AI Tracking, Attribution & Analytics
Attribution is broken. Not philosophically — practically. Most marketing teams are running attribution setups that were designed for a cookie-first, multi-touch, MTA-friendly world, and almost none of those conditions still apply. iOS 14.5 broke half of it. GA4 broke the other half. Most attribution software is a lagging indicator of what actually matters. This pillar is where we document what actually works now — the models, the tools, the workflows, the calculations — from running attribution at sprint scale for our own marketing and for client deployments. Server-side from day one, on your infrastructure, measuring real revenue per channel, not platform-reported ROAS.
attribution fundamentals
Sub-clusterserver-side tracking
Sub-clustermulti-touch models
Sub-clustertools & comparisons
Sub-clusterdefinitions
Sub-clusterIf your tracking is wrong, every decision you make is wrong. Real CAC, real LTV, real ROAS — measured per channel, not estimated by a vendor.
— Attribution thesisAttribution
Walking through the practical differences between attribution models in GA4 and how switching from last-click to data-driven changes where you should be spending your budget.
Read post:Attribution
Step-by-step guide to deploying a GTM server container, routing GA4 events through your first-party domain, and verifying data integrity.
Read post:Attribution
Most UTM setups create noise, not signal. This is the naming convention and governance model that turns campaign parameters into a clean attribution system.
Read post:Attribution
Browser-side GA4 loses 30–40% of conversions to ad blockers and ITP. Server-side tracking routes events through your own domain, restoring the data you're missing.
Read post:Attribution
Data-driven attribution — what it actually measures
Data-driven attribution (DDA is Google’s machine-learning-based approach to distributing credit across marketing touchpoints. Instead of applying a fixed rule (giving all credit to the last click or s
Read post: Data-driven attribution — what it actually measuresAttribution
Multi-touch attribution without enterprise prices
Multi-touch attribution is a statistical methodology, not an enterprise software feature. Any team with GA4, BigQuery export, and basic SQL can implement it for under £200/month. Here’s the full stack, the three SQL queries that do the work, and what it actually costs.
Read post: Multi-touch attribution without enterprise pricesAttribution
Marketing Attribution Models: A Decision Framework for Teams Who Don't Have Unlimited Data
A marketing attribution model assigns credit for conversions to the touchpoints that influenced them. The six models that matter in 2026 each suit a different data maturity level. Which one fits depends less on the model and more on the data infrastructure you already have.
Read post: Marketing Attribution Models: A Decision Framework for Teams Who Don't Have Unlimited DataMarketing Operations
Why most AI marketing automation projects stall at month 3
Read Why most AI marketing automation projects stall at month 3Marketing Operations
The AI marketing operations playbook that's built for reality
Read The AI marketing operations playbook that's built for realityContent Strategy
ReadAI Models
ReadMarketing Operations
AI marketing automation and the central hub for unified system thinking.
Explore Marketing OperationsContent Strategy
AI content production at scale, with quality and brand voice as constraints.
Explore Content StrategyAI Search
Showing up where buyers ask AI — Google AI Overviews, ChatGPT, Perplexity, Claude — plus the AI-era SEO that still feeds them.
Explore AI SearchAI Models
Comparing Claude, ChatGPT, Gemini and others — and picking the right model for each marketing job we run for clients.
Explore AI ModelsBook a discovery call
Sixty minutes. No slide deck. We tell you which sprint tier fits and when the attribution function starts running.
Get in touch →