How I approach GTM, pixels, server-side tagging, event design, and the ownership problem behind tracking drift.
Tracking architecture is not the same as installing tags.
The architecture question is: what should the business know, what should each platform receive, under which consent state, and who owns the system when campaigns, agencies, platforms, and legal requirements change?
How I approach it
I start with inventory, not opinion. What is visible in GTM? What fires outside GTM? Which vendors are present? Which events look conversion-worthy? Which identifiers or values appear to be sent? Which parts are browser-side, server-side, or undocumented?
From there, the work becomes architectural:
- define the event model around the business model, not around platform defaults
- separate configured tracking from observed firing
- decide what belongs in browser, server-side GTM, CRM, or warehouse flows
- document consent states and platform-specific implications
- design for maintenance, not just launch
The goal is a tracking stack that marketing can use, legal can understand, and future teams can maintain.
Where this connects
Tracking architecture connects directly to media signal quality, analytics trust, privacy governance, and AI-assisted operations. If the data layer is weak, every optimization layer built on top inherits the weakness.