When we take over a new account, we run into at least one of these three mistakes almost every time — often within the first few minutes of looking. These aren't exotic edge cases. They're things that happen during normal setup when nobody's flagged what to watch out for.
Mistake 1 — A purchase is tracked as a regular pageview, not a key event
The most common thing we see. The site has GA4 installed, tracking works, but the "thank you for your order" page behaves like any other page in the system — never marked as a key event (conversion).
The consequence: GA4 knows people bought something, but never reports it to Google Ads as a conversion. Campaigns end up optimizing toward "clicks" instead of "purchases" — which in practice means the algorithm is targeting people who click, not people who buy.
How to spot it: Admin → Key Events. If your "purchase" or "order submitted" event isn't listed there, this is your problem.
We cover the correct setup process in detail in our GA4 setup guide.
Mistake 2 — An internal filter is excluding real customers
GA4 lets you exclude a company's internal traffic so employees don't skew the numbers. The problem shows up when the filter is set by IP range for an entire office building or coworking space — where other companies also sit, or where employees work from home through a shared VPN.
The consequence: The data looks smaller than reality, and nobody knows why, because the filter runs quietly in the background.
How to spot it: Admin → Data Filters. Check the IP range against who's actually at that address. When in doubt, set the filter to "Testing" mode first — it logs the impact without removing data, so you can see the effect before losing anything irreversibly.
Mistake 3 — Test and production properties get mixed up
Businesses often create a test GA4 property to experiment before going live. The problem happens when the live site's tracking code stays pointed at that test property by mistake — or the reverse, where reporting happens off the test property while the real, production property collects actual data nobody's watching.
The consequence: Someone is looking at numbers that don't represent the site's real traffic — budget or strategy decisions get made on bad data without anyone realizing it.
How to spot it: Check the Measurement ID in the site's code (View Source → search for G-) against the ID of the property you're actually viewing in GA4.
Why these three keep happening
All three share one thing in common: GA4 doesn't throw an error when they happen. The system keeps running, data keeps collecting, dashboards keep filling with numbers — the numbers just don't mean what you think they do. That's exactly why they're so easy to miss, even when someone's checking the account regularly.
If you're dealing with the broader question — why conversions aren't importing into Google Ads at all even though GA4 itself looks fine — we cover seven common causes in detail here.
In MARTY, we check for these three mistakes — and a handful of others — automatically every time a new account gets connected, before they can skew a budget decision. Join the waitlist.