How to tell if your GA4 is actually tracking correctly

If you can't answer that quickly, or the answer is "they never really have", you're not alone. Most businesses assume GA4 works because it shows numbers. The dashboard loads, the line goes up and to the right, and everyone moves on. But numbers on a screen are not the same as truth. GA4 will happily report figures that are wrong, and it will do it with total confidence.

I see this constantly. A team makes a budget decision, a channel call, or a campaign cut based on GA4 data that was never reconciled against anything. The tool said 400 conversions, so 400 conversions it was. Nobody asked whether those 400 matched the 260 actual enquiries sitting in the CRM.

This piece is the practical companion to the "why your data is wrong after a migration" conversation. That one explains the causes. This one is the self-audit: a concrete google analytics audit you can run yourself to work out whether your setup is telling you the truth.

Why "it shows numbers" is a trap

GA4 is a collection engine, not a verification engine. It records whatever it's told to record, tags it however it's been configured, and reports it back without checking any of it against reality. If the configuration is wrong, the data is wrong, and the interface gives you no warning. There's no red light.

The problem compounds because the numbers are usually plausible. Nobody questions a figure that looks roughly right. A conversion count that's 40% too high doesn't announce itself. It just sits there quietly, shaping decisions, until someone finally lays it next to the CRM and the gap falls out.

That reconciliation gap is the single most common issue I run into. So let's start there.

Six signs your GA4 is misreporting

Run through these. If more than one lands, your data is steering decisions it shouldn't be.

• GA4 and your CRM disagree. Pull GA4 conversions for last month and put them next to actual enquiries or sales in your CRM for the same window. A small variance is normal. A gap of 30% or more, in either direction, means something structural is broken.
• Conversions are double-counted or missing entirely. A thank-you page that fires on every reload inflates your count. A key event that was never configured, or that broke when the form was rebuilt, means real conversions vanish. Both happen more often than teams realise, and they often happen at once.
• Internal and bot traffic is polluting the data. If your own team, your developers, and a steady drip of bot hits are all counted as real users, your engagement and conversion rates are quietly diluted. Small sites feel this hardest.
• Cross-domain tracking is broken. If your booking system, checkout, or enquiry form lives on a different domain and cross-domain isn't set up correctly, GA4 treats the jump as a new session from a new user. Your best converting journeys get shredded into disconnected fragments, and paid channels get robbed of the credit.
• Key events were never properly configured. Plenty of accounts I open have two or three events doing real work and a dozen that were switched on, never tested, and never trusted. If you can't say what each key event actually measures, it isn't measuring anything useful.
• Consent settings are silently wiping data. A consent banner that blocks GA4 until acceptance, combined with a chunk of visitors who never click, means a slice of your traffic is never recorded at all. The data looks complete. It isn't.

Notice the pattern. None of these throw an error. Every one of them produces numbers that look fine and are wrong.

What the reconciliation gap looks like in practice

A concrete one. A services business came to me convinced their organic search was carrying the whole operation. GA4 showed organic driving the clear majority of their leads, so that's where attention and budget were going. Paid was on the chopping block.

We put GA4 next to the CRM for the same three months. GA4 reported a bit over 500 form conversions. The CRM held roughly 300 genuine enquiries. Right away, a 200-lead gap nobody could explain.

Pulling it apart took an afternoon. The thank-you page fired its event on reload, so anyone who refreshed or hit back was counted twice. The enquiry form sat on a separate booking subdomain with cross-domain tracking half-configured, so a large share of conversions were being credited to "direct" and to organic landing pages rather than the paid campaigns that actually started the journey. And internal traffic was never filtered, so the team's own testing padded the numbers further.

The correction rewrote the story. Paid wasn't the underperformer, it was a meaningful share of real pipeline that GA4 had been quietly handing to organic. They kept the budget they were about to cut. That's the cost of unaudited data: not a messy dashboard, but a wrong decision made with full confidence.

How a proper analytics audit works

A real google analytics audit isn't clicking through the interface looking for something that seems off. It's a structured pass that ends with data you can actually trust. The shape of it:

• Reconcile against the source of truth first. Before touching configuration, line GA4 up against the CRM or sales system for a fixed window. The size and direction of the gap tells you what kind of problem you're chasing.
• Audit every key event end to end. Trigger each one deliberately and confirm it fires once, at the right moment, with the right parameters. Anything that can't be explained gets fixed or removed. No orphan events.
• Check cross-domain and referral exclusions. Map every domain in the real user journey, including payment and booking systems, and confirm sessions hold together across the jumps. This is where attribution most often quietly breaks.
• Filter internal, developer, and bot traffic. Exclude your own people and known bots so engagement and conversion rates reflect actual customers, not the team testing forms.
• Verify consent and tag firing. Confirm what happens before and after consent, and quantify how much data the current setup is dropping so you know what you're not seeing.
• Document the corrected baseline. The audit ends with a known-good configuration and a reconciled baseline, so the next time the numbers move, you can tell a real change from a tracking fault.

Work through that and you move from "the dashboard shows numbers" to "I can defend these numbers in a budget meeting". That's the whole point. Analytics only earns its place when you can act on it without a caveat.

If your GA4 has never been reconciled against your CRM, that's the audit worth booking first, and it's exactly what the Analytics and Tracking Audit is built to sort out.