Open Shopify and GA4 side by side on the same month and the two revenue figures will not match. They have never matched, they are not supposed to match, and the effort most brands pour into making them match is effort taken from the decisions the data was supposed to support.
The useful question is not why the numbers differ. It is which number belongs in which conversation, and how far the gap can drift before it stops being physics and starts being a broken tag.
The two systems are counting different things
Shopify counts orders on its own servers. Money moved, the order exists, it is in the ledger whether or not a browser cooperated. GA4 counts a purchase event fired by JavaScript in the customer's browser on the confirmation page. Every order that reaches Shopify has to survive a second, far more fragile journey to reach GA4.
Three things eat that journey. Ad blockers hide somewhere between 15 and 30 percent of web traffic from analytics tools, and the people running them skew younger, more technical and higher income, which for a lot of DTC brands is the segment you most want to understand. Consent is the second. Under GDPR a customer who declines tracking never fires the event, and in the EU the share of users declining or blocking is high enough to move the number on its own. The third is plain mechanics: iOS privacy restrictions, a customer closing the tab before the confirmation page finishes loading, a payment method that redirects away and never comes back.
Add those together and 10 to 12 percent is the normal gap on a properly configured store. Twenty to thirty percent is common. With a highly technical audience it can reach 50 percent without anything being broken at all.
When the gap is telling you something is wrong
Track the ratio, not the difference. Divide GA4 transactions by Shopify orders for the same 30 day window and watch that single number month on month. Between roughly 0.70 and 0.90 you are in normal territory. A sudden drop means a tag broke, usually during a theme change, an app install or a consent banner update. A number above 1.05 means you are double counting.
Duplicates are the more dangerous failure because they flatter you. The usual cause is two systems firing the same purchase event: Shopify's native Google channel integration alongside a Google Tag Manager container, or a third-party app that also ships its own tag. GA4 deduplicates only when the transaction_id matches exactly, so if one source sends the Shopify order ID and another sends its own reference, both land and both count.
A brand that under-reports by 25 percent makes cautious decisions. A brand that over-reports by 15 percent scales a channel that was never working. Undercounting costs you growth. Overcounting costs you cash.
There is a specific reason to check this now. Shopify has been retiring the old checkout customisation path, with Plus stores cut off in August 2025 and everyone else in August 2026. Additional Scripts and checkout.liquid customisations stop firing at that point. The risk sits in the migration window, where a developer adds a proper custom pixel and leaves the legacy snippet in place, and you get clean textbook two times duplication until somebody notices.
What GA4 is genuinely good at
Once you stop asking GA4 to be a finance tool, it becomes a useful one. Its strength is relative comparison across channels and behaviour inside the funnel, both of which survive a consistent undercount. If GA4 misses a quarter of your orders roughly evenly, the shape of the picture still holds even when the absolute figure does not.
Attribution is the other half. GA4 now runs data-driven attribution as its default, looking at up to 50 touchpoints across the 90 days before a conversion and distributing credit by modelled contribution rather than by rule. For accounts with reasonable conversion volume, that surfaces meaningfully more assisted contribution than last click does, particularly for the upper-funnel channels that last click has always made look worthless. Check your lookback window while you are in there. Acquisition conversions default to 30 days, which understates early touchpoints for any brand with a considered purchase and a long research cycle.
Five reports carry most of the decisions: traffic acquisition for channel comparison, ecommerce purchases for product level demand, purchase journey for where the funnel leaks, funnel exploration for custom paths such as a quiz or a subscription upgrade, and cohort exploration for whether the customers a channel brings back actually return. Everything else in the interface is interesting rather than actionable.
The case for server-side tracking
Sending order data from the server rather than the browser removes ad blockers and browser restrictions from the equation. Capture typically moves from around 90 percent to above 98 percent. For a brand spending meaningfully on paid media, that is not a vanity improvement in a dashboard.
The conversions you were losing were never spread evenly. They clustered in the audiences most likely to block tracking, which means every bidding algorithm reading that signal was being fed a systematically distorted view of who converts. Fixing capture improves the signal going out to the ad platforms as much as it improves the reporting coming in. Consent Mode v2 sits alongside it, letting the modelled portion of blocked traffic be estimated rather than simply discarded.
What this looks like in practice
Write the rule down and hold everyone to it. Shopify and your payment processor own every financial number: revenue, refunds, anything that reaches the P&L or a board pack. GA4 owns channel comparison, funnel behaviour and assisted contribution. A post-purchase survey owns the question of where customers say they heard about you. Three tools, three jobs, no arguments in the Monday meeting about which screenshot is right.
Then run the reconciliation monthly and log it. GA4 transactions over Shopify orders, one number, plotted over time. You are not chasing 1.00. You are watching for the month where the ratio moves more than a few points, because that is a tag, an app or a consent change, and catching it in week one instead of week nine is the entire value of the exercise.
Audit the firing sources once while you are at it. List every system that could send a purchase event, keep exactly one, and confirm it sends the real Shopify order ID as the transaction_id. Most of the duplicate problems I see in DTC accounts are a tag somebody installed for a two-week test in 2024 and never removed.
Finally, stop reading GA4 revenue against ad spend to calculate ROAS. GA4 undercounts orders and platforms overcount their own contribution, so the two errors push in opposite directions and the answer can be wrong by a factor that changes the decision. Contribution margin per order, built from Shopify revenue and real costs, is the only version of that maths worth acting on.
Inside the system
How we build this for brands
For the brands we work with, we build profit and cash-flow dashboards from live Shopify and ad data so the financial numbers come from the ledger and the behavioural numbers come from analytics, without either one pretending to be the other. A reporting agent runs against that data weekly and surfaces leakage or risk before it compounds, including the tracking ratio itself, so a broken tag shows up as an alert rather than as a quiet three-month hole in a channel report.
Underneath it sits the unglamorous work: one purchase event, the real order ID, server-side where the spend justifies it, and a consent setup that is both compliant and not silently deleting a third of your data. Part of this runs live for portfolio brands today; the full system is what we deploy when we take a brand on.
Analytics Audit
Find out which of your numbers you can actually trust
We reconcile your GA4, Shopify and ad platform data into one view, find the duplicate events, the broken tags and the consent gaps, and tell you which reported wins are real. You get the true ratio, the leakage, and the order to fix it in.
Book Your Analytics AuditFrequently asked questions
Why is GA4 revenue lower than Shopify?
Shopify counts paid orders from its own servers. GA4 counts purchase events fired in the customer's browser. Anything that stops that browser event firing removes the order from GA4 while Shopify still banks it. Ad blockers, declined cookie consent, iOS privacy restrictions, customers closing the tab before the confirmation page loads and tagging errors all take a slice. A 10 to 12 percent gap is normal for most stores, 20 to 30 percent is common, and with a young or technical audience it can reach 50 percent.
What is a normal GA4 to Shopify revenue discrepancy?
Treat 10 to 30 percent as the normal band, with 10 to 12 percent typical for a well configured store on client-side tracking alone. Below that band and you are likely double counting somewhere. Above it and you either have a broken tag, a consent banner blocking more than it should, or an audience that runs ad blockers heavily. Track the ratio monthly rather than reacting to a single week.
How do I fix duplicate purchase events in GA4 on Shopify?
Check your GA4 transaction count against Shopify orders across 30 days. A ratio above 1.05 points at duplication rather than noise. The usual causes are two systems firing the same event, typically Shopify's native Google channel app plus a Google Tag Manager container or a third-party app, or the purchase tag firing again when a customer refreshes the confirmation page. GA4 only deduplicates when the transaction_id matches exactly, so make sure every source sends the real Shopify order ID, then remove the redundant tag rather than filtering it later.
Should DTC brands use server-side tracking for GA4?
If paid media is a material share of your acquisition, yes. Server-side tracking sends order data from the server rather than the browser, which takes ad blockers and browser restrictions out of the equation and typically lifts capture from roughly 90 percent to over 98 percent. The value is not a prettier revenue number. It is that your channel comparison and your ad platform signal both improve, because the conversions that were previously invisible cluster in specific audiences rather than spreading evenly.
Which GA4 reports should a DTC brand actually use?
Five carry most of the decisions. Traffic acquisition for channel comparison, ecommerce purchases for product level demand, purchase journey for where the funnel leaks, funnel exploration for custom paths such as quiz or subscription flows, and cohort exploration for whether the customers a channel brings in come back. Everything else is interesting rather than actionable, and the revenue figure itself should come from Shopify or your payment processor.
Can I use GA4 as my source of truth for revenue?
No. Use Shopify or your payment processor for anything financial, including revenue, refunds and the numbers that reach your P&L. GA4 is a behavioural and relative performance tool. It is good at telling you that organic search is trending up faster than paid social, and unreliable at telling you exactly how many pounds either produced. Brands get into trouble when a GA4 number lands in a board deck next to a Shopify number and nobody reconciles the two.
About the author
Caner Veli is a DTC operator who has helped 350+ brands fix broken growth engines. He built Liquiproof from zero to 3,000+ global retailers in under 6 years. He now runs the same playbook, supported by AI systems he built himself, for DTC and CPG brands.