Measurement
Why is my GA4 data wrong?
GA4 data is usually not wrong, it is measuring something different than you assume. Common causes are misconfigured conversion events, consent mode gaps, cross domain tracking issues, bot traffic, or sampling in reports. Check the event configuration and consent setup before doubting the numbers themselves.
Last reviewed 2026-08-16
The detail
The longer answer
When someone says GA4 data is wrong, they almost always mean it disagrees with another number they trust more, like a CRM total, an ad platform's own reporting, or last year's Universal Analytics numbers. GA4 is rarely lying. It is measuring a different thing than the number you are comparing it to, and the gap is usually explainable once you know where to look.
The most common cause is event configuration. GA4 does not automatically know what a conversion is. If someone set up a purchase event or a lead event incorrectly, pointed it at the wrong page, or left a duplicate event firing twice, your numbers will be off in a very specific and traceable way. This is the single most frequent source of bad GA4 data we see, and it is entirely fixable with a proper audit of the event and conversion setup.
The second cause is consent and privacy tooling. If your site runs a cookie consent banner and it is wired to GA4 through consent mode, a meaningful share of visitors who decline analytics cookies will not be measured at all, or will be measured with modeled, estimated data rather than a raw hit. This is by design and required in many jurisdictions, but it means GA4 totals will structurally undercount actual traffic and conversions compared to server-side or CRM numbers that do not depend on a cookie choice.
Cross domain and cross subdomain tracking is another frequent culprit, especially for businesses that use a separate checkout platform, a booking tool, or a payment processor on a different domain. If cross domain linking is not configured, GA4 will see a visitor land on your marketing site, then leave and 'arrive' on the checkout domain as a brand new session with no referrer. This shows up as inflated direct traffic and broken conversion attribution, not as an obviously broken report.
Bot and spam traffic can also skew numbers, though GA4's default bot filtering catches the obvious cases. More subtle inflation comes from internal traffic (your own team browsing the site), staging environments sending data to the production property, or third party scripts that fire page view events on load without a real visitor present. None of these produce an error message. They just quietly add noise that makes your totals look larger or smaller than reality.
Finally, some of what looks like wrong data is a reporting artifact rather than a measurement problem. GA4's standard reports apply thresholding and, in some cases, sampling once your data volume or your use of certain dimensions crosses a threshold. Two people looking at the same underlying data through different report configurations can see different numbers, and neither is technically incorrect.
The fix in every case is the same: pick the specific discrepancy, trace it to its source using GA4's DebugView and the underlying event stream, and correct the configuration rather than distrusting the platform wholesale. Data that looks wrong for a reason you cannot name is a signal to audit the setup, not to abandon the tool.
Key points
What to take away
- GA4 measures whatever events and conversions were configured, not what you assume it measures by default.
- Consent mode causes real undercounting for visitors who decline analytics cookies, which is expected behavior, not a bug.
- Missing cross domain tracking inflates direct traffic and breaks attribution when checkout or booking lives on a separate domain.
- Internal, staging, and bot traffic can add invisible noise to totals without triggering any warning.
- Standard GA4 reports can apply thresholding or sampling, so two report views of the same data can legitimately differ.
- DebugView and the raw event stream, not the summary reports, are the right place to trace a specific discrepancy to its source.
Common misconception
What people get wrong
GA4 is just unreliable and undercounts everything compared to the old Universal Analytics.
GA4 uses a different data model and different default privacy behavior than Universal Analytics, so totals will differ even with identical traffic. Most large gaps trace back to a specific, fixable configuration issue like event setup or consent mode, not to GA4 being inherently less accurate.
Related questions
Questions that come up next
Why analytics data goes wrong, how to measure return, and what a rate really tells you.
Where this gets applied
The work behind this answer
Each link explains why it is relevant, not just where it goes.
How Lingows handles this
In practice
When a client tells us their GA4 numbers do not match what they expect, we do not start by rebuilding the property. We start by tracing one specific discrepancy through DebugView and the event stream to find the actual cause, whether that is an event firing twice, a missing cross domain configuration, or consent mode doing exactly what it is supposed to do.
From there we fix the configuration and document what the corrected numbers represent, so the client's team can explain the data to their own stakeholders instead of just trusting a dashboard.
Want this handled properly on your own site
Start with a diagnosis. You get the roadmap and the quote before anyone builds anything.