the payment step means the gateway or a wallet script. Thank-you page slow means post-purchase upsell, and nobody cares, leave it.
Why Shopify gets slow during Black Friday when Shopify itself is fine
Shopify's checkout runs on Shopify's infrastructure and it holds up under genuinely brutal load. Almost every "Shopify checkout lagging under high traffic" ticket we've worked turns out to be one of four things, and three of them are the merchant's.
External HTTP calls made during checkout. Carrier-calculated shipping, third-party tax calculation, third-party fraud scoring, inventory availability lookups against an ERP. Shopify waits on these. Your customer waits on Shopify. A rate endpoint that answers in 300ms in October can sit at 3 seconds on Friday morning because the carrier's own API is under the same load as everyone else's.
Script weight in the storefront path. Theme app extensions, app embeds, legacy ScriptTag injections and whatever the growth team added in September. 180 KB of additional app JavaScript at an effective mobile throughput of 1.5 Mbps is 180 × 8 ÷ 1500 = 0.96 seconds of download before a single byte is parsed. On a Pixel 6a on a congested cell tower in a mall parking lot, parse and execute roughly doubles that.
Your own traffic shaping. A 9am Klaviyo send to the full list, a simultaneous SMS blast, a TikTok Shop promo and a Meta budget uncap all landing in the same ten-minute window. You built the spike. Stagger the sends next time.
Shopify. Occasionally. It happens. The status page will say so.
The kill order: how to disable Shopify apps during a sale without losing orders
Rule one: never uninstall an app mid-sale. Uninstalling can strip configuration, orphan metafields, and leave dead markup in the theme. Reinstalling on Saturday is a worse day than the one you're having. Disable the feature, not the app.
Here's the sequence, lowest risk first. Everything in tiers one through three can be reversed with a single toggle.
Tier 1 — zero order impact, kill immediately. Session recording and heatmaps (Hotjar, Clarity, FullStory and friends). Live chat widgets, unless chat is genuinely staffed and converting. Social-proof and "23 people viewing" popups. Review widgets that load above the fold on PDPs can usually be deferred rather than killed, but on a bad morning, kill them. None of these touch the order path. You lose analysis you were never going to read.
Tier 2 — marketing pixels you can afford to lose for ninety minutes. Settings → Customer events, and toggle off the custom pixels you don't need for attribution. Keep whatever feeds your server-side conversion API, because that's what your Meta bidding runs on. Kill the rest: secondary analytics, affiliate trackers, retargeting pixels for channels you aren't spending on today. A/B testing scripts go here too, and they're often the heaviest single thing on the page because the sane ones block render to avoid flicker. You should not be running a test on Black Friday anyway.
Tier 3 — cart and PDP app features. Upsell and bundle widgets that make an AJAX call on every page load. Recommendation carousels fed by a third party. Currency and geolocation popups. These live in the theme editor's App embeds panel, or in the app's own dashboard. Toggling an app embed off and saving does change the live theme's settings immediately, so it is technically a live change, but it isn't a code deploy and it isn't covered by a sensible code freeze. One click back.
Tier 4 — checkout UI extensions. Open the checkout editor and hide the blocks you added for this campaign: the gift-message field, the upsell block, the custom trust banner, the delivery-date picker. Each one is a network round trip during a step where the customer already has their card out. If you're on Plus with multiple checkout profiles, publish the lean profile you should have built in October. If you didn't build one, edit the live one and write down every change you make.
Carrier-calculated shipping is the biggest single lever, and it costs you money
This is the honest trade-off in the whole runbook. If the checkout hangs between the address step and the shipping step, the cause is almost always a rate request going out to UPS, USPS, FedEx, or a rate-engine app sitting in front of them, and timing out or crawling.
Do the arithmetic before you decide. Say the shipping step's p95 goes from 0.4s to 3.1s. At 1,200 checkouts an hour, that's 2.7 × 1,200 = 3,240 seconds, 54 minutes of collective waiting per hour, spread across people who are three taps from a competitor's tab. Switching to pre-built flat rates clears it instantly.
What it costs: you're now eating the difference between the flat rate and real cost on heavy, far-shipping orders. Model your two or three worst zone-8 scenarios in advance and set the flat rate so the blended margin survives. A merchant doing mixed-weight apparel can usually set flat rates that lose a couple of dollars on the tail and come out ahead. A merchant shipping 40 lb items cannot, and should instead cut the rate app's timeout and accept a fallback rate. Either way, decide the number now, not at 9am.
What you don't switch off, no matter how slow it feels
- Tax calculation. Economic nexus doesn't take Friday off. Turning off Avalara or changing tax settings mid-sale to shave 200ms creates a filing problem that lasts years. Leave it.
- Payments and wallets. Shop Pay, Apple Pay and Google Pay are faster than your form. Removing a BNPL option mid-sale changes the mix of who can afford the cart.
- Fraud and risk tooling. Chargebacks arrive in January with interest.
- Accessibility overlays or remediation you've deployed. ADA demand letters are a real line item for US storefronts and the filing volume doesn't care that it was your busiest day.
- Shopify Functions. Discount and shipping Functions run on Shopify's side under a hard resource cap. They're rarely the cause of user-visible latency, and a half-disabled discount stack is the fastest way to turn a slow day into a pricing incident.
Pause the background jobs nobody thinks about
Not customer-facing, still a cause of storefront degraded performance during a peak sale. Your ERP or 3PL inventory sync hammering the Admin API eats rate limit and can collide with inventory reservations at checkout. A full catalogue price push mid-sale is worse. Bulk operations from a PIM, a feed export, a nightly report that somebody scheduled for 9am Pacific without thinking about time zones.
Pause them for three hours. Inventory drift of three hours on BFCM is a reconciliation job for Monday. A checkout that times out is lost revenue.
A BFCM monitoring checklist you can run while the sale is live
Two screens, one person watching, fifteen-minute cadence.
- Shopify's live view: sessions, carts, checkouts, orders. The ratio matters more than the absolute number. Checkouts flat while sessions climb is your first signal, usually ten minutes before anyone complains.
- Synthetic checks every 60 seconds against home, a top collection, a top PDP and cart, from at least two US regions. If you only run one, run it from the opposite coast to your office.
- Real-user Core Web Vitals, segmented to mobile. Desktop numbers will lie to you on BFCM.
- Status pages, pinned in one tab group: Shopify, your payment providers, your BNPL providers, your email and SMS platform, your top three apps.
- A manual end-to-end test purchase every half hour on a real phone. Automation misses the thing where the Apple Pay sheet takes four seconds to appear.
- Support inbox volume. Three "is your site down?" emails in ten minutes beats any dashboard.
Between sales, this is a tooling problem rather than a staffing one. A monitor that tracks PageSpeed and flags regressions as apps get added is the cheap version of a person watching a screen: SwiftStore is the one we built for it. The deeper work of getting app-script weight down before the peak is a separate project, and the time for it is Q3, not now. We cover the approach in Shopify speed optimization.
Who to call when Shopify checkout is slow
Write this list down this week and put it somewhere that isn't Slack, because if the day is bad enough you'll be on a phone.
Shopify support is your first call for anything that looks platform-side. If you're on Plus, you have a direct escalation path and you should have confirmed who's on it before the holiday, not after. If you're on Advanced or below, you're going through standard support, and you should know that going in rather than discovering it at 9:15.
Your agency or developer, if they're on call. Ask about this explicitly when you engage anyone: is someone reachable on BFCM Friday, by phone, and at what cost. A lot of shops say yes and mean "we'll read email on Monday."
App vendors. Check now whether each critical app's support is a human or a form. The ones with 24/7 coverage during BFCM usually announce it in early November. The ones that don't, you handle by toggling off and asking questions later.
Payment provider and carrier account reps. Mostly you're confirming a known incident rather than getting a fix, but confirmation is worth something when the CEO wants an explanation.
If the storefront is actually down, not slow
Different problem, shorter list. Check the status page. Check whether it's DNS, which you'll spot because the domain fails but the myshopify.com URL resolves. Check whether a custom script or a theme change somebody pushed "just a tiny fix" twenty minutes ago is the cause, and if a change went out in the last hour, roll it back before you diagnose anything.
Then communicate. Post on the channels where your traffic is: an Instagram story, a tweet, a line in the email you're about to send. "We're getting hammered, hold tight, your code works all weekend" buys you more goodwill than silence and it extends the window so people come back. If you took payment and can't confirm orders, say so before customers start filing disputes.
Turning it all back on
Keep a running note with a timestamp for every toggle you flip. Not in your head. The failure mode we see most often isn't the incident, it's the Tuesday after, when attribution looks wrong for three days because a pixel never went back on and nobody remembers switching it off.
Restore in reverse order, one tier at a time, with ten minutes between tiers and a look at the dashboard after each. If the slowdown returns at tier three, you've found your culprit and you now know what to fix in January.
If you want a second pair of eyes on the storefront before the weekend, our free audit covers script weight and the checkout path specifically, and we work US hours from Bengaluru. For merchants who'd rather have someone on call than a PDF, the US engagement page explains how we handle peak-season cover.


