Skip to content

Shopify, GA4 and Meta Disagree: Which Number Is Right?

Shopify counts orders, GA4 counts the purchase events its tag recorded, and Meta counts the purchases it can link to its own ads. All three can be right at once. A budget decision needs a fourth number none of them shows: what each channel caused.

Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free

By , Founder & CEOPublished 10 min read

Run the numbers for your store: the free ad platform over-reporting checker.

Shopify, GA4 and Meta never agree because each one counts something different. Shopify counts orders. GA4 counts the purchase events its tag recorded and splits the credit for each one among channels by a rule. Meta counts the purchases it can link to one of its own ads, inside its own window. All three can be right at the same time. A budget decision needs a fourth number that none of them shows: what each channel caused.

What does each system count, and when?

ShopifyGA4Meta
What it countsOrders placed in your shopPurchase events its tag sentPurchases it can link to one of its ads
Which ordersPending, unpaid and canceled orders are in its reports; test orders are notOnly orders whose purchase event reached GA4Only orders its pixel or server event reported
The revenue it showsTotal sales: gross sales minus discounts and sales reversals, plus taxes, duties, shipping and feesPurchase revenue: the value the tag sent, minus refundsThe value its pixel or server event sent
Who gets the creditThe channel of the visit, under the model you pick: last non-direct click, last click, first click, any click or linear (Shopify)The touches before the purchase, split by GA4's model, which defaults to data-driven attributionIts own ad, when the purchase came within 7 days of a click or 1 day of a view under the default setting
When it books the saleThe day the order was placed; a return shows on the day it was processedThe day the event arrived; credit looks back up to 90 days by defaultThe day of the ad impression or of the purchase, by the report's setting

None of them is wrong by its own rules. They answer three different questions: was an order placed, which recorded source came before this purchase, and can this purchase be linked to my ad?

Why the numbers never match: five causes, one check each

1. VAT and shipping are in one number and not in another

Shopify defines Total sales as "gross sales - discounts - sales reversals + taxes + duties + shipping charges + fees" (Shopify). Google's rule for the purchase value is the items alone: "Don't include shipping or tax". Shopify's Google & YouTube app sends a purchase value that "does not include shipping charges and taxes". So on the same order, GA4 shows less than Shopify's Total sales before a single sale is lost. Meta shows whatever value its pixel or server event sends, so check what yours sends.

Check. Pick one recent order with no discount and no refund. Write down the items excluding VAT, the VAT, the shipping and the total paid. Find the same order in GA4 by its transaction ID and read Purchase revenue. Pass: you know which of those amounts GA4 holds, and you set it against Shopify's Net sales ("gross sales - discounts - sales reversals"), not against Total sales. The GA4 VAT check walks you through it.

2. One purchase is sent twice

Shopify warns that a store with more than one Meta pixel "can result in duplicate or incorrect data in your reports". Meta removes a duplicate only when the pixel's event ID matches the Conversions API's event_id and the event names match, and events are "only deduplicated if they are received within 48 hours". Then it discards the later copy. In GA4, the transaction_id parameter "helps you avoid getting duplicate events for a purchase".

Check. Count GA4's purchases and Shopify's paid orders for the same closed week. Pass: GA4 is at or below Shopify. Fail: GA4 is above it, so some orders send the purchase event more than once, often from a second tag in the theme next to an app. For Meta, ask whoever set up the Conversions API whether the server event carries the same event_id as the pixel. If it does not, Meta has nothing to match the two copies on.

3. Meta counts views, and some conversions are estimates

Meta's Marketing API documentation says its default attribution setting means 7-day click and 1-day view (Meta). So a purchase made within a day of an ad someone only saw can count for Meta. Shopify credits the channel a visit came from (Shopify), so an ad that was seen and never clicked leaves it nothing to credit.

Some conversions are not observed at all. When a visitor declines analytics cookies under consent mode, GA4 receives a cookieless ping, and that data "is used for behavioral and conversion modeling, to fill the gaps in your data". In Google Ads, "In the 'Conversions' column, Google reports both modeled and observed conversions" (Google Ads). Modelled conversions estimate purchases Google is unable to observe directly, so they will never match an order list one by one.

Check. Report Meta purchases for one closed week twice: once with 7-day click alone, once with 7-day click and 1-day view, Meta's default. Through the API this is the action_attribution_windows setting. The difference is what Meta credits to people who saw an ad and did not click. Keep that line apart in every comparison with Shopify.

4. Each one counts on its own clock

Meta's documentation gives the example of a person who saw an ad on 1 January and bought on 2 January: the purchase shows on 1 January when the report counts by impression time, and on 2 January when it counts by conversion time (Meta). GA4 gives credit to touches up to 90 days before a purchase by default (Google). Shopify shows a return "as a negative number on the date the return was processed". And Shopify notes that "Discrepancies might be introduced because of different reporting time zones".

Check. Pick a closed week, Monday to Sunday, that ended more than seven days ago, so late purchases and returns have landed. Set every report to those seven days in your shop's time zone, and Meta to count by conversion time. Pass: the gap that is left stays the same when you run the same week again a week later.

5. Some buyers are never recorded

Shopify's own page on discrepancies says "Some visitors might not allow cookies or JavaScript" and that customers "can use browser extensions to block Google Analytics from tracking their sessions and purchases". Under consent mode, when a visitor declines analytics storage, "No Analytics cookies are set, accessed, or read from the device". Shopify still records the order. GA4 gets a cookieless ping, or nothing if its tag was blocked.

Check. Divide GA4's purchases by Shopify's paid orders for the same week. The platform over-claim tracker does it week by week. Far below 1: consent, a blocked tag or a purchase event that did not fire. Above 1: go back to cause 2.

Then add the platforms up

Each ad platform reports on its own ads. Nothing in Meta's count removes an order that Google Ads also counts, and the other way round. Add Meta's purchases to Google Ads' and TikTok's for one week and set the total against Shopify's paid orders. Above 1 means some orders are claimed more than once. The ad platform over-reporting checker adds up what Meta, Google and TikTok report and sets the sum against your store's revenue.

Which number goes to your accountant?

Shopify's, filtered to paid orders. It is the only one of the three built from orders, and your books start from orders and the money that reached your bank. Two things to fix before you hand it over. Shopify's reports include "Pending, unpaid, and canceled orders", so filter to paid orders. And Total sales include taxes and shipping, while Net sales are gross sales minus discounts and sales reversals. Give your accountant the order list and let them pick the measure.

Never hand over an ad platform's purchase value. It is a claim on your sales, counted inside that platform's window, and the claims of different platforms overlap.

Which number goes into a budget decision?

None of the three as reported. Shopify says what you sold, not which channel sold it. GA4 assigns credit by a rule: attribution, in Google's words, is "the act of assigning credit for important user actions to different ads, clicks, and factors", and its models give direct visits no credit "unless the path to key event consists entirely of direct visits". Meta's count is credit as well, set by Meta's own window.

Credit is not cause. Gordon, Zettelmeyer, Bhargava and Chapsky compared 15 advertising experiments at Facebook with observational methods (Marketing Science, 2019) and found that the observational methods "often fail to produce the same effects as the randomized experiments". The number a budget needs is what each channel caused: the sales that would not have happened without it. A holdout test measures that for one channel at a time (incrementality testing). Causal attribution estimates it for every channel at once, from the paths GA4 already holds.

Get the file that holds the paths

Your GA4 property already holds the paths behind all three numbers, in one export:

  1. In GA4, open Advertising, then Attribution paths.
  2. Set the date range to 12 months or more.
  3. Pick the purchase key event.
  4. Click Share this report, then Download file, then CSV.

Which defaults change what the file holds, and how to check them: how to export attribution paths from GA4.

See what each channel caused. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes. The read is €99 once per upload, excl. VAT, with a full refund within 30 days. The first finding is free: which channels sent the buyers GA4 files under Direct. Get your first finding free

Sources, 17 November 2026: Sales reports (Shopify); Marketing reports (Shopify); Analytics discrepancies (Shopify); Meta pixel (Shopify); Recommended events: purchase (Google for Developers); Shopify event parameters (Google for Developers); Analytics dimensions and metrics (Google); Get started with attribution (Google); Select attribution settings (Google); Consent mode on websites and mobile apps (Google); Understand your conversion tracking data (Google Ads); About modeled online conversions (Google Ads); Ad Account, Insights (Meta for Developers); Handling duplicate Pixel and Conversions API events (Meta for Developers); A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook (Gordon, Zettelmeyer, Bhargava and Chapsky, Marketing Science, 2019)

Frequently asked questions

  • Why doesn't GA4 revenue match Shopify?
    Because they count different amounts and different orders. Google's rule for the purchase value is the items alone, without shipping or tax, while Shopify's Total sales add taxes, duties, shipping and fees. GA4 also misses buyers who decline analytics cookies or block its tag. Set GA4's Purchase revenue against Shopify's Net sales for the same closed week, not against Total sales.
  • Why does Meta report more purchases than Shopify?
    Meta counts the purchases it can link to its own ads, by default within 7 days after a click or 1 day after a view. That includes buyers who only saw an ad, and nothing in its count removes an order another platform also counts. Meta drops a browser and a server copy of one purchase as a duplicate only when their event names and event IDs match.
  • Which number should I use for my marketing budget?
    None of the three as reported. Shopify says what you sold, and GA4 and Meta say who gets the credit under their own rules. A budget needs what each channel caused: the sales that would not have happened without it. A holdout test measures that for one channel; causal attribution estimates it for every channel from your GA4 paths.
  • Which number goes to my accountant?
    Your shop's orders, filtered to paid ones. Shopify's reports include pending, unpaid and canceled orders, and its Total sales include taxes and shipping, so hand over the order list and let your accountant pick the measure. An ad platform's purchase value is a claim on your sales, not a sale.
  • Are modelled conversions real purchases?
    They are estimates. Google says modelled conversions estimate conversions it is unable to observe directly, and its Conversions column reports modelled and observed conversions together. GA4 models from cookieless pings when a visitor declines analytics cookies. Modelled conversions will never match your order list one by one.

Go deeper: Incrementality testing, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.

Keep reading

Terms in this article

Browse the full glossary

Your platforms guess.
We run the math.

Upload a GA4 export and see what each channel caused, next to last-click, in 1–2 minutes. The read is yours to keep.

The free finding runs in your browser. Your file never leaves it. The full read is €99, refundable within 30 days, no questions asked. Prices exclude VAT.