← Back to framework
Tools / Bucket 2

Analytics and measurement

Multi-touch attribution, reading across platforms, and spotting when numbers lie. Five layers of truth from one campaign.

Tools and platforms

/Google Analytics 4

Website and app analytics. Tracks user journeys, conversions, and events. The move from Universal Analytics to GA4 changed how sessions and attribution work. Event-based, not pageview-based.

analytics.google.com
/Platform reporting (Meta, Google, TikTok)

Each ad platform has its own reporting dashboard. They each claim credit for the same conversions. The skill is reading across all three without double-counting.

/Holdout and incrementality testing

Withhold ads from a control group and measure the difference. The only reliable way to know if advertising caused the sale or if the customer would have bought anyway.

/Dashboards and data visualisation

Tools like Looker Studio, Tableau, or even spreadsheets. The tool matters less than the questions you ask of it. A dashboard shows what happened, not why.

lookerstudio.google.com

What AI does vs what you do

AI handles
  • Pull data from every platform automatically
  • Build dashboards and visualisations
  • Spot trends and flag anomalies
  • Surface correlations across data sets
  • Calculate statistical significance
You decide
  • Decide which questions actually matter
  • Determine causation vs correlation
  • Design holdout tests to prove incrementality
  • Interpret why things happened, not just what
  • Spot when a metric looks good but means nothing

Skills to develop

/Correlation vs causation thinking
/Holdout test design
/Incrementality thinking
/Question formulation (asking the right question matters more than the tool)
/Contextual analysis (what else was happening that week?)
/Reading across platforms without double-counting

AI is excellent at spotting correlations. But it takes human judgment to figure out why. This is called incrementality thinking, and it is one of the most valuable analytical skills you can develop.