North Star & Inputs

The Aim step of the strategy process: the one outcome every downstream choice is judged against. Next comes the diagnosis: where the leverage to move it actually sits.

Reading Data introduces north star + inputs as a framework for turning a number into a decision: one outcome metric that captures real value, and the two or three inputs you can actually move that drive it. That’s the mechanics. This page is the strategy layer built on top of it: how a company picks its one north star, and how that choice shapes what every team works on quarter to quarter.

First of three. This page picks the one company outcome and the inputs that drive it. Goals & Success Metrics turns an input into a goal a team commits to, and Setting Good Metrics is the craft of writing one that survives contact with a team.

Why exactly one#

A north star is the single outcome metric that captures the real value the business creates for the people it serves: not a proxy for it, the thing itself as closely as you can measure it. It is deliberately singular.

The discipline is in the “exactly one.” A team with five north stars has none: five competing priorities, each justifying a different roadmap, with no way to adjudicate between them. A second “north star” removes your answer to the only question that matters: if we could only move one number this quarter, which one? A single north star forces that answer in advance, before a real tradeoff forces it for you.

This doesn’t mean nothing else gets tracked. It means everything else (including every input metric below) earns its place by explaining how it moves the one number that outranks it.

Choosing the one#

The north star has to be the outcome, not an activity that leads to it. Signups aren’t the north star of a marketplace; successful transactions might be. Orders started aren’t the north star of an online store; orders that actually complete are closer. The tests from Setting Good Metrics apply at full strength here: it must tie to a decision, and it must be hard to game, because everyone in the company will eventually be measured against it.

It’s usually lagging by nature: value realised is a downstream result, not something you observe the moment it starts. That’s fine. The north star doesn’t need to be actionable day to day; the inputs do that work.

The inputs#

The inputs are the two or three metrics that drive the north star and that a team can actually move this quarter. Two properties matter, and both are non-negotiable:

• They drive the outcome. Not correlate with it: you should be able to state the mechanism. If you can’t explain why moving this input moves the north star, it’s not an input, it’s a coincidence.
• They’re actually movable. An input a team can’t act on is the north star wearing a disguise: restating the goal instead of decomposing it into something to work on.

If you can’t name the inputs, you can’t act on the outcome: only report it after the fact. Naming them turns “we want more value delivered” into “we improve activation, then retention, then expansion,” each with an owner.

Keep the list to two or three. More than that stops being a set of levers and starts being a wishlist: everything looks like an input if the bar is low enough, and a long list gives every team a metric to point to instead of forcing a real priority.

North Star vs. Diagnosis. This page is only about choosing the one metric and its inputs, the aim. It stops there. Decomposing those inputs to find where the leverage actually is, and sizing it, is the next step: the Diagnosis. North Star says what to move; Diagnosis says where to push.

Connecting to goals and bets#

The north star and its inputs are not the roadmap; they’re what the roadmap is judged against. At strategic planning, every candidate initiative carries a metric goal, a specific number the team wants to move, before Discover or Define has even started. That metric goal should trace back to one of the inputs, and from there to the north star. An initiative that can’t be connected to an input is either mis-scoped or solving a problem that doesn’t matter yet, the same question quarterly planning already asks: does this fit where we’re going?

Bets is where this gets formalised: turning “grow the input metric” into a goal a team commits to for the quarter, with the north star as the reason it matters and the inputs as the language the goal is written in.

What teams actually own#

The north star belongs to the company; no single team can move it alone, and holding one team accountable to it invites gaming. The inputs are what get distributed: each should have an owning team, the one whose quarter-to-quarter work is judged by whether that number moved.

This is the same shape as the capacity allocation in Half Strategy: categories get a budget, initiatives get sized, and each initiative’s metric goal should point at an input the owning team is accountable for. When a team asks “what are we actually trying to move this quarter,” the answer should be an input, not the north star itself, and not a made-up local metric with no line back to it.

Example#

Illustrative example, threaded through this documentation: an online store.

North star (completed orders per week): orders that actually ship and reach the customer (payment captured, order fulfilled), not orders merely started or added to cart. Say it sits at ~1,900 / week today. The whole game is to grow that number without widening the top of the funnel to fake it.

It’s driven by three inputs a team can actually move, each a real metric with a value you can point at today and a target for the quarter:

Picture it as one screen: the north star on top, these three underneath. Push completion and attach up and load time down, and completed orders rises for real, because each input is a genuine step toward a completed order, not a vanity proxy sitting next to one.

Why this north star and not “orders started”: orders started is an activity metric: it climbs if the funnel gets wider even while nothing downstream improves, and it can’t fall when the real numbers are bad. It tells you the product is being opened, not that anyone was helped. “Completed” can’t be inflated the same way.

Related: The Framework, Diagnosis, Bets, Reading Data, Setting Good Metrics, Half Strategy