The GA4 Audit Hub

The GA4 audit most teams skip.

GA4 is not broken. It is biased by defaults nobody on most marketing teams has ever audited. This hub is the operating reference for the cleanup — the architecture argument, the 8-point audit checklist, the FAQ, and the path to a measurement layer your finance partners will actually believe.

Jump to the 8-point audit ↓ Read the anchor essay →
On this page
The argument in one paragraph

Why default GA4 is misleading you.

None of the issues uncovered by a real GA4 audit are bugs. They are defaults — designed reasonably for the median GA4 user, which is not an enterprise marketing organization spending eight figures a year inside Google. Data-driven attribution that nobody can audit, modeled conversions filling in for consent-mode gaps, enhanced measurement counting scrolls as engagement, cross-domain stitching that quietly over-credits Direct: each one is a sensible default at small scale that becomes a structural measurement bias at enterprise scale.

The fix isn't replacing GA4. It's treating measurement as an ongoing engineering discipline — with a budget, a roadmap, and an executive owner — and starting with an audit that names every place the data has been biased before anyone in leadership saw it. This hub is built around that audit.

Read the full essay: Your GA4 Setup Is Quietly Lying to You — A Teardown →

Where the bias lives

The five default GA4 behaviors silently distorting your measurement.

Each of these is a reasonable default in isolation. Stacked, they compound into a measurement environment where the answer to "is the data trustworthy?" has to be "it depends what you're using it for."

Default 01
Data-driven attribution, unaudited
GA4 has shifted most properties to data-driven attribution. The methodology is closed — there's no published audit trail for how the model assigns credit. For most decisions that's fine; for budget allocation decisions it isn't, because nobody on the team can defend the number when finance asks.
Default 02
Modeled conversions, opaque ratio
Consent-mode v2 and signal loss mean GA4 increasingly reports modeled conversions alongside observed ones. The reported ratio is rarely visible at the executive level. Teams making budget decisions on conversion counts are often making them on a number with a meaningful modeled component nobody told them about.
Default 03
Enhanced measurement counts everything
Scrolls, outbound clicks, file downloads, and video starts all fire automatically and feed into engagement metrics. That makes a footer-scroll session look engaged and a 40-second read look not engaged. The engagement signal becomes a noisy proxy that then feeds Google Ads audiences and bidding.
Default 04
Cross-domain stitching, invisible failures
GA4 stitches sessions across domains and devices using its own logic. Some stitches succeed. Many fail. Failed stitches show up as new sessions and direct traffic — which is why "Direct" is so often the top-converting channel in enterprise properties. It usually isn't direct. It's broken stitching.
Default 05
Conversion event hygiene drift
Conversion events accumulate over time. Without a quarterly review, you end up with events flagged as conversions that are firing from bots, internal traffic, or test environments. The model is then learning from a polluted signal you never explicitly approved.
The diagnostic

The 8-point GA4 reality check.

Run this before the next quarterly business review. If the team can't answer all eight items in writing, you're making decisions on a dataset you haven't audited.

The 8-point GA4 reality check

Designed to be completed in one focused 90-minute analytics session, with the GA4 admin and the channel lead in the room.

Take it with you

The expanded GA4 audit checklist.

The expanded version of this audit — with the specific GA4 reports, settings paths, and decision thresholds — is included in the 2026 Signal Architecture Framework. Designed to be walked into a quarterly analytics review.

The Expanded GA4 Audit

Embedded inside the full Signal Architecture Framework, alongside the Signal Flow Map, the Meridian Readiness Checklist, and the Cost Per Decision Worksheet.

Get the framework → Read the anchor essay →
The compounding cost

What an un-audited GA4 actually costs you.

GA4 measurement biases don't just live in the dashboard. They feed Google Ads audiences. They train Performance Max's bidding. They influence the conversion values your team optimizes against. A scroll event being counted as engagement at the GA4 layer becomes a low-quality audience signal at the activation layer. A modeled conversion ratio nobody verified becomes a misallocated budget at the planning layer. The bias compounds, and the compounding shows up in the P&L months after the original setup decision.

The cleanest way to put a dollar value on this is the Cost Per Decision calculator. The decisions made on bad data take just as long, cost just as much, and unwind in the same way — only the direction of the unwind is unpredictable. Cleaning up the GA4 layer is the cheapest decision-quality investment most marketing teams can make in 2026.

Frequently asked

Five questions about GA4 measurement, answered straight.

Is GA4 inherently broken for enterprise measurement?
No. GA4 is a well-engineered tool with reasonable defaults for a small-business median user. The issue at enterprise scale is that those defaults — modeled conversions, data-driven attribution without audit trails, enhanced measurement counting everything — quietly bias the data. The fix is treating measurement as an ongoing engineering discipline, not replacing the tool.
How often should an enterprise property be audited?
Once a quarter at minimum. Defaults change, conversion events accumulate, audience definitions drift, consent rates shift, and cross-domain configurations break when teams launch new properties. The audit is the discipline; the dashboard is the symptom.
What's the relationship between GA4 and Marketing Mix Modeling?
They answer different questions. GA4 measures what happened inside the tracked customer journey — useful for in-flight tactical decisions and audience activation. Marketing Mix Modeling measures the incremental impact of channel investment on a business outcome, using aggregated data. The two coexist; smart teams use GA4 for tactics and MMM for budget allocation.
Why is "Direct" so often my top-converting channel?
Almost always cross-domain or cross-device stitching that's failing more often than the team realizes. Failed stitches show up as new sessions with no referrer, which GA4 categorizes as Direct. The fix is configuration work — linker parameters across all customer-journey domains, a properly tested handoff between marketing site and storefront — not a marketing investment in "Direct."
Does Consent Mode v2 mean GA4 conversion data can't be trusted in the EU?
It means it can't be trusted unless you know the consent acceptance rate and the modeled conversion ratio. With both numbers visible, GA4 EU data is usable for trend analysis but rarely usable for absolute targeting or budget allocation. The honest move is to layer in MMM or geo-experiment evidence for decisions that touch real money.

Stay Sharp.

Get the next framework, hub, and experiment the moment it's published.