GA4 Audit Checklist: The Checks I Make Before Trusting the Data

Most GA4 audits start in the wrong place — someone opens the Realtime report, sees numbers moving, and concludes the setup is fine. It rarely is. One analysis of GA4 properties found that over 80% of accounts have incorrect or incomplete event tracking — the setup looks active because standard events like page_view are firing, while the events that actually matter for revenue decisions are missing, misfiring, or only half-implemented.

When I audit a GA4 property, I go through the same four checks in the same order every time — not because it's exhaustive, but because it’s where the highest-impact errors actually live.

Start with Key Events, not dashboards

Before looking at a single report, I check what’s actually marked as a Key Event and whether each one represents genuine purchase intent. It’s common to find Key Events polluted with page views, about-page visits, or generic sign-up events — none of which tell you anything about revenue, but all of which dilute every conversion report built on top of them. I’ve cleaned up GA4 setups where removing three or four irrelevant Key Events did more for report accuracy than any dashboard rebuild could have.

Custom events are usually where the discipline breaks down first. A click, form submission, or video-view event gets defined once, then quietly redefined or partially reused later — the event name stays the same while what it actually measures drifts. If nobody can tell you, from memory, exactly what a given Key Event represents and why it’s marked as one, that’s the first thing to fix.

Validate ecommerce events against real orders

Next, I compare GA4’s purchase events and revenue values directly against the store’s actual order data — not against another dashboard, against the source system itself. A variation of more than 5% between GA4 purchase data and your database or ERP almost always signals a tracking configuration error, not normal reporting variance. Duplicate purchase events, missing transaction IDs, incorrect ecommerce parameters, and cross-domain tracking gaps are the recurring causes.

On Shopify specifically, this check often surfaces the same root cause we wrote about in Why Broken Shopify Product Data Wastes Your Ad Spend: inconsistent product identifiers between the storefront and the order system mean GA4’s ecommerce events are technically firing correctly while still reporting against the wrong product record.

Why GA4, Google Ads, and Shopify disagree

By this point, a client has usually already noticed that GA4, Google Ads, and Shopify all report different numbers for the same week — and assumed one of them is simply broken. Properties with misconfigured conversion events report inflated or deflated ROAS by as much as 30% in connected Google Ads accounts, which makes this discrepancy expensive, not just confusing.

Some of the mismatch is structural and won’t fully disappear: each platform uses a different attribution window and handles cookie consent and ad blockers differently. The audit’s job is to separate that expected structural variance from an actual configuration error — and to agree, in writing, which system your team treats as the source of truth for which decision.

Is your attribution model actually right?

Once the event-level data is trustworthy, I check whether the attribution model in use actually fits the business — data-driven attribution needs enough conversion volume to learn from, and a short-consideration impulse purchase has a very different attribution shape than a considered B2B sale with a multi-week sales cycle. A model mismatched to the sales cycle doesn’t throw an error; it just quietly overcredits one channel at another’s expense, indefinitely.

How often to actually run this

A full audit roughly every six months, with lighter monthly checks on conversions, filters, and integrations, is a reasonable cadence for most ecommerce accounts — tightening up around major events like a platform migration, a new marketing channel launch, or a Shopify theme change, all of which are common triggers for silent tracking breakage.

If you’re not sure where your own GA4 setup currently stands, this is exactly the audit I run as part of my GA4 consulting work— working directly with you, not handed off to an account team.