Let me guess: your organization finally migrated to GA4, checked the box on the "compliance mandate," and assumed everything was working correctly.
Spoiler alert, it's probably not.
I see this pattern constantly with large organizations, government agencies, and higher ed institutions. The GA4 implementation gets rushed through by a well-meaning IT team (or worse, a marketing coordinator who Googled "how to set up GA4"), and six months later, someone notices the data doesn't make sense. Sessions are missing. Conversions aren't tracking. Year-over-year reports are impossible.
Here's the thing about GA4: it's deceptively simple to install but incredibly complex to implement correctly. And when you're dealing with multiple departments, legacy systems, procurement bureaucracy, and stakeholders who think "cookies" are still just something you eat at meetings? The margin for error expands exponentially.
Let's walk through the seven most common GA4 implementation mistakes I encounter during audits, and more importantly, how to fix them before they cost you budget, credibility, or both.
1. You Set It to "Default" and Forgot About Data Retention
GA4's default data retention period is two months. Let that sink in for a second.
Your institution spent months navigating procurement to get analytics in place, your team finally started building custom reports, and GA4 is quietly deleting your granular event data after 60 days. Want to run a year-over-year enrollment comparison? Too bad. Need to analyze seasonal trends in grant applications? The data's gone.
The Fix: Navigate to Admin > Data Settings > Data Retention and change it to 14 months immediately. This setting is not retroactive, so if you've already been running on the default, you've lost that historical data permanently.
For government and higher ed clients subject to records retention policies, document this configuration change and the rationale. Your compliance team will thank you later.
2. You're Tracking the Same Events Multiple Times (And Don't Know It)
Duplicate event tracking is the silent killer of GA4 data integrity. It happens when:
- GA4 loads multiple times on the same page (common with tag management conflicts)
- You're using both Google Tag Manager and platform-specific tracking (looking at you, WordPress plugins)
- Multiple departments set up their own "custom" tracking without coordinating
I recently worked with a state university where the admissions team, IT department, and external agency had all implemented GA4 tracking independently. Their "form_submit" event was firing three times per actual submission. When they reported a 200% increase in applications, they were actually seeing the same growth as the previous year, just counted three times.
The Fix: Conduct a complete GA4 audit to identify all tracking sources. Use the DebugView in GA4 to watch events fire in real-time and catch duplicates. Establish a single source of truth for implementation, ideally Google Tag Manager with proper governance documentation.
For large organizations, create a tracking specification document that requires approval before any new events are added. Yes, it slows things down. It also prevents catastrophic data pollution.
3. You're Creating Custom Events in the GA4 Interface (When You Shouldn't Be)
This one trips up even experienced analysts. GA4's interface has a friendly "Create Event" button that seems helpful, until you realize you're creating duplicate events on top of what's already being sent from GTM or your tracking code.
Here's the distinction that matters: custom events appear automatically in GA4 once they're sent from your tracking implementation. You don't need to "register" them in the interface. The "Create Event" function is only for modifying or combining existing events, not for initial setup.
The Fix: Use the GA4 interface only to register custom parameters as custom dimensions or metrics. Your actual event implementation should happen in Google Tag Manager (or your tracking code), and those events will populate in your reports automatically.
Think of it this way: GTM is the kitchen where you cook the data; GA4 is the restaurant where you serve it. You don't cook the meal twice.
4. You Burned Through Your Custom Dimension Limit Without a Plan
GA4 gives you 50 custom dimensions per property. Sounds like plenty, right?
Then your enrollment team wants to track 12 different application stages. Your development office needs custom dimensions for donor segments. Your communications team wants to track content categories. Your IT team adds technical parameters "just in case."
Before you know it, you're at 47 custom dimensions, most of which aren't actually being used in any reports, and you can't add the one critical dimension you actually need for the president's dashboard.
The Fix: Treat your custom dimensions like a limited natural resource, because they are. Before implementation, create a dimension strategy document that identifies:
- Must-have dimensions tied to specific reporting needs
- Shared parameters that work across multiple use cases
- A formal request and approval process for new dimensions
For large organizations, consider implementing a quarterly review where unused dimensions get deprecated to free up space. Document why each dimension exists and who owns it. This prevents the "I think someone was using that" paralysis when you need to make changes.
5. Your Google Tag Manager Setup Doesn't Account for Your Complex Website
Government and higher ed websites are rarely simple. You've got:
- CMS platforms integrated with CRM systems
- Authenticated user portals alongside public pages
- Legacy systems duct-taped to modern frameworks
- Dynamic content loading without proper page view triggers
- Custom JavaScript that conflicts with tracking code
When your GTM container doesn't account for this complexity, tags fire at the wrong time (or don't fire at all). I've seen cases where GA4 was only tracking 60% of actual traffic because tags weren't firing on dynamically loaded content.
The Fix: Test your GTM implementation across every section of your website, especially:
- Password-protected portals (student/employee systems)
- Form-heavy pages (applications, donations, contact forms)
- Single-page applications or JavaScript-heavy interfaces
- Pages with third-party integrations (payment processors, chat widgets)
Use GTM's Preview mode religiously, and implement JavaScript error tracking to catch initialization failures. For complex enterprise sites, consider bringing in a GTM specialist for the initial architecture, it's cheaper than fixing broken data six months later.
6. You're Comparing GA4 Metrics to Universal Analytics (Or Your CRM) and Losing Your Mind
"Why don't our session numbers match?"
"Where did our bounce rate go?"
"Why is GA4 showing different conversion numbers than our CRM?"
These questions haunt every GA4 implementation in large organizations. The answer? GA4 fundamentally changed how it defines and calculates metrics. Sessions work differently. Bounce rate became "engagement rate." Attribution models shifted.
Then you add CRM data to the mix, which typically shows unfiltered data while GA4 applies filters (internal IP addresses, bot traffic, privacy settings), and the discrepancy reports start flying up to leadership.
The Fix: Stop trying to make GA4 match Universal Analytics. It won't, and it shouldn't. Instead:
- Create documentation explaining the metric definition changes
- Build baseline reports in GA4 before comparing to historical data
- For CRM discrepancies, document known filter differences (like internal traffic exclusions)
- Educate stakeholders that "different" doesn't mean "broken"
The bigger challenge in government and higher ed? Convincing leadership that the new numbers are correct, not the old ones. Build a narrative around why GA4's methodology is actually more accurate for modern user behavior.
7. You're Mixing Scopes in Your Reports (And Getting Nonsense Results)
GA4 operates with three scope levels: user, session, and event. Each dimension and metric belongs to one of these scopes, and when you mix them incorrectly in custom reports, you get results that look impressive but are fundamentally meaningless.
For example, combining user-scoped dimensions (like "First User Source") with event-scoped metrics (like specific button clicks) creates a report that looks like it's telling you where engaged users come from, but it's actually overcounting because it's aggregating across incompatible scopes.
The Fix: Before building any custom report, verify the scope of every dimension and metric you're using. GA4's interface tries to warn you about incompatible combinations, but it's not foolproof.
For enterprise reporting, create a dimension and metric reference guide that lists the scope of commonly used elements. Train your team on scope concepts before they start building reports. It's tedious upfront, but it prevents the "why is this number impossibly high" conversations later.
What This Actually Means for Your Organization
If you're reading this and recognizing three or more of these mistakes in your current GA4 setup, you're not alone. Almost every large organization I audit has at least five of these issues, and most don't realize it until they're making decisions based on flawed data.
The good news? These are all fixable. The bad news? They require strategic attention, not just technical tweaks. Someone in your organization needs to own GA4 governance, and that someone needs authority to say "no" to ad-hoc tracking requests that would compromise data quality.
For government agencies and higher ed institutions, the stakes are particularly high. Your data informs budget allocations, enrollment strategies, and policy decisions. Getting it wrong doesn't just waste marketing dollars: it impacts your mission.
Start here:
- Run a comprehensive GA4 audit to identify which of these seven mistakes exist in your implementation
- Prioritize fixes based on which reports and decisions are most critical to your organization
- Establish governance documentation before making changes
- Test everything in a development environment first
And if you're staring at this list thinking "we don't have the internal capacity to fix this": that's exactly why specialized GA4 consulting exists. Sometimes the ROI Shield requires calling in reinforcements.
Your data is too important to leave broken. Fix it now, before your next board presentation relies on numbers that don't add up.
