TRACKING SPECIALIST — BARIS ASA · LONDON
Synctag
Start the check → TRACKING SPECIALIST — BARIS ASA · LONDON

READING ROOM › SYMPTOM FILES › SYM-096

SYMPTOM FILE SYM-096 FILED 18 NOV 2026 · REVISED 18 NOV 2026

GA4 purchases counted twice on Shopify: the Google & YouTube app and a GTM tag both send them.

ALSO FILED AS
  • “Shopify GA4 purchases doubled after installing the Google & YouTube app”
  • “Google & YouTube app and GTM both send the GA4 purchase”
  • “Shopify GA4 duplicate purchase custom pixel and Google channel”
  • “GA4 double counting Shopify orders”
  • “shopify_event_name parameter GA4”
  • “Shopify GA4 purchases doubled after adding GTM”

Baris Asa, tracking consultant, London · Last checked: November 2026

If your Shopify store has the Google & YouTube app connected to a GA4 property, and GTM also sends a purchase to that same property, each order can be counted twice. The app sends one copy from Shopify’s checkout; a GA4 tag in GTM, usually in a custom pixel, sends the other. GA4 only drops the repeat when both carry the same transaction_id. To prove the purchase is counted twice by this pair, look at one order’s purchase hits: the app’s copy carries a shopify_event_name parameter, the GTM copy doesn’t. The fix is one sender: for most stores, keep the app and remove GA4 from GTM.

On file as F-101 . Shopify says it plainly: adding Google Analytics to GTM in addition to its built-in integration can result in duplicate tracking. This note is that one pair of senders, when the purchase itself doubles. For every app and pixel a Shopify store can run, and what each one sends, see Shopify tracking apps: what each sends. If you haven’t yet found where your second purchase comes from, start with the general hunt in GA4 purchase event fires twice.

How do you know both are installed?

First the symptom. In GA4, open the Transactions report for the last full week. Purchases close to double your Shopify orders, or a Transaction ID column that mixes two ID formats, point here. Then three places to check; step 3 can go to whoever holds the GTM login.

  1. The app. In Shopify admin, open the Google & YouTube app and go to its Settings. If a Google Analytics property is connected, note its measurement ID (it starts with G-).
  2. Customer events. In Settings, Customer events, look at the custom pixels. A pixel that loads GTM (search its code for GTM-) is the usual second sender. Search for G- and gtag too: Shopify turned old Google Analytics tags that weren’t moved to the app by February 2025 into custom pixels, and one of those doubles the purchase the same way. With the upgraded Thank you page, theme code no longer runs after checkout (what replaced the old checkout scripts), so a GTM purchase on a current Shopify store almost always comes from here.
  3. The container. In GTM, find the GA4 event tag that sends purchase and the Google tag it uses. If the measurement ID matches the one in step 1, both routes write to the same property.

Same ID in steps 1 and 3 means two senders for every order. Different IDs means two properties, each counted once. That isn’t this fault; start with the general hunt.

How do you tell the two copies apart?

Google lists shopify_event_name among the parameters the app sends on every event, purchase included, and suggests it for filtering the app’s data in GA4. Google doesn’t document the value it carries, so read it as present or absent. A GTM tag doesn’t send that parameter unless someone added it. That one field splits the two copies.

Same day. Place a test order with the browser’s Network panel open and filter for collect. Each GA4 hit with en=purchase is one copy. Search each one for shopify_event_name: the hit that carries it is the app, the hit without it is GTM. Two purchase hits for one order, one of each, confirms the pair.

Over a few days, because a browser test can miss a copy (see the cautions below):

  1. In GA4, Admin, Custom definitions, create an event-scoped custom dimension for the parameter shopify_event_name. Google says reporting on a new custom dimension starts after 24 to 48 hours.
  2. Then build a free-form exploration: Transaction ID and shopify_event_name as rows, Ecommerce purchases as the value, for a few days after the dimension went live.
  3. Read the rows. A purchase with any value in shopify_event_name came from the app. A purchase with (not set) came from somewhere else.
Transaction IDshopify_event_nameEcommerce purchasesSender
ID Apresent1Google & YouTube app
ID B(not set)1GTM in a custom pixel
EXPLORATION · purchases by sender, one order · worked example, illustrative, not a client account

One order, two IDs, two purchases kept. Which Shopify identifier each sender uses depends on the setup; what matters is that they differ, so GA4 has nothing to match. If both senders use the same ID, GA4 already drops one purchase and only the funnel events double, which is the overlap described in Shopify tracking apps: what each sends. A purchase double with matching IDs is somewhere else: go back to the general hunt.

Two cautions on testing. Google lists Tag Assistant among the features that may not work in a custom pixel, and Shopify’s GTM pixel tutorial says Tag Assistant’s Troubleshoot tag feature isn’t compatible with custom pixels, so GTM Preview is a poor witness for the GTM copy.

Trade press reported in 2026 that the app’s purchase can also reach GA4 from Shopify’s servers; the Google Analytics help pages I read don’t confirm it. Either way the exploration reads what GA4 kept, whichever route a copy took, so I trust it over a live test.

Which one do you switch off?

Switch off a whole route, not just the purchase. The app sends page_view, view_item, add_to_cart, begin_checkout and more. A GA4 setup in GTM next to it doubles those too, which inflates funnel steps even when nobody notices the revenue.

Keep the app (most stores). Google calls it the recommended method for Google tagging on Shopify. It also says running Google tags inside a Shopify custom pixel is not a supported implementation, and that its support can’t help with one. In GTM, pause or delete every GA4 tag that writes to the connected property, then publish. Leave the custom pixel in place if it still runs tags for other platforms. GTM in a Shopify custom pixel covers what that sandbox can and can’t see.

Keep GTM (only when you rely on it). If your reports depend on events or parameters the app doesn’t send, keep GTM and disconnect Google Analytics in the Google & YouTube app’s settings. Google says that integration can be disconnected there. Then the GTM route has to carry every ecommerce event on its own, so check the funnel the next day.

Don’t keep both and hope. Matching IDs on both routes would let GA4 drop the repeat. But Google doesn’t say which Shopify field the app sends as transaction_id, so you would be copying a value you can’t see documented. One sender is the fix.

Whichever you keep, check Google Ads in the same sitting. If Ads counts the sale through the app’s conversion action and through a GA4 import, both set as primary, the double moves to Ads. That is F-104 : GA4 and Google Ads numbers don’t match.

Picking the route that stays, then checking a full week of GA4 purchases against Shopify orders, is what I do as a Shopify tracking specialist for stores and the agencies behind them.

How do I know it’s fixed?

  • A closed week of GA4 purchases sits at or a little below the orders in Shopify, test orders left out of both. A few per cent below is normal: some buyers block scripts, and purchases without analytics consent are only partly modelled back. Above is the fault coming back. This is the check that counts.
  • In the Transactions report, new orders carry one ID format.
  • The exploration shows one row per new order: a shopify_event_name value if you kept the app, (not set) if you kept GTM.

Past data stays doubled. Note the switch-off date and compare from the first full week after it.

Common questions

Why does GA4 count my Shopify orders twice when I use the Google & YouTube app and GTM?

Both send a purchase to the same GA4 property for the same order. The app sends one from Shopify’s checkout_completed event; a GA4 purchase tag in GTM, usually inside a custom pixel, sends another. GA4 keeps both unless they carry the same transaction_id. Shopify’s own help warns that adding Google Analytics through GTM on top of the built-in integration can result in duplicate tracking.

How can I tell which GA4 purchase came from the Google & YouTube app?

The app adds a parameter called shopify_event_name to the events it sends, purchase included. Register it as an event-scoped custom dimension, then build an exploration with Transaction ID and shopify_event_name as rows and Ecommerce purchases as the value. Rows where shopify_event_name is (not set) came from somewhere else, usually GTM.

Should I keep the Google & YouTube app or GTM for GA4 on Shopify?

For most stores, the app. Google calls it the recommended method for Google tagging on Shopify and says running Google tags inside a Shopify custom pixel is not a supported implementation. Keep GTM for GA4 only if you rely on events or parameters the app doesn’t send, and then disconnect Google Analytics in the app’s settings so only one sender is left.

If I disconnect Google Analytics in the Google & YouTube app, do I lose Google Ads conversion tracking?

No. In the app, Google Analytics depends on conversion tracking, not the reverse: Google’s Merchant Center help says you deactivate Google Analytics before you can remove conversion tracking, and that the Google Analytics integration can be disconnected in the app’s settings. After disconnecting, open Google Ads, Goals, Conversions, and check the app’s purchase action is still recording and is the only primary purchase action, so the sale is counted once in Ads.

Will removing the duplicate fix my past GA4 revenue?

No. Purchases already recorded stay in the reports. Write down the date you switched one sender off and compare GA4 with Shopify from the first full week after it.

Sources

Read on 18 November 2026:

← back to the reading room