GA4 Missed 68% of One Retailer's Christmas Orders: Audit Finds Five Failure Modes

An audit of Google Analytics 4 at eight U.K. ecommerce brands found GA4 recorded just 32% of one retailer's web orders over the Christmas period — 66 percentage points below the store's own records. The analyst behind the audit, a digital agency practitioner who has worked with all eight brands, compared GA4's Q4 figures for 2024 and 2025 against store order data ahead of this year's peak season.

Between Christmas Day 2025 and Jan. 4, 2026, GA4 told the retailer — labeled Retailer A — that revenue was down 68% year over year. The store's own order records showed a fall of 13%. Across the quarter, GA4 captured 81% of Retailer A's web orders in Q4 2024 and 41% in Q4 2025.

Six of the eight stores allowed order-by-order matching. At three, GA4 told a very different story from the store. Two had cookie banners switching themselves off for days or weeks at a time, with no signal in GA4 that data was missing. A third recorded no purchases for five weeks, loaded them late and counted hundreds twice. A fourth was flooded with bots in September. The auditor's verdict: most are setup problems that could happen at any store.

What can go wrong — five failure patterns

• A steady gap. One brand's GA4 captured 76% of web orders in both Q4 2024 and Q4 2025; another ran between 72% and 84% every week of Q4 2025. That is the shape cookie consent produces. The fix: use GA4 for trends, not totals.
• A cookie banner that switched itself off. Retailer A's sessions dropped out in blocks ending on the 14th of each month; Retailer B's blocks ended on the 20th. Retailer B's GA4 recorded 94% of store orders in Q4 2024 and 36% in Q4 2025, and showed revenue down 69% while net sales rose 18.5%. The cause: consent plans with monthly page-view allowances. Retailer B captured up to 68% of orders in weeks the banner ran and 10% to 22% in weeks it switched off. Retailer A's allowance ran out Dec. 1 — the Black Friday cycle — and GA4 did not recover until Dec. 15.
• A blackout and late backfill. From Oct. 28 to Dec. 4, 2025, Retailer C's store recorded 2,215 orders; GA4 recorded none. On Dec. 9, GA4 logged 1,261 orders in one day, 1,000 with order numbers from inside the gap.
• The same order counted twice. Google's guidance says GA4 "deduplicates purchase events with the same transaction ID" — but only from the same user. At Retailer C, 657 of 3,365 order numbers in Q4 2025 carried more than one purchase event, adding 735 phantom purchases. The quarterly total of 4,100 against 4,501 store orders looked almost fine; the weekly detail did not.
• Visitors who weren't shoppers. Retailer D's GA4 recorded 23,187 sessions on Sept. 27, 2026, and 25,622 the next day, against a mid-September baseline of 300 to 900. Almost all were direct, desktop Chrome visits averaging about a second. Search Console clicks did not move.

How bad data distorts budget decisions

Retailer A's Q4 web orders rose from 5,874 to 7,686 while GA4 showed them falling by a third. GA4 put Google Ads' return on spend at falling from 8.07 to 4.79 and organic search revenue nearly halving. A premium multi-touch attribution tool matched to store orders told the opposite story: Google Ads' return rose from 4.42 to 7.93 on 20% less spend, and revenue attributed to organic and direct visits rose 58%. Retailer C's GA4 put Google revenue, paid and organic, down 86% while the store's net sales rose 4.8%.

The auditor's rule: judge channels together against store orders, because models can differ but should not point in opposite directions. Cutting either channel on GA4's numbers would have pulled budget from channels that were working.

What should marketers do before peak?

The audit prescribes a weekly check. Divide GA4's Google organic sessions by Search Console clicks — in Retailer C's blackout, Google sent 4,244 clicks in the week of Nov. 3 while GA4 recorded 57 organic sessions. Export weekly web orders for Q4 2024 and Q4 2025, pull GA4 purchases as distinct transaction IDs (not the transactions metric, which counts refunds), and divide. A step change, a near-zero week or a week above 100% signals a tracking problem.

For stakeholders, lead with the capture rate: "GA4 recorded 41% of our web orders this quarter, down from 81% last year" is a sentence any finance director understands. The auditor also recommends moving purchase tracking server-side — sending the purchase event from the store's own system through the Measurement Protocol, keyed on the order number and carrying the shopper's client and session IDs — while keeping the browser tag for behavior. Google says the Measurement Protocol is meant "to augment automatic collection," not replace it.

The final word is blunt: GA4 is the best free tool most ecommerce teams have for behavior, but it is not a record of sales. Count orders where they happen — in the store.