How a tracking audit unfolds: protocol, duration, intermediate deliverables

A tracking audit unfolds in four distinct phases over a 2-4 calendar week window. Phase 1 — technical crawl and tag inventory (2-3 days). Phase 2 — GTM container audit (2-4 days). Phase 3 — GA4 and BigQuery configuration review (1-3 days). Phase 4 — acceptance tests on critical conversion journeys (1-2 days). The same method applies whether you have a brochure site or a 30-million pageview platform — only scope adapts.

Phase 1 — Technical crawl and tag inventory

During the first 2-3 days, I document every tag actually deployed on your site using Screaming Frog in tag detection mode on key pages, completed by ObservePoint on large sites to automate hundreds of page scans.

For each detected tag, I record: publisher, apparent firing condition, parameters pushed, CMP pre-consent compliance, cookie lifetime. At the end of this phase, you receive a CSV inventory of tags present, classified by publisher and Consent Mode category. Many sites have never had this inventory formalized.

What I ask at this stage: read access to your GTM container (web + server-side if applicable), read access to your GA4 property, list of subdomains in scope.

Phase 2 — GTM container audit

During the next 2-4 days, I review the GTM container’s internal organization. I document: tag naming, trigger and variable structure, misconfigured built-in variables (a classic: Click URL instead of Click Element URL), orphaned or duplicated tags, firing rules with their exceptions. If you have an sGTM container, I audit its routing, claims, event transformations, Meta CAPI integrations or Enhanced Conversions.

I note each anomaly with its severity (critical, major, minor) and estimated correction effort. This phase produces most of the findings — typically 60-70% of anomalies surface in the GTM container, not GA4.

Phase 3 — GA4 configuration review

During 1-3 days depending on complexity, I document GA4 property by property: data streams, recommended events, conversions, audiences, custom parameters, custom dimensions, referral exclusions, internal filters, attribution model. I verify Enhanced Ecommerce events are correctly implemented if applicable.

If BigQuery is connected, I audit exports too: ga_session_id, user_pseudo_id presence, event_params cleanliness, custom dimension completeness in events_* tables.

Phase 4 — Acceptance tests on critical journeys

The phase that distinguishes an honest audit from a theoretical one. During 1-2 days, I replay conversion journeys in real conditions: private browsing, mobile (iOS Safari and Android Chrome), with consent accepted then refused.

For ecommerce: product search → product page → add cart → checkout → confirmation. For B2B lead gen: key page → form → confirmation → CRM follow-up. For media: article → scroll → engagement → subscription conversion. I note at each step which events actually fire, with which parameters, and where the gaps are with the theoretical tagging plan. This phase surfaces the highest-impact anomalies — those nobody had seen because no one was testing consent-refused mobile journeys.

The restitution

At the end of the four phases, I draft the report — details on the page about what an audit report contains — and schedule a 1-hour video restitution with your team to prioritize the action plan.

Frequently asked questions

Do you work in our environments or yours?

Always yours. I never use mirror or test accounts — production reality is what matters. NDA signed if your legal team needs it.

Need to pause campaigns during the audit?

No. The audit runs on production tracking without modifying anything — it’s diagnostic, not intervention.

Do you implement the fixes?

The report is designed so a developer or marketer can implement directly — each anomaly is described with its precise fix. If you want me to lead implementation, that’s a separate engagement.