There is nothing more frustrating than logging into your Shopify or CRM dashboard, seeing 100 sales, and then looking at Google Ads only to find it claims you got 40.
If your GA4 conversion count disagrees with your actual revenue or your gut instinct, you aren't alone. After auditing dozens of ad accounts, I've noticed the exact same tracking problems showing up again and again.
This checklist covers the most common, and most damaging, issues. Work through it before you trust any Smart Bidding decisions to your data.
1. Check for Duplicate Conversion Events
Open GA4, navigate to Reports → Events, and filter by your main conversion event (like purchase). Now, cross-check that count against your backend platform's order count for the exact same time period.
The usual suspect for overcounting: Multiple Google Tag Manager (GTM) triggers firing on the same page. Check your GTM container for duplicate triggers. A simple Page View trigger and a DOM Ready trigger can easily both fire on the same thank-you page, doubling your numbers.
The Fix: Use GTM's Preview mode and walk yourself through a test conversion. Count how many times the event fires in the debugger. It should be exactly once.
2. Verify the Thank-You Page URL
Your conversion trigger probably relies on a URL match, something like /thank-you or /order-confirmation. But if your platform updates its URL structure (for instance, Shopify deciding to append a ?ref=... query parameter), your hardcoded trigger can break silently overnight.
The Fix: Use contains matching for the path rather than equals for the full URL. Make it a habit to test your checkout flow after every major platform update.
3. Confirm Enhanced Ecommerce Data
If you're running Shopify or WooCommerce, your purchase event needs to be passing specific data back to Google: transaction_id, value, currency, and an array of items.
If any of these are missing, Smart Bidding literally cannot optimize for revenue. It will fly blind and optimize only for conversion volume, meaning you might get lots of cheap sales instead of profitable ones. You can read more about setting up ecommerce events in the official Google documentation.
The Fix: In GTM Preview mode, expand the purchase event and verify that all parameters are populating with real, dynamic values.
4. Check Your Session Attribution
GA4 uses a completely different attribution model than the old Universal Analytics. By default, it relies on data-driven attribution. If you recently migrated, your channel-level numbers will look different. That's expected behavior, not a bug.
The real red flag: If your Direct traffic proportion is suddenly higher than 20%, it's usually a glaring sign that you're missing UTM parameters on your ad campaigns or email blasts.
5. Verify the Google Ads Link
In GA4, head to Admin → Product Links → Google Ads Links. Confirm your Google Ads account is actually linked and, crucially, that Personalized Advertising is enabled.
Without this link, Google Ads cannot import your hard-earned conversions from GA4. Your Google Ads conversion column will sit completely empty or rely solely on its own legacy tag-based conversions.
6. Check Import Status in Google Ads
Jump over to Google Ads. Go to Tools → Conversions, and find your GA4-imported conversions. The status should proudly say Recording.
If it shows No recent conversions, the link might have broken, or the event isn't firing. A surprisingly common cause is that the GA4 property was recreated or switched out by a well-meaning developer, quietly severing the link.
7. Hunt Down Self-Referrals
If your checkout flow redirects users through a payment processor (like Stripe, PayPal, or Klarna), the session can restart when they land back on your return URL. GA4 sees this as a new session, attributing your hard-won conversion to (direct) or paypal.com instead of the actual traffic source that drove the sale.
The Fix: Add your payment processor's domain to GA4's referral exclusion list (Admin → Data Streams → Configure Tag Settings → List Unwanted Referrals).
8. Ensure Server-Side Events Aren't Doubling Up
If you've upgraded to a sophisticated setup with both client-side GTM tags and server-side GTM containers sending the exact same event, you're going to double count everything.
The Fix: Inside your server-side GTM, use the send_to parameter so server events route exclusively to GA4, and client events route exclusively to Google Ads. Alternatively, implement robust deduplication logic in your server container.
9. The Incognito Ad Blocker Test
Walk through a real conversion in an incognito window with a popular ad blocker like uBlock Origin enabled. If the conversion fails to show up in GA4's Real-Time report, you have a client-side blocking problem.
If your audience uses ad blockers heavily (like B2B tech audiences), your conversion data is systematically under-counting reality.
The Fix: The only bulletproof solution is implementing robust server-side tracking for your most critical conversion events.
10. Compare GA4 vs. Platform Data for 30 Days
Pull the last 30 days of conversions from GA4 and hold it up against your ultimate source of truth (your Shopify orders, closed CRM deals, or phone system logs).
What is acceptable? A discrepancy of ±3–5% is normal. Browsers crash, ad blockers interfere, and attribution windows differ.
What requires an investigation? A discrepancy of 10% or more in either direction means something in the pipeline is fundamentally broken.
What's Next?
If you spotted issues on more than three of these checks, your Smart Bidding algorithm is optimizing based on unreliable data. This means your reported ROAS is almost certainly an illusion. Compare bidding strategies in our guide on Target CPA vs. Target ROAS.
Explore our full-service GA4 & GTM Conversion Tracking Setup and Google Ads Management services, or Book a free call with us to get a complete tracking audit.