Black Friday & Cyber Monday Tracking Checklist 2026: 131 Checks for GA4, GTM, Meta CAPI, Google Ads and Server-Side
Short answer: Before Black Friday and Cyber Monday, make sure every purchase is counted once, with the right value, on every checkout type, in GA4, Google Ads, Meta and TikTok. Then freeze tracking changes two days before the sale, place a real test order on a phone every morning of the sale, and compare store orders with each platform every 1–2 hours. The 15 checks below catch most of the problems I find in pre-sale audits.
Most Black Friday “performance problems” I audit are really tracking problems. A pixel that stopped after a Shopify update. Purchases counted twice. A gclid stripped by a redirect app. The ads keep spending, Smart Bidding and Meta’s algorithm learn from bad data, and nobody notices until the weekend is over.
This is the list I go through so that doesn’t happen: 131 checks across 14 areas, each with how to check it, where, what “broken” looks like, how urgent it is and when to do it.
Prepared by Jobair Mahmud, tracking and analytics specialist (Top Rated on Upwork, 175+ projects). Platform facts were checked in November 2026. Platforms change fast: if something looks different in your account, trust the account.
Short on time? Do these 15 first
These are the Critical checks that catch most of the problems. Find the full row by ID further down.
- SHP-01 Purchase tracking does not rely on Additional Scripts or checkout.liquid
- SHP-05 Each sales channel app is connected once, with no duplicate Purchase sources
- SHP-09 Shop Pay, Apple Pay, Google Pay and PayPal orders all fire purchase
- GTM-01 The GTM container loads only once per page
- GTM-06 Purchase sends one unique
transaction_idto every platform - EVT-04
add_to_cartfires from quick add and the cart drawer, not only the product page - EVT-09 GA4 purchase has the right value, currency, coupon and items
- EVT-10 Reloading the thank-you page does not create a second purchase
- META-02 Conversions API sends the same events as the Pixel
- META-03 Meta browser and server events are deduplicated (same
event_id) - GADS-01 Only Purchase is a Primary conversion in Google Ads
- GADS-02 Only ONE primary purchase action (not GA4 import + Ads tag)
- GADS-05
gclidsurvives redirects, popups and the cart - CNS-02 Consent Mode v2 default state loads before any tag
- LIVE-01 A real test order on mobile every morning of the sale
Timeline (Black Friday = Fri 27 Nov 2026, Cyber Monday = Mon 30 Nov)
| When | What happens |
|---|---|
| Now (T-7) | Access, inventory, platform setup. Anything that needs approval or data history. |
| T-3 | End-to-end testing on real devices, every checkout type. |
| T-1 | Freeze changes. Baselines, alerts, dashboards, team on-call plan. |
| Sale days | Monitor, compare, log. Roll back, don’t hot-fix. |
| After the sale | Clean up, reconcile, save data, write the retro. |
Priority: Critical = if this is broken, sale data or ad optimisation is wrong; fix before anything else. High = causes real data loss or bad decisions; fix before the freeze. Medium = makes reporting better and next year easier.
| Area | Checks | Critical |
|---|---|---|
| 1. Access & change control | 8 | 3 |
| 2. Shopify checkout & Customer events | 11 | 6 |
| 3. Data layer & Google Tag Manager | 11 | 2 |
| 4. GA4 property setup | 11 | 1 |
| 5. GA4 ecommerce events | 13 | 4 |
| 6. Meta Pixel + Conversions API | 15 | 5 |
| 7. Google Ads conversion tracking | 11 | 4 |
| 8. TikTok & other ad pixels | 5 | 0 |
| 9. Consent & privacy | 7 | 1 |
| 10. Server-side tracking | 6 | 2 |
| 11. UTMs, codes & attribution | 9 | 0 |
| 12. Dashboards, alerts & speed | 7 | 1 |
| 13. Sale days war room | 9 | 3 |
| 14. After the sale | 8 | 0 |
| Total | 131 | 32 |
1. Access & change control
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| ACC-01 | Admin access on every platform you’ll touch during the sale | Log in today, with the account you’ll use on Friday: Shopify (staff), GTM (Publish), GA4 (Editor), Meta Business Manager (Admin), Google Ads (Admin), TikTok Business Center, Merchant Center. | Friday 9 PM, tracking breaks and the only person who can publish GTM is asleep. | Critical | T-7 |
| ACC-02 | A second person has emergency access | Add one backup admin to each platform. Check their 2FA actually works. | One lost phone or one sick person = no fixes for 4 days. | High | T-7 |
| ACC-03 | GTM container backup exported | GTM > Admin > Export Container. Save the JSON with today’s date. | No clean rollback point when a late change breaks checkout tags. | Critical | T-3 |
| ACC-04 | No unpublished GTM workspace changes | Workspace Changes = 0, or every pending change is documented and owned. | A small fix gets published with a half-built test tag. | High | T-3 |
| ACC-05 | Tracking change freeze date agreed | Pick a freeze (I use T-2). After that, only fixes for broken tracking. Post it in the team chat. | A “quick tweak” on Thursday night breaks purchase for the whole weekend. | Critical | T-3 |
| ACC-06 | Tag inventory: every pixel, where it’s installed, who owns it | One list: tool, events, install method (theme code / app / GTM / Customer events / server). | Two apps and GTM all send Meta Purchase. Nobody knows which to switch off. | High | T-7 |
| ACC-07 | Dead pixels and old apps removed | Check theme.liquid, app embeds and Customer events for old agency pixels and abandoned test tags. | Extra JS slows mobile pages and sends duplicate events. | Medium | T-7 |
| ACC-08 | Every checkout path that leaves your domain listed | PayPal, Klarna, Afterpay, Shop Pay, subscription apps, third-party checkouts. Each needs a test order. | Redirect payment methods never get tracked. | High | T-7 |
2. Shopify checkout & Customer events
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| SHP-01 | Purchase tracking no longer depends on Additional Scripts or checkout.liquid | Settings > Checkout: Thank you / Order status pages on the new version; Additional Scripts empty or read-only. Plus stores lost these on 28 Aug 2025, non-Plus on 26 Aug 2026. | Purchase tags silently stopped after the auto-upgrade. | Critical | T-7 |
| SHP-02 | Every checkout pixel lives in Settings > Customer events | Each app and custom pixel shows “Connected”. | “Disconnected” = zero checkout events since that day. | Critical | T-7 |
| SHP-03 | Custom pixels use the right privacy permission | Customer privacy > Permission (Marketing / Analytics / Preferences) and “Data sale” match what the pixel does. | Pixel blocked for consented visitors, or firing for visitors who said no. | High | T-3 |
| SHP-04 | Custom pixel purchase listens to checkout_completed | Code uses analytics.subscribe('checkout_completed', ...), not a page view on the thank-you URL. | Purchase fires again every time the order status page is reopened. | Critical | T-7 |
| SHP-05 | Native sales channels connected once, right data sharing level | Facebook & Instagram app: Maximum (Pixel + CAPI). Google & YouTube app: GA4 + Ads linked. TikTok connected. Not also sending the same events from GTM. | Platforms show 2–3× the real orders. | Critical | T-7 |
| SHP-06 | One revenue definition, used everywhere | e.g. subtotal after discounts, excluding tax and shipping. Same field in GA4, Meta, Google Ads, TikTok. | GA4 AOV $118, Meta $131, Google Ads $109. Nobody trusts any dashboard. | High | T-7 |
| SHP-07 | Currency sent on every value event | currency on view_item, add_to_cart, begin_checkout, purchase. Multi-currency: presentment vs shop currency, consistently. | EUR or GBP orders read as USD. | High | T-7 |
| SHP-08 | Discounted order tested in every active Shopify Market | Use the real BF code in at least two markets. Check value after discount and currency. | Revenue inflated by the full discount. | High | T-3 |
| SHP-09 | Express checkouts fire purchase | One test order each with Shop Pay, Apple Pay, Google Pay, PayPal. | Express buttons skip begin_checkout; PayPal return drops purchase. | Critical | T-3 |
| SHP-10 | Post-purchase upsell revenue tracked correctly | Accept an upsell on a test order. Extra purchase or value update, never double counted. | Upsell revenue missing or the order reported twice. | Medium | T-3 |
| SHP-11 | Not on Shopify? Checkout and thank-you pages excluded from cache | WooCommerce / custom: caching plugin and CDN exclude cart, checkout and order-received URLs. | Cached thank-you page sends the same order for different customers. | Critical | T-7 |
3. Data layer & Google Tag Manager
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| GTM-01 | GTM container loads once per page | Tag Assistant: one GTM-XXXX per page. Check theme code and apps. | Every GA4 and Ads event counted twice. | Critical | T-7 |
| GTM-02 | Ecommerce data layer follows the GA4 schema | Preview on product, cart, checkout: items with item_id, item_name, price, quantity, item_variant, item_brand, item_category. | Products show “(not set)”; remarketing can’t match. | High | T-7 |
| GTM-03 | ecommerce: null pushed before each ecommerce event | Check the push before add_to_cart, begin_checkout clears the previous object. | Items from lists leak into add_to_cart. | Medium | T-7 |
| GTM-04 | item_id matches Merchant Center and Meta catalog IDs | Compare 3 SKUs: data layer ID vs feed ID (SKU vs variant ID vs shopify_US_xxx). | Dynamic remarketing and PMax can’t match products. | High | T-7 |
| GTM-05 | Prices are numbers, not text | 49.00, not “$49.00” or “49,00”. | Revenue shows 0 or the event is rejected. | High | T-7 |
| GTM-06 | Purchase sends a unique transaction_id | Shopify order ID or number, the same value in every platform. | No dedup; refreshing the thank-you page = new sale. | Critical | T-7 |
| GTM-07 | Each tag fires exactly once per action | Click Add to cart once in Preview and count fires. | Add to cart counted twice. | High | T-7 |
| GTM-08 | Tags have consent settings configured | Advanced settings > Consent settings on every tag. | Tags fire before consent where that isn’t allowed. | High | T-3 |
| GTM-09 | Tested on iPhone Safari, not only desktop Chrome | Real iPhone: add_to_cart and purchase reach GA4 DebugView and Meta Test Events. | Desktop passes, Safari drops events. | High | T-3 |
| GTM-10 | Hero banners send view_promotion / select_promotion | Add promotion_id and promotion_name to banner clicks. | No way to know which offer drove sales. | Medium | T-3 |
| GTM-11 | Final version published with a clear name | e.g. “BFCM 2026 – final – 24 Nov”, with a description. | Nobody knows which version to roll back to. | Medium | T-1 |
4. GA4 property setup
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| GA4-01 | Site sends data to the correct property | Measurement ID in Tag Assistant matches Admin > Data streams. | Sale data lands in a property nobody looks at. | Critical | T-7 |
| GA4-02 | Data retention set to 14 months | Admin > Data retention > Event data retention. | Next November you can’t compare in Explorations. | High | T-7 |
| GA4-03 | Internal traffic filters Active | Admin > Data filters: Active, not Testing. | Your own test orders mixed into conversion rate. | Medium | T-7 |
| GA4-04 | Payment gateways in unwanted referrals | paypal.com, checkout.stripe.com, klarna.com, afterpay.com and others you use. | Sales credited to “paypal.com / referral”. | High | T-7 |
| GA4-05 | Cross-domain measurement set where needed | Configure domains, then check _gl appears between domains. | Each domain switch starts a new “direct” session. | High | T-7 |
| GA4-06 | Purchase is a key event; page_view and scroll are not | Admin > Key events. Keep the list short. | Junk “conversions” imported into Ads. | High | T-7 |
| GA4-07 | Google Ads and Merchant Center linked | Admin > Product links. | No GA4 audiences in Ads, no product data. | High | T-7 |
| GA4-08 | BigQuery export switched on (not retroactive) | Admin > BigQuery links > daily export. | No raw event data after the sale. | Medium | T-7 |
| GA4-09 | DebugView shows your test device | Walk the full funnel and watch every event with parameters. | “Tested” in Realtime, which hides missing parameters. | High | T-3 |
| GA4-10 | Custom dimensions registered | e.g. customer_type, payment_type, discount_code. | Parameters collected but never visible. | Medium | T-7 |
| GA4-11 | Annotations for sale start, code changes and email sends | Add annotations in GA4 reports. | In January nobody remembers why traffic jumped. | Medium | T-1 |
5. GA4 ecommerce events (walk the funnel)
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| EVT-01 | view_item_list on collections, search and homepage rows | Include item_list_id / item_list_name. | Can’t tell if the BF collection page worked. | Medium | T-3 |
| EVT-02 | select_item on product card clicks | Same item_list_name as the list. | Broken path from list to product. | Medium | T-3 |
| EVT-03 | view_item on every product page with price and variant | Change variant: price and item_variant update. | Old (non-sale) price in reports. | High | T-3 |
| EVT-04 | add_to_cart from product page, quick add, upsells and cart drawer | Test each Add button separately. | Quick add (your main BF button) sends nothing. | Critical | T-3 |
| EVT-05 | remove_from_cart | Remove in the drawer and on /cart. | Cart value never goes down. | Medium | T-3 |
| EVT-06 | view_cart on cart page and drawer open | Fires once per open, full items array. | No cart-to-checkout data. | Medium | T-3 |
| EVT-07 | begin_checkout from cart AND express buttons | Express buttons often skip it. | More purchases than checkouts. | High | T-3 |
| EVT-08 | add_shipping_info and add_payment_info | On Shopify, from Customer events. | Can’t see if the drop is at shipping or payment. | Medium | T-3 |
| EVT-09 | purchase: transaction_id, value, tax, shipping, currency, coupon, items | Compare every field with the Shopify order. | Value includes tax in GA4 but not in Ads; coupon empty. | Critical | T-3 |
| EVT-10 | purchase fires once | Reload the thank-you page and reopen the order link from the email. | Each reload = new revenue. | Critical | T-3 |
| EVT-11 | search with search_term | Enhanced measurement or custom event. | You miss what shoppers searched for. | Medium | T-3 |
| EVT-12 | Refund plan ready | App, Measurement Protocol or manual import. | December returns are invisible. | Medium | T-1 |
| EVT-13 | GA4 matches Shopify within your normal gap | Compare yesterday’s purchases and revenue; write down the gap %. | You can’t spot a real break on Friday. | Critical | T-1 |
6. Meta Pixel + Conversions API
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| META-01 | Pixel fires all standard events | PageView, ViewContent, AddToCart, InitiateCheckout, AddPaymentInfo, Purchase (Pixel Helper, Test events). | Retargeting audiences empty. | Critical | T-7 |
| META-02 | Conversions API sends the same events | Events Manager > Overview: Browser and Server for each event. | iOS and ad-blocker purchases lost. | Critical | T-7 |
| META-03 | Browser and server events deduplicated | Same event_name + event_id; check deduplication details. | Purchases counted twice; budget scaled on fake ROAS. | Critical | T-7 |
| META-04 | Event Match Quality for Purchase is good (aim 7+, minimum 6) | Events Manager > event > Event Match Quality. | Weaker optimisation when CPMs are highest. | High | T-7 |
| META-05 | Customer information parameters on server events | Hashed em, ph, fn, ln, ct, st, zp, country, external_id + fbp, fbc, IP, user agent. | CAPI adds little. | High | T-7 |
| META-06 | fbc captured from fbclid and kept to checkout | Land with ?fbclid=test123, complete checkout, check fbc. | Ad clicks not linked to purchases on Safari. | High | T-3 |
| META-07 | Purchase value, currency and content_ids correct | content_ids match the catalog, value after discount. | Advantage+ catalog ads can’t match items. | Critical | T-3 |
| META-08 | Only one source sends Purchase | App OR GTM OR tracking app, unless deduplicated by design. | Duplicates dedup can’t fix. | Critical | T-7 |
| META-09 | test_event_code removed from production | Search server container / CAPI app config. | Events show in Test events only. | High | T-1 |
| META-10 | Diagnostics tab clean | Fix missing params, duplicates, domain or value issues. | Meta warned for weeks and nobody looked. | High | T-7 |
| META-11 | Domain verified | Brand safety > Domains: Verified. | Problems with event configuration. | Medium | T-7 |
| META-12 | Catalog has BF sale prices, no rejected sale products | sale_price + sale_price_effective_date; check rejections. | Ads show full price, site shows BF price. | High | T-3 |
| META-13 | Team knows Meta’s 2026 attribution change | 7-day view and 28-day view windows were removed on 12 Jan 2026. Default is 7-day click + 1-day view. Compare years on the same window. | “ROAS dropped vs last year” panic. | High | T-1 |
| META-14 | Server events sent in real time | Timestamps close to the order time. | Optimisation reacts hours late. | Medium | T-3 |
| META-15 | BFCM audiences built before launch | Purchasers 180 days, ATC 7/14 days, IC 7 days, engagers, email list. | Audiences still filling when the sale ends. | Medium | T-7 |
7. Google Ads conversion tracking
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| GADS-01 | Only Purchase is Primary for bidding | Goals > Conversions > Summary; everything else Secondary. | PMax spends the BF budget on add-to-carts. | Critical | T-7 |
| GADS-02 | Only ONE primary purchase action | Not both an Ads tag purchase and a GA4-imported purchase. | Double counted conversions, wrong tROAS. | Critical | T-7 |
| GADS-03 | Dynamic value and transaction_id on the purchase tag | Tag Assistant on a test order. | Every order valued at 1. | Critical | T-7 |
| GADS-04 | Enhanced conversions on, diagnostics healthy | Since 2026, web and leads are one setting. Confirm it’s on and user data is sent. | Toggle on, but no hashed email reaches Google. | High | T-7 |
| GADS-05 | Auto-tagging on, gclid survives the journey | Visit ?gclid=test123 and click through redirects, popups, cart. | Redirect app strips gclid. | Critical | T-7 |
| GADS-06 | Conversions with cart data | Items + merchant ID, feed country, feed language. | No product-level insight for Shopping and PMax. | Medium | T-7 |
| GADS-07 | Purchase counting = Every | Conversion action settings. | Repeat buyers counted once. | Medium | T-7 |
| GADS-08 | Consent Mode signals visible for EEA / UK traffic | Diagnostics show consent status. | Remarketing and conversions shrink in Europe. | High | T-7 |
| GADS-09 | Merchant Center: sale prices and promotions submitted early | sale_price + effective date; Promotions need review. | Promotion still “under review” on Friday. | High | T-7 |
| GADS-10 | Seasonality adjustment scheduled | Tools > Bidding > Seasonality adjustments (short events, big CVR jump). | Smart Bidding under-bids on Friday. | High | T-1 |
| GADS-11 | Know how to apply a data exclusion | Tools > Bidding > Data exclusions. Practise once. | A 6-hour outage teaches bidding your campaigns stopped working. | High | T-1 |
8. TikTok & other ad pixels
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| TT-01 | TikTok Pixel and Events API live and deduplicated | Same event_id in TikTok Events Manager. | Double counted purchases. | High | T-7 |
| TT-02 | TikTok funnel events fire with value | ViewContent, AddToCart, InitiateCheckout, AddPaymentInfo and the purchase event, with value, currency, content_id. | Campaign optimises for an event that never fires. | High | T-3 |
| TT-03 | Advanced matching on | Pixel settings + Events API payload. | Weak match, weak optimisation. | Medium | T-7 |
| TT-04 | ttclid captured and passed to server events | Land with ?ttclid=test, purchase, check the server event. | TikTok clicks not linked to orders. | Medium | T-3 |
| TT-05 | Pinterest, Snap, Microsoft Ads: same 4 checks | Fires? Deduplicated? Correct value? Respects consent? | A small channel quietly double counts. | Medium | T-3 |
9. Consent & privacy
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| CNS-01 | Cookie banner doesn’t cover Add to cart on mobile | Test on a small phone where the banner shows. | Conversion rate drops and you blame the offer. | High | T-3 |
| CNS-02 | Consent Mode v2 default loads before any tag | Tag Assistant > Consent: default, then update after the click. | Tags fire with no consent state. | Critical | T-7 |
| CNS-03 | ad_storage update fires on “Accept” | Since 15 June 2026 Google uses Consent Mode as the single control for GA4 data shared with Ads. | Fewer conversions, smaller remarketing lists, no visible error. | High | T-7 |
| CNS-04 | Advanced vs Basic consent mode decision documented | Write down which one and why. | Nobody knows why EU conversions are low. | Medium | T-7 |
| CNS-05 | CMP wired into Shopify’s Customer Privacy API | Accept / reject and check Customer events follow it. | Banner says rejected, pixels still fire. | High | T-7 |
| CNS-06 | Meta / TikTok pixels respect consent | Reject cookies, check Pixel Helper. | Compliance risk on your busiest week. | High | T-7 |
| CNS-07 | Privacy policy lists every new BFCM tool | Review app, popup, SMS tool, pixel. | Policy out of date. | Medium | T-3 |
10. Server-side tracking
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| SST-01 | Server container can scale for BF traffic | Cloud Run: raise max instances, keep min instances. Stape: plan request limit vs expected traffic. | Request quota hit on Saturday; server events stop. | Critical | T-3 |
| SST-02 | Tracking subdomain works, SSL not about to expire | Open the health URL; check certificate dates. | Expired certificate = every server event fails. | Critical | T-7 |
| SST-03 | Hosting billing is valid | Card not expired, no unpaid invoice, budget alert set. | Account suspended during the sale. | High | T-7 |
| SST-04 | GA4 client claims requests, transport_url correct | sGTM Preview: requests claimed, tags return 200. | Browser sends to the server, server sends nowhere. | High | T-7 |
| SST-05 | Server error logs checked | 4xx/5xx from Meta, GA4, TikTok, Google Ads in the last 7 days. | Errors for weeks nobody reviewed. | High | T-3 |
| SST-06 | Webhook purchase deduplicated with the browser purchase | Same event_id / transaction_id. | Every order counted twice. | High | T-7 |
11. UTMs, codes & attribution
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| UTM-01 | UTM naming rules written and shared | Lowercase; source = platform, medium = channel, campaign = bfcm_2026_offer. | “Facebook”, “facebook”, “fb” = three sources. | High | T-7 |
| UTM-02 | Every BF email and SMS link tagged | Check Klaviyo / Omnisend / Postscript settings; click every link in a test email. | Email revenue shows as direct. | High | T-3 |
| UTM-03 | Paid social ads use dynamic URL parameters | e.g. utm_campaign={{campaign.name}}&utm_content={{ad.name}}. | Can’t see which creative sold. | High | T-3 |
| UTM-04 | Influencers and affiliates have unique UTMs AND codes | One link + one code per partner. | Can’t pay or judge partners fairly. | Medium | T-7 |
| UTM-05 | Short links and QR codes keep UTMs, gclid, fbclid | Click each and check the final URL. | Traffic becomes direct. | High | T-3 |
| UTM-06 | Discount code to channel map ready | Code → channel → owner. | Post-sale fight about who sold what. | Medium | T-7 |
| UTM-07 | GA4 custom channel group for SMS, affiliates, influencers | Admin > Channel groups. | Big Unassigned bucket. | Medium | T-7 |
| UTM-08 | Post-purchase survey live | “How did you hear about us?” on the thank-you page. | Word-of-mouth sales invisible. | Medium | T-7 |
| UTM-09 | Source of truth agreed | Shopify = revenue, GA4 = behaviour, ad platforms = their optimisation signals. | Saturday spent arguing about numbers. | High | T-1 |
12. Dashboards, alerts & speed
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| MON-01 | Live dashboard built and tested | Shopify orders, GA4 purchases, Meta purchases, Google Ads conversions, spend, revenue. | Building reports on Friday instead of watching them. | High | T-1 |
| MON-02 | Baseline numbers saved | Last 4 weeks: daily orders, CR, AOV, mobile share, plus BFCM 2025. | Can’t tell “slow” from “broken”. | High | T-1 |
| MON-03 | Normal gap between Shopify and each platform known | Calculate from the last 30 days. | Panic over a normal gap, or a real break ignored. | Critical | T-1 |
| MON-04 | GA4 alerts for purchase drops | Insights > Create, sent to 2 people. | Tracking dies at 1 AM, nobody knows. | High | T-1 |
| MON-05 | Checkout funnel exploration saved by device | view_item > add_to_cart > begin_checkout > purchase, mobile vs desktop. | Can’t see where people drop. | Medium | T-1 |
| MON-06 | Page speed checked after all BF apps are added | PageSpeed Insights on homepage and a top product page, mobile. | Tracking perfect, LCP 5s, buyers leave. | High | T-3 |
| MON-07 | Alert channel and on-call person defined | One group, one person per time block for 4 days. | Everyone assumes someone else is watching. | High | T-1 |
13. Sale days war room (Fri–Mon)
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| LIVE-01 | Morning test order on mobile, every sale day | Buy the cheapest product with the live code, check every platform within 30 min, refund. | A plugin auto-updated overnight and broke tracking. | Critical | Sale days |
| LIVE-02 | Compare orders across platforms every 1–2 hours | Gap above normal + 10 points = investigate. | Problem found Monday instead of Friday. | Critical | Sale days |
| LIVE-03 | Watch Meta Events Manager | Overview per event and Diagnostics. | Server events stopped and nobody looked. | High | Sale days |
| LIVE-04 | Watch Google Ads conversion action status | “No recent conversions” or “Inactive” on Purchase. | Bidding runs blind for hours. | High | Sale days |
| LIVE-05 | If tracking breaks: roll back, don’t hot-fix | Restore the last good GTM version first. | Live debugging creates a second problem. | Critical | Sale days |
| LIVE-06 | Log every outage window | Start and end time, for data exclusions and honest reporting. | Bad data mixed into BFCM results. | High | Sale days |
| LIVE-07 | Re-test after any site change | New banner app, landing page or theme swap = re-test add_to_cart and purchase. | Cyber Monday page launched without tracking. | High | Sale days |
| LIVE-08 | Cyber Monday links and codes tagged | Same UTM and code-map rules. | Cyber Monday revenue lands in direct. | High | Sale days |
| LIVE-09 | End-of-day snapshot | Export orders, revenue, spend, platform conversions. | No record of what you saw when you made decisions. | Medium | Sale days |
14. After the sale
| ID | Check | How to check | Red flag | Priority | When |
|---|---|---|---|---|---|
| POST-01 | Seasonality adjustment ended, data exclusions applied | Window matches the real sale dates; exclusions for logged outages. | Bidding expects BF conversion rates in December. | High | After |
| POST-02 | Refunds and cancellations sent to analytics | GA4 refund events or import. | BFCM looks better than it really was. | Medium | After |
| POST-03 | Final numbers reconciled | Gap per platform and the reason for each. | Wrong numbers reported to the owner or client. | High | After |
| POST-04 | BFCM data exported and saved | GA4, ad platforms, Shopify orders. | No clean data for next year. | Medium | After |
| POST-05 | BFCM first-time buyers tagged as a cohort | Shopify tag or segment for 60/90-day repeat rate. | You never learn if BF buyers are worth it. | Medium | After |
| POST-06 | Sale-only scripts and apps removed | Countdown timers, urgency popups, temporary pixels. | “Ends tonight” banner still live in January. | Medium | After |
| POST-07 | Parked fixes shipped | Everything noted as “after sale”. | Same issues at Christmas and next BFCM. | Medium | After |
| POST-08 | One-page tracking retro written | What broke, how fast you found it, what to fix next time. | Same mistakes next November. | Medium | After |
After Cyber Monday: Christmas, Boxing Day and January
The sale doesn’t end on Monday. Most stores change offers, shipping and pages again in December, and each change can break tracking:
- Re-run the 15 first checks before every new promotion (Christmas gift guides, Green Monday, Boxing Day, January sale), and after any theme or app change.
- Last-order dates and shipping changes often come with new landing pages, banners and popups. Test add_to_cart and purchase on them.
- Gift cards and gift wrap: decide whether gift card purchases count as revenue in ads (usually not for ROAS bidding), and make sure add-ons don’t create a second purchase.
- Returns in January: send refunds to GA4 so your Q4 report shows net revenue, not just orders.
- Keep seasonality adjustments short. Use them only for the real spike days, not for the whole of December.
For agencies: running this across client stores
If you manage several ecommerce clients, run the 15 first checks on every account in the same week, freeze changes on the same day, and give one person ownership of the live monitor for all of them. I work white-label with agencies on exactly this: pre-sale audits, fixes under your brand and NDA, and on-call cover across the sale weekend. See white-label tracking for agencies, the agency Q4 tracking playbook and the agency onboarding tracking audit checklist.
Frequently Asked Questions
When should I finish my Black Friday tracking setup?
At least a week before the sale (T-7) for access, platform setup and anything that needs approval, like Merchant Center promotions. Test on real devices at T-3, and freeze tracking changes about two days before Black Friday.
What are the most important Black Friday tracking checks?
Make sure each purchase is counted once (unique transaction ID, deduplicated Meta Pixel and CAPI events, one primary purchase conversion in Google Ads), that express checkouts like Shop Pay and PayPal fire purchase, that the gclid survives your redirects, and that Consent Mode v2 loads before any tag.
How do I know if my tracking breaks during the sale?
Know your normal gap between Shopify orders and each platform before the sale. On sale days, compare every 1–2 hours. If the gap grows by more than about 10 percentage points above normal, investigate. Place a real test order on a phone every morning.
What should I do if tracking breaks on Black Friday?
Roll back to the last good GTM version instead of debugging live, log the start and end time of the outage, and apply a data exclusion in Google Ads so Smart Bidding doesn’t learn from the broken period.
Does server-side tracking help on Black Friday?
Yes, if it’s set up well. Server-side tracking and Meta CAPI recover purchases that browsers block. But the server must be able to scale for sale traffic, the SSL certificate and billing must be valid, and server events must be deduplicated with browser events.
Why does Meta show lower ROAS than last Black Friday?
Partly because Meta removed the 7-day view and 28-day view attribution windows on 12 January 2026. Compare 2025 and 2026 using the same attribution window before deciding performance dropped.
Short on time before Friday?
If you find Critical items you can’t fix yourself, I can audit and fix your GA4, GTM, Meta Pixel + CAPI, Google Ads and TikTok tracking before the sale. Send me the check IDs that are red and I’ll tell you what it takes.
Free to use and share. Please keep the credit.
Need this implemented, audited, or white-labeled?
Whether you are an eCommerce brand owner fixing signal loss or a digital marketing agency needing a senior analytics engineer under strict NDA, I deliver 24–48h SLA turnaround with 8.5+ EMQ match rates.


