Data & Attribution
The Quarterly Data Quality Audit
Most bad decisions made from data are not made from the wrong analysis. They are made from the right analysis on broken inputs.
8 min read
Data & Attribution
Attribution is a modelling problem, not a reporting one. Incrementality tests, UTM governance, warehouse-first measurement and the KPI trees that keep teams pointed at the same number.
Data & Attribution
Incrementality Testing: The Only Number Your CFO Should Trust
Platform-reported ROAS answers "who touched the sale?" Incrementality answers "would it have happened anyway?" Only the second one supports a budget decision.
8 min read
Data & Attribution
Predicting Lifetime Value Without Fooling Yourself
An LTV number that nobody can falsify is not a forecast. It is a permission slip for overspending.
9 min read
Data & Attribution
Geo Holdouts: The Most Practical Causal Test
When user-level tracking fails, geography is still a reliable way to build a control group.
9 min read
Data & Attribution
Dashboards People Actually Open
If a dashboard cannot change what someone does tomorrow, it is a screensaver with numbers.
8 min read
Data & Attribution
Write the Tracking Plan Before the Code
Nobody regrets writing a tracking plan. Everyone regrets the six months of data they had to throw away.
8 min read
Data & Attribution
First-Party Data Strategy That Survives Consent
The question is no longer how much data you can collect. It is how much you can justify collecting.
9 min read
Data & Attribution
UTM Governance and KPI Trees: Boring Work That Makes Data Usable
Nobody gets promoted for standardising campaign tags. But every reliable marketing dashboard sits on top of exactly that work.
6 min read
Data & Attribution
Mix Modelling Is No Longer Only for Giants
Mix modelling went from a luxury for large advertisers to a practical option for anyone with tidy weekly numbers.
10 min read
Data & Attribution
Every Attribution Model Is Wrong in a Different Way
Choosing an attribution model is not choosing the truth. It is choosing which distortion you can live with.
9 min read
The briefing
One deep-dive per week. No growth-hack listicles.
Channel breakdowns, measurement teardowns and the numbers behind them — sent Thursdays.