End-to-end
measurement
infrastructure. Not
just another tag
implementation.
Four tiers, in sequence. Each builds on the last. You decide how far to go.
I structure engagements as discrete tiers with decision gates between them. You buy what you need, see the results, and decide whether to go further. The first two tiers have non-negotiable components: the things that must be done properly or not at all. Everything else is modular and scoped to your specific situation.
Measurement Foundation
Server-side GTM moves your tracking off the browser and onto a server you control. That's where consent stops being broken, cookies stop expiring on you, and match rates start looking like the platforms claim they do.
Tag Audit & Data Layer Spec
I audit your existing container and produce a documented spec that becomes the blueprint for the server-side implementation. Implementing sGTM on top of a messy client-side container just moves the mess server-side.
Server-Side GTM Deployment
Full sGTM setup with custom domain and same-origin config for extended cookie life. Hosted on Stape by default; EU-hosted or self-hosted on Google Cloud for enterprise.
Warehouse Event Sink
Your sGTM event stream flows directly into a data warehouse from day one. Without it, Tier 1 is a dead end, and you should hire a GTM specialist, not me.
Consent Management Implementation
CMP deployment integrated with GTM consent mode, enforced server-side. Cookiebot by default, or bring your own.
Event Delivery Configuration
GA4, Meta CAPI, Google Ads EC standard. Additional platforms like TikTok, LinkedIn, Pinterest, or Snap, scoped to your channel mix.
Real-Time Event Architecture
Implementation designed with awareness of future real-time streaming (Pub/Sub, webhooks), no decisions that close doors.
Server-side tracking that's properly consented, delivering higher-quality data to your ad platforms and streaming raw events into your warehouse.
Data Infrastructure
Raw events landing in a warehouse is the easy part. Turning them into a documented, tested, queryable layer your team and your AI tools both trust is the work.
Warehouse Configuration
BigQuery project setup with proper dataset structure (raw, staging, marts), IAM, and cost controls. The organizational foundation that prevents your warehouse from becoming an ungoverned swamp.
Dataform Transformation Models
A standardized medallion pipeline: raw → staging → marts. Campaign performance, channel attribution, LTV, session events. The single most valuable deliverable of the engagement. Clients with existing dbt implementations stay on dbt, with a migration conversation available.
Automated Data Quality Framework
Dataform assertions that continuously validate schema, null rates, volume anomalies, and source freshness. Silent data failures are how teams make decisions on broken numbers for weeks.
Data Dictionary & Schema Docs
AI-generated, human-reviewed docs of every model, column, and dependency. Ships as a deliverable and lives in the repo, makes your data self-describing for your team and for any AI that needs to work with it.
AI Access Layer
A scheduled BigQuery export to Cloud Storage in parquet or JSON, a clean handoff point for Claude, ChatGPT, Gemini, or your own internal systems, no direct warehouse access required.
Data Pipeline Deployment
Ingestion of ad spend, CRM, and e-commerce data into BigQuery. BigQuery Data Transfer Service by default; Airbyte for sources it doesn't cover.
Automated Campaign Taxonomy
Dataform models using BigQuery's native AI to standardize campaign names, UTMs, and ad-group naming. Runs continuously, kills the eternal spreadsheet-mapping exercise.
A single source of truth for your marketing data. Clean, documented, AI-enriched, accessible to any downstream tool or AI system.
See it live →Business Intelligence
Dashboards are a means to an end. The end is your team answering their own questions, in their tools, on their timeline.
Dashboard Design & Build
Executive summary, channel performance, and campaign drill-down dashboards on the Tier 2 mart layer. Looker Studio or Metabase, depending on your embedding and access needs.
Natural Language Data Exploration
Gemini-powered querying in BigQuery for technical users. For marketing teams: the Tier 2 AI Access Layer means you query data with whatever AI tool your team already uses.
Automated Narrative Reporting
A RAG pipeline built natively in BigQuery that generates written performance summaries delivered to Slack or email on your schedule. Replaces the weekly meeting where someone reads a dashboard aloud.
Your marketing team self-serves answers instead of asking the data person. Your executives get a written summary every Monday that tells them what changed and why.
See it live →Attribution & Advanced
Multi-touch attribution, geo-lift incrementality, MMM. The rigorous methods that let you say “this channel is incremental,” not “this channel correlates with revenue.” Reserved for clients who’ve completed Tiers 1 and 2 with me. The methodology is only as trustworthy as the pipeline it sits on.
Multi-Touch Attribution Modeling
Shapley value or position-based models in Dataform on your event data. Replaces platform-reported attribution with logic you own, can inspect, and can trust.
Geo-Lift Incrementality Testing
The most rigorous, privacy-friendly method for measuring whether a channel is driving incremental revenue or just capturing existing demand. Directly measures causation rather than inferring it from correlation.
Media Mix Modeling
Bayesian MMM using Google's open-source Meridian framework, running on BigQuery. A budget-allocation model with its assumptions out in the open.
Measurement Strategy & Roadmap
Ongoing advisory on what to test next, which channels to validate, and how to build a culture of measurement. The retainer component, not maintaining infra, but guiding the questions.
Attribution you own, methodology you can defend to your CFO, a testing roadmap that compounds what you've learned.
Not sure where you'd start? See your session first.
Every page on this site is instrumented with the same Tier 1 foundation. Open your session on any page to see it flowing through the stack. The best pitch I have is the one running underneath you.