Metric trees
A metric tree is a stated model of cause and effect: the outcome the business is judged on at the top, the drivers that move it beneath, and at the bottom the levers a team can actually pull. It is the drawing that tells you which Key Results are worth having.
The obvious bit
Every company has a number it is judged on. Almost none can show, on one page, what moves it.
So Key Results get chosen the way most things get chosen: from what is already in a dashboard, from what somebody can pull without asking, from what looked good last time. That is selection by availability, and it is why so many teams hit their numbers while the business does not move.
A metric tree fixes the order. Draw the model first. Choose the measures second.
What a tree actually is
Three layers, and the discipline is in keeping them apart.
The outcome. The one or two numbers the business is genuinely judged on. Revenue, retention, margin, the thing your board opens with.
The drivers. What arithmetically produces that outcome. Not what influences it vaguely: what it decomposes into. Revenue is not driven by “brand”, it decomposes into volume and price, and volume decomposes further.
The levers. The things a team can actually change inside a cycle. This is where Key Results live, and nowhere else.
The test of a good tree is arithmetic, not opinion. If you cannot say roughly how much a lever moves the driver above it, the link is a hope rather than a model.
Why it beats a dashboard
A dashboard shows you what is happening. A tree tells you what to do about it, because it carries the connections. When the outcome slips, a dashboard offers you forty numbers and no argument. A tree lets you walk down the branch and find the driver that moved.
It also makes the trade visible. Two teams proposing goals on the same driver is duplication. Two goals on drivers that cancel each other out is worse, and it is invisible without the tree. This is where cross-functional collision gets caught before the quarter starts rather than during it.
The failure modes
The tree that is really an org chart. Branches drawn by department rather than by cause. It looks tidy and it teaches nothing, because outcomes do not respect reporting lines.
The tree with no arithmetic. Boxes connected by arrows nobody can defend. If “engagement” sits above “revenue” with no stated relationship, that arrow is decoration.
The tree that is never wrong. A model that cannot be disproved is not a model. Each link is a belief, so it should carry a confidence level and something that would show it was wrong. That is the same discipline we apply to strategic beliefs, one level down.
The tree built once. The tree is the drawing of your business, so it changes when the business does. Review it when the strategy changes, not annually out of habit.
Our opinion
Most companies do not need a bigger dashboard. They need the one page that says what is connected to what, and they need it before goal-setting rather than after.
It is also the cheapest institutional memory you can build. When the person who understood the model leaves, the tree stays, and the next person can argue with it rather than guess.
One line to keep: if you cannot draw the line from a Key Result to the number the business is judged on, you have found a measure, not a lever.
What is the difference between a metric tree and a KPI dashboard? +
A dashboard reports numbers side by side. A metric tree states the relationships between them: what decomposes into what, and roughly how much. The tree tells you where to act; the dashboard only tells you something moved.
How deep should a metric tree go? +
Deep enough to reach levers a team can move inside a cycle, and no deeper. If the bottom layer still needs another decomposition before anyone can act on it, keep going. If it has become a task list, you have gone too far.
Do Key Results come from the tree? +
Yes, and only from the bottom of it. A Key Result on a driver nobody can move is a report. A Key Result on a lever with a stated link to the outcome is a bet worth making.
From the ZOKRI OKR Handbook, the methodology we install and maintain. Written by Matt Roberts.
A UK-based strategy and OKR consultant and two-time SaaS founder with a venture-backed exit, Matt turns strategy into execution for teams scaling from tens to thousands. He co-founded ZOKRI in 2018, having previously co-founded Linkdex, a venture-backed enterprise SaaS platform he led to a trade sale. He writes the methodology behind these notes.
A metric tree is usually the first artefact we build with a leadership team, because everything downstream depends on it. Two sessions, and the arguments it starts are the point.