Black Friday analytics checklist: what to watch in GA4, day by day

A Black Friday analytics checklist has two halves. Before the sale: make sure GA4 records every purchase correctly, because data you fail to collect on the day is gone for good. During the sale: open one or two specific reports per day and ignore everything else, because GA4’s standard reports lag behind what is happening on your site.

Black Friday 2026 falls on Friday 27 November, the day after US Thanksgiving (Thursday 26 November). Cyber Monday is Monday 30 November. The checklist below is built around those dates.

Before the sale: the setup checklist (finish by mid November)

A fix shipped on Black Friday morning only covers the hours after it, so finish these before the first promotional email goes out.

1. Place a test order and watch it arrive

Turn on debug mode with Tag Assistant, place a real low-value order (or one with a 100% discount code, if your checkout fires the tag on free orders), then open Admin > Data display > DebugView. DebugView shows the events from your device as they come in.

Check that the purchase event arrives once, and that it carries:

• transaction_id, which Google’s event reference marks as required and which helps GA4 avoid counting the same purchase twice
• value and currency (currency is required whenever you send a value)
• items, also required

An ecommerce event missing a required parameter is treated as a custom event and does not show up in the Ecommerce purchases report.

Then reload the confirmation page. If a second purchase appears, fix it now.

2. Check the funnel events as well as the purchase

GA4 does not collect ecommerce events on its own: your site or your store platform has to send them (Google’s help). The full list of ecommerce events is on Google’s ecommerce measurement page. For Black Friday, the ones you need are view_item, add_to_cart, begin_checkout, add_shipping_info, add_payment_info and purchase.

The Checkout journey report (Reports > Monetization) needs begin_checkout, add_shipping_info, add_payment_info and purchase, one per step. If you are on Shopify, note this line from Google’s documentation: Shopify collects begin_checkout, add_payment_info and purchase automatically, but you have to set up add_shipping_info yourself. Without it, the report has a hole in the middle.

3. Stop payment pages from stealing the credit

If buyers leave your site to pay (PayPal, a bank’s 3D Secure page, a hosted checkout) and come back to a thank-you page, GA4 can credit the payment domain as a referral. Add those domains under Admin > Data streams > your web stream > Configure tag settings > Show all > List unwanted referrals. You need Editor access, and the limit is 50 domains per data stream (Google’s help). Do it now, not on Saturday when PayPal shows up as your second biggest referrer.

4. Tag every Black Friday link with UTMs

Every email, SMS, influencer link and paid social ad needs utm_source, utm_medium and utm_campaign. Google’s URL builder guidance is blunt on two points: values are case sensitive (BlackFriday and blackfriday become two campaigns), and missing parameters show up as “(not set)”. Pick one lowercase campaign name and share it with whoever builds your emails.

5. Mark the key events you care about

purchase is a key event by default. If you also want to count newsletter sign-ups or begin_checkout as key events, mark them now under Admin > Data display > Events (click the star next to the event; you need Editor access). Marking an event only affects data from that moment on, never the past.

6. Write down last year’s numbers

Pull last year’s Black Friday week (Monday 24 November to Monday 1 December 2025, since Black Friday 2025 was Friday 28 November) and note sessions, purchases, conversion rate and revenue per day. You will want them on Saturday morning.

7. Set an hourly purchase alert

GA4 custom insights can evaluate a condition every hour on web data and email you when it triggers. Create one with hourly evaluation on the metric that matters most to you, with either the “has anomaly” condition or a threshold such as a percentage decrease, and turn on the email notification. It is cheap insurance against a checkout that silently stopped firing the tag at 2am. We cover the setup in GA4 custom insights.

During the sale: what to check each day

GA4 data arrives at three speeds (Google’s data freshness page): Realtime within minutes, intraday data within 2 to 6 hours on standard properties, and daily data after about 12 hours. Processing can take 24 to 48 hours, and numbers may change during that window. So each day below uses the report that is actually reliable on that day.

Day (2026)	What is happening	Open this	Watch for
Launch day of your early deals	First email or ad goes out	Realtime, then DebugView on one test order	Purchases appearing in the key events card; your promotion’s landing page showing up in the page views card
Day after launch	First full day of data	Traffic acquisition, filtered on your campaign	”(not set)” or duplicate campaign spellings; fix the links before the big days
Thu 26 Nov (Thanksgiving)	US holiday, early deals running	Realtime, hourly	The custom insight emails; a flat zero on purchases while traffic is up
Fri 27 Nov (Black Friday)	Peak	Realtime, every few hours	Key events still flowing; a channel that should be busy and is not
Sat 28 Nov	Friday is partly processed	Checkout journey for Friday	The step where the drop-off jumped compared with a normal Friday
Sun 29 Nov	Friday is mostly settled	Ecommerce purchases for Friday	Which products sold, which promoted ones did not
Mon 30 Nov (Cyber Monday)	Second peak	Realtime, plus Traffic acquisition for the weekend	Email and paid channels against last year’s notes
Thu 3 Dec onward	Black Friday weekend and Cyber Monday settled	All of the above, plus your store backend	The reconciliation below

Use Realtime for “is it working”, not for “how much”

The Realtime report shows the last 30 minutes, with cards for active users, first user source, views by page title, event count and key events. Use it to confirm that purchases are recorded and your deal pages are getting views, not to judge the day: Google describes it as best effort, cards stop at 700 rows, and its source card only shows how users were first acquired, so Google recommends the acquisition reports for source data.

Compare like with like

When you compare in the date picker, Previous period (match day of week) lines a Friday up with a Friday (Google’s help). For last year, Previous year compares the same calendar dates, which puts this Friday 27 November next to a Thursday, Thanksgiving 2025. The match-day-of-week version for last year no longer exists, so set the 2025 range by hand.

After the sale: reconcile, then read

Wait until GA4 has had 48 hours on Cyber Monday before you write anything down. Then:

1. Count orders in two places. Compare GA4 purchases per day with your store’s paid orders per day. GA4 will usually show fewer, because it only counts orders whose purchase event reached it: anything that stops the tag (a refused consent banner, a blocker, a thank-you page that fails to load) loses the order. What matters is whether the gap on Black Friday looks like the gap on a normal week. A gap that suddenly widened points to a tracking break, not to buyers. If you run Shopify, Shopify Analytics vs GA4 explains which number to trust for what.
2. Read revenue from the store, attribution from GA4. The store is right about money. GA4 is better at telling you which campaign brought the buyer.
3. Look at the week after. A clean Black Friday followed by a slump that runs well past the usual post-sale dip is the real story for December. If traffic falls more than you expected, the traffic drop checklist walks through the diagnosis in order.

Where Dialytics fits (and where it does not)

Dialytics is our product, so read this with that in mind. It is not a live dashboard: it checks GA4 once a day for alerts and once a week for the digest, and posts in Slack.

Every morning it compares yesterday’s sessions, users, new users and key events with the same weekday over roughly the last 12 weeks, and posts at most one alert. On Saturday 28 November it may well flag Black Friday itself as unusual, since no recent Friday looked like it. The useful alerts come before and after: a Tuesday in early November where key events dropped while sessions held, or a December Friday that fell well below a normal Friday. The Monday digest on 30 November covers the week of Black Friday in plain language, and you can ask follow-ups by mentioning the bot in the thread.