Before Black Friday, check two numbers

what your ad platforms say you sold, and what Shopify says you sold. if they're far apart, something is counting wrong, and the ad platforms bid on that number

do the check run the free scanner

why now

Meta and Google Ads decide where to spend from the purchases your tags report. If each purchase is sent twice, they think a campaign sells twice what it does and spend on that basis. If purchases go missing, they pull back from what works. The check takes about ten minutes, and the time to do it is before budgets go up for Black Friday.

Written by Andrew Maury. I sell tracking audits and run a free scanner. The numbers below come from my own scans, and the method is on the study page.

jump to: what I found · the check · where the second copy is · timing

what I found in a random sample

A random sample of 386 Shopify storefronts in the Majestic Million, scanned October 13 to 14, 2026.

73 of the 307 stores running GA4 (24%) sent each page view twice to the same property

In at least 62 of them, one copy came from Shopify's own Google & YouTube channel and the other from a second install of the same tag. Page views are doubled, and every visit reads as engaged, so engagement rate looks too good and bounce rate too low. Sessions and visitors aren't doubled.

8 of the 310 stores whose checkout loaded (about 1 in 40) sent Meta's checkout event twice

InitiateCheckout went out from two installs with different event IDs, so Meta can't merge them. Both copies run inside checkout, so purchases are very likely doubled too.

My scanner doesn't place orders, so it can't see purchases directly. A doubled page view tells you the same tag is installed twice. The two numbers below tell you whether purchases are doubled as well.

the check

  1. Pick a window. Use the last 28 full days. Shorter windows let timezone differences move orders between days.
  2. Shopify. Go to Orders, filter to the Online Store sales channel and those dates, and write down the count. That filter leaves out POS, draft and marketplace orders, which never load a page and so never fire a tag. If you sell subscriptions, leave out the rebills too, for the same reason.
  3. Meta. In Events Manager, open your pixel and find the Purchase row on the Overview for the same dates. Write down the total events. Use this, not the purchases in Ads Manager, which are attributed and windowed.
  4. GA4. Go to Reports > Engagement > Events and find the purchase row for the same dates. Write down the event count. Don't use Monetization > Ecommerce purchases, which counts items rather than orders.
  5. Compare each with the Shopify number.
    • A little below orders is expected: ad blockers, declined cookie consent, people who close the tab right after paying. I won't give you a "normal" percentage, because nobody has published one with a method behind it.
    • Well above orders means something is sending purchases twice.
    • Far below orders means purchases are going missing, often because a tag doesn't run at checkout. If you're not on Plus and your purchase tracking lived in theme code or additional scripts, it stopped on 26 August 2026, when Shopify retired them on the thank-you page.

The GA4 vs Shopify orders page goes through the gap in more detail, including how to match orders one by one.

if it's high, where the second copy usually is

  • Shopify admin > Settings > Customer events. Every app pixel and custom pixel is listed there. Two that send to the same platform is the usual cause.
  • The sales channel apps. The Google & YouTube app and the Facebook & Instagram app each install their own copy. A tag in the theme code or in Google Tag Manager on top of one of them makes two.
  • Apps with their own tracking field. In one store I checked, the cookie-consent app had a GA4 ID filled in. It loaded GA4 on its own, next to Shopify's channel.

timing

A dead Universal Analytics tag is safe to remove any time. Removing one of two live installs a few weeks before Black Friday is a judgment call. If purchases are clearly doubled and you're spending on ads, raise it with whoever runs them now. Otherwise, write it down and fix it after Cyber Monday, when a mistake costs less.

what these numbers are and aren't

A random sample of Shopify storefronts in the Majestic Million, which leans toward larger, established stores. It is not a sample of all Shopify stores. The scan reads public pages only: it never signs in, never places an order, and can't see anything sent server-side. Method and limits are on the study page. Contains data from the Majestic Million (https://majestic.com), licensed under CC BY 3.0.

or let the scanner look

The free scanner loads your storefront, cart and checkout and lists every tag it sees firing twice. Enter your two numbers and it tells you which way the gap points. It reads public pages only, so it can't see server-side sends or the purchase itself. If you'd like me to go through it with you, that's the $250 Scan Read.

run the free scanner read a scan with me, $250