AI Council on YouTube · Open on YouTube
Experimentation, causal thinking and metric design · Proficient · Lesson 63 of 99 · 28 min video, then about 2 hours of practice
Building metric trees that connect daily work to outcomes
After this you can decompose an outcome metric into a tree of input metrics, identify which inputs teams can influence, and use the tree to explain a metric movement.
Why it matters
A metric tree breaks an output (revenue) into the inputs that produce it (new customers times conversion times order value, and so on), down to levers teams control. It turns 'revenue is down' into a structured investigation and tells each team which number is theirs. The trap is building a tree of arithmetic identities without asking which nodes anyone can actually move.
How to use this resource
Abhi Sivasailam, a growth and analytics leader, on designing metric trees and operationalising them; the clearest talk on the subject.