GA4 audit · GTM · e-commerce

First establish which GA4 data you can trust.

I review the complete measurement chain—from the dataLayer and GTM to GA4, e-commerce, Consent Mode and advertising integrations. I show where the data distorts decisions and what should be fixed first.

I do not promise a one-to-one match between GA4 and a CRM or store. I separate implementation errors from differences caused by definitions, consent, delays and the limitations of the platform.

Evidence, not a checklisttests, screenshots and reproducible steps
Priority by decision impactrevenue, conversion and reporting risks come first
An implementable planowners, dependencies and effort estimates

When an audit is useful

The data exists. The problem is knowing whether it is comparable.

An audit is the right first step when measurement is already running, but the results are disputed or nobody can explain the source of a mismatch.

Conversions do not reconcile with the CRM or store

A shared definition, reference system or explanation of the difference is missing.

The e-commerce funnel has gaps or duplicates

purchase, transaction_id, item data and refunds do not form a consistent view of sales.

The GTM container grew without documentation

Legacy tags, DOM dependencies and unknown triggers make every change risky.

Attribution breaks across domains

Self-referrals, payment gateways, subdomains or incorrect UTMs change session and conversion sources.

The CMP and Consent Mode work separately

It is unclear which tags run for each consent state and when they are triggered.

The team debates numbers instead of acting

Marketing, sales and finance use different definitions of the same conversion.

Audit outcome

A report that leads from a symptom to a repair decision.

Every material issue is described with evidence, impact, a recommendation and an acceptance check. The audit can enter a backlog instead of becoming a presentation that is filed away.

Findings register

Issue, environment, reproduction steps, evidence and area of ownership.

Data reconciliation

Comparison of critical events and values with an agreed reference system.

Business priorities

Critical, important and optimisation items, with the consequence for decisions.

Repair roadmap

Sequence, dependencies, required roles and indicative effort.

Core review areas

01

Measurement architecture

Property, streams, domains, filters, referral exclusions and systems of record.

02

Events and conversions

Names, parameters, trigger conditions, duplicates and conversion definitions.

03

E-commerce and dataLayer

Funnel, item data, transaction_id, currencies, values and edge cases.

04

GTM, consent and integrations

Tags, triggers, CMP, Consent Mode, Google Ads and technical dependencies.

05

Quality and reconciliation

Completeness, consistency, delays and comparison with the store, CRM or backend.

An audit does not automatically include rebuilding measurement. When recommendations require changes, the implementation scope is agreed only after the causes and dependencies are known.

Explore GA4 implementation

Process

From a business question to a verified backlog.

  1. 01

    Problem qualification

    We establish the decisions that depend on the data, symptoms, sites, integrations and reference systems.

    Outcome: scope, access and criteria
  2. 02

    Evidence collection

    I review the configuration, container, dataLayer and data flow, then reproduce the critical scenarios.

    Outcome: documented observations
  3. 03

    Testing and reconciliation

    I test duplicates, omissions, consent and domains, and compare critical results with the reference system.

    Outcome: causes and confidence level
  4. 04

    Priorities and review

    I order recommendations by impact, dependency and effort, then review them with the people who will implement the changes.

    Outcome: report, meeting and roadmap

A tangible result

How I describe an individual finding.

This is an illustrative structure, not data from a specific client. The report separates fact, interpretation and recommendation.

P1 · criticalpurchase fires again after returning from the payment gatewayIllustrative audit finding
Symptom
GA4 revenue is higher than the order system for payments that redirect customers outside the store.
Evidence
The same transaction_id is sent again when the return page is refreshed.
Impact
Campaigns learn from inflated purchase counts and the reported ROAS is too optimistic.
Recommendation
Send purchase after order confirmation and prevent the identifier from being emitted again.
Acceptance check
One purchase per transaction_id across purchase, return and refresh scenarios.

Choosing the first step

Not every measurement problem starts with an audit.

The current state is unknown

GA4 audit

When data exists, but it is unclear what works, why numbers differ and what should be fixed first.

Review the scope
Definitions are missing

Measurement plan

When KPIs, events, parameters, the dataLayer and ownership must be agreed before configuration.

Explore the plan
The scope is known

GA4 implementation

When specifications and issues are known, and configuration, changes, QA and documentation are required.

Explore implementation
Measurement is live

Data quality monitoring

When critical data requires ongoing controls, alerts and a response process after releases.

Explore monitoring

Conditions for a useful audit

What I take on and what I need from your team.

My responsibility

  • test plan and the list of data needed for verification,
  • review of GA4, GTM, the dataLayer, consent and agreed integrations,
  • reproduction of critical scenarios and evidence capture,
  • assessment of impact, priority and implementation dependencies,
  • report and a meeting to review the findings.

Client responsibility

  • access limited to the agreed accounts and environments,
  • a person familiar with business goals, campaigns and measurement history,
  • a reference system such as a CRM, order platform or backend,
  • the ability to perform a test purchase, lead or other business scenario,
  • participation from the people who will decide how recommendations are implemented.

Honest limitations

An audit increases confidence, but it does not turn GA4 into an accounting system.

I document confidence levels and the limits of each conclusion instead of hiding them behind a promise of “100% accurate data”.

  • GA4 does not have to equal the backend.Differences can result from definitions, cancellations, delays, consent and modelling.
  • Historical gaps usually cannot be reconstructed.An audit identifies the cause and improves collection from the point a change is released.
  • Conclusions depend on access to evidence.Without a reference system, implementation can be assessed but the scale of a mismatch may remain uncertain.
  • Consent creates unavoidable gaps.Correct configuration cannot make every session and conversion observable.
  • A technical audit is not legal advice.I test tag behaviour; the client or its adviser assesses the legal basis.
  • Recommendations have a shelf life.Changes to the app, checkout, CMP and campaigns require later testing or monitoring.
Krzysztof Surowiecki

One accountable consultant

I connect technical diagnosis with business consequence.

I lead the audit personally—from questions about decisions and systems of record, through GA4 and GTM testing, to the discussion with the people who will implement the fixes. Priority therefore reflects more than the number of errors.

20+years of experience
150+analytics and data projects
50+clients

The figures cover the wider analytics and data practice, not GA4 audits alone.

Before we talk

Frequently asked questions.

How long does a GA4 audit take?

It depends on the number of sites, containers, integrations, consent variants and the availability of reference data. After a short qualification, I propose a scope and schedule instead of applying one duration to every environment.

What access do you need?

Usually, agreed access to GA4 and GTM, the ability to test the site and comparative data from a CRM or order system are enough. Google Ads, the CMP, BigQuery and other systems are included only when they belong to the scope.

Will GA4 match the backend after the audit?

I do not guarantee a one-to-one match. The goal is to understand the differences, remove implementation errors and agree tolerances for critical metrics.

Does the audit include fixing the issues?

The core deliverable is diagnosis, priorities and a repair plan. I can estimate implementation after the audit, when the causes, development dependencies and acceptance criteria are known.

What exactly will I receive?

A findings register with evidence, a data quality and reconciliation summary, priorities, recommendations, a roadmap and a review meeting. The format is designed to move directly into a backlog.

How is the audit priced?

Pricing depends on the number of sites, containers, e-commerce scenarios, integrations, reference systems and the required testing depth. After qualification, I identify the core scope and optional elements.

First step

Describe one mismatch that blocks a decision today.

I will respond with questions about the site, reference system, GTM, consent and the people needed for the audit. If an audit is not the right first step, I will say so before preparing an estimate.

Before ordering an audit

Check the basics yourself.