Grade on settled cash every week, know the three distortions and the one that runs the other way, reprice on evidence, diagnose properly when a number moves, and resist the obvious additions.
Everything up to here produces numbers. This part is about the fact that most of those numbers are wrong in specific, knowable, directional ways, and about what to do once you can see through them.
This part reads a ledger, so it assumes settled payments already carry the campaign that produced them, which is Part 9's whole job. If that isn't built yet, the two numbers this part needs most, this week's settled cash and which campaign it came from, don't exist, and the weekly scorecard has nothing to read.
Everything up to here produces numbers. This part is about the fact that most of those numbers are wrong in specific, knowable, directional ways.
The weekly scorecard reads settled cash and the campaign that produced it. Nothing else qualifies as the number to grade on.
I built analytics models for banks and miners before this. Every step in The Book Is the Funnel is one we ran on our own funnel first, failures included. This part is sold on its own because most people need one piece of it, not all sixty at once.