GA4 may be missing half of your ecommerce orders

Ahead of the year-end campaign season, ecommerce budget decisions are made by looking at GA4 reports. An audit published by Search Engine Journal questions how far those reports can be trusted: Christian Goodrich put the GA4 data of eight UK ecommerce brands next to the stores' own order records. The author notes that he works for a digital agency that has worked with the brands.

Why does GA4 undercount orders?

Six brands could be matched, and at three of them GA4 told a very different story from the store records. At one brand, GA4 recorded 81 per cent of web orders in the last quarter of 2024 and 41 per cent in the last quarter of 2025. At two brands the cookie consent banner stayed off for days once the plan's monthly page-view limit ran out, and GA4 gave no warning. At another, GA4 recorded no purchases for five weeks, then loaded the orders late and counted hundreds of them twice.

At one brand, GA4 showed 4,100 purchases in the last quarter of 2025 against 4,501 store orders. The total looked reasonable, but double counts and gaps had cancelled each other out. In September another brand's sessions were inflated by bot traffic, while Search Console clicks did not move.

How does missing data distort budget decisions?

Retailer A's web orders rose from 5,874 to 7,686 while GA4 showed a decline and cut Google Ads return from 8.07 to 4.79. In an attribution tool matched to orders, the same return rose from 4.42 to 7.93. In the author's view, cutting channels on GA4's numbers would have meant pulling budget from channels that were working.

What should ecommerce teams in Türkiye do?

• Compare the last two years' quarters week by week: divide GA4 purchases with distinct transaction IDs by store orders. A sudden drop, a near-zero week or a week above store orders is a tracking problem.
• Size your cookie consent plan's monthly limit for peak-season traffic, and make sure limit warnings reach marketing as well as developers.
• Divide GA4 Google organic sessions by Search Console clicks each week; when tracking breaks, the ratio drops suddenly.
• Send the purchase from the store system when the order is created, using the Measurement Protocol, rather than from the browser, so a page refresh cannot create a duplicate. The author suggests trialling the change in a separate property ahead of time, not in November.
• Agree the source rule before the budget meeting: orders and revenue come from the store, behaviour from GA4. Measurement and conversion tracking work builds that split into reporting.