How-To · Tracking Setup
Is Your Tracking Lying to You? A 10-Minute GA4 Sanity Check
Duplicate tags, self-referrals, missing key events and your own visits inflating the numbers. Here's how to spot the usual GA4 problems and fix them.
Check five things. Make sure the GA4 tag fires once per page, your important actions are marked as key events, your own team's visits are filtered out, payment and booking tools aren't showing up as referrals, and a test visit appears correctly in DebugView. Most misleading GA4 numbers trace back to one of these.
Key Takeaways
- A bounce rate in the single digits usually means the tag fires twice.
- Mark lead and sale actions as key events, or GA4 won't count them as wins.
- Filter out internal traffic and list unwanted referrals like payment pages.
- Test with a real visit in DebugView before trusting any change.
Bad data is worse than no data. With no data, you know you're guessing. With bad data, you make confident decisions based on numbers that were never true.
The good news: most GA4 problems fall into a handful of patterns, and you can check for all of them in about ten minutes.
1. Is the Tag Firing Once?
The most common GA4 problem is the tag installed twice: once in the site's theme and again through Google Tag Manager or a plugin. Every page view gets counted double, and engagement numbers look suspiciously amazing.
How to check: Open your site with Google's Tag Assistant connected, load a page and look at how many times the page view fires. Once is right. Twice means you have two installs to untangle.
2. Are Your Wins Marked as Key Events?
GA4 collects lots of events, but it only treats the ones you mark as key events as wins. Google describes a key event as an action that's especially important to your business. If form submissions or bookings aren't marked, your reports will show visits but no results.
How to check: In Admin, open Events or Key events and confirm your lead and sale actions are on the list. Then compare last month's count with the leads you actually received. They won't match perfectly, but they should be in the same neighborhood.
3. Are You Counting Yourself?
Your team visits the site constantly: checking pages, testing forms, showing it off. That inflates traffic and muddies tests.
How to fix: GA4 lets you define internal traffic by IP address, then turn on a data filter that excludes it. Start the filter in Testing mode so you can confirm it catches the right visits before you activate it. Once a filter is active, its effect on the data is permanent, so double-check first.
4. Are You Referring Yourself?
If your own domain, a payment provider or a booking tool shows up under Referral, sessions are probably being broken in two when visitors hop over and back. The sale then gets credited to "paypal.com" instead of the ad or search that really brought the person in.
How to fix: Add those domains to GA4's list of unwanted referrals in your data stream's tag settings.
5. Did a Real Test Visit Show Up Right?
Before you trust any change, watch it happen.
- Turn on debug modeUse Tag Assistant or the GA debugger so your visit shows up in DebugView.
- Walk through a real pathLand on a page, click around and submit a test form.
- Watch the events arriveIn Admin, open DebugView and confirm each event appears once, with the right details.
- Annotate the fixAdd a GA4 annotation on the date so future you knows when the numbers changed.
The Sanity Check, as a List
- One GA4 tag per page, not two
- Lead and sale actions marked as key events
- Internal traffic filter tested, then active
- Payment and booking domains listed as unwanted referrals
- UTM tags on links in emails and social posts
- A test visit confirmed in DebugView
Frequently Asked Questions
My bounce rate is 2%. Is that good?
Probably not. A bounce rate that low usually means something is firing extra events or the tag is installed twice. Real sites rarely look that perfect.
Why does GA4 show fewer leads than my inbox?
Some visitors block analytics or decline cookies, and some leads arrive by phone or email without a tracked event. A small gap is normal. A big gap means a key event isn't firing.
Will fixing tracking change my old numbers?
No. Fixes apply to data collected from then on, which is exactly why an annotation on the fix date helps.
Sources
- Events and key events, Google Analytics Help
- Filter out internal traffic, Google Analytics Help
- Monitor events in DebugView, Google Analytics Help
- About annotations, Google Analytics Help