Product metrics and goals
Metric tree
Also called Driver tree, KPI tree
A diagram that breaks a top metric into the metrics that drive it, level by level, so you can see how each team's numbers roll up.
How it's worked out
Each metric = the sum or product of the metrics beneath it, where the maths allows
A tree makes the logic explicit: weekly active users are new users plus returning users, and returning users are last week's users times retention. Where the maths holds, you can see how much a change at the bottom moves the top.
Not every link is arithmetic. Some metrics support others without adding up to them, and a good tree says which is which.
In practice
Revenue = orders × average order value; orders = visitors × conversion; conversion = add-to-cart rate × checkout completion.
Common mistake
Treating correlations as if they were formulas, and expecting a branch to move the top by an exact amount.
Where it comes from
Driver trees go back to the DuPont company's return on investment formula of the early twentieth century, which broke one figure into the margins and turnover behind it.