Key Result

A Key Result is a measurable outcome, one to four per Objective and two on average, that tells you whether you are actually progressing. The preferred shape is simple: increase or decrease a named metric from X to Y.
The from-X-to-Y formula does quiet, important work. It forces a baseline (where are we now?), a target (where are we going?), and a direction, which together turn a vague aspiration into something you can be right or wrong about. "Improve activation" is a wish. "Increase first-session activation from 34% to 55%" is a Key Result: it names the metric, admits the starting point, and commits to a number someone can check at quarter end.
Choose metrics that move, and that matter
The best Key Results use fast-moving metrics that genuinely matter, chosen with a metric tree so you can see how a given number connects to the outcome above it. Two failure modes bracket the choice. A metric that matters but barely moves in a quarter (annual revenue, brand awareness) gives you no feedback in time to act. A metric that moves easily but does not matter (page views, tickets closed) gives fast feedback on the wrong thing. Aim for the intersection: leading indicators close enough to the real outcome that moving them is progress, not theatre.
When you cannot set a target yet
Sometimes you genuinely do not know where to set Y. That is a signal, not a failure. If you lack a baseline, use a baseline Key Result for a cycle: measure honestly, establish the number, and set a real target next quarter. If the metric that matters is too slow to move inside a quarter, use a proxy metric that leads it. And if you lack the insight to write any meaningful Key Result at all, the honest move is to do the discovery first, through dual track discovery, not to dress the research up as a "discovery Key Result" and grade it. Learning work and delivery work follow different rules.
Avoid task-based Key Results
The most common mistake is the task-based Key Result: "launch the new pricing page", "hire two engineers", "run the campaign". These are activities, not outcomes, and they can all be completed while the thing you actually wanted, more revenue, faster delivery, happier customers, does not move at all. A Key Result should describe the change in the world, not the work you did hoping to cause it. The work belongs in your initiatives; the Key Result is the evidence the work paid off.
Stretch, narrative, and honesty
Where a target is a deliberate stretch, label it as an aspirational target so nobody mistakes an ambitious reach for a broken commitment, and give each Key Result a short narrative explaining the belief behind it. This is where honesty is won or lost: unlabelled stretch targets, quietly graded as failures, teach teams to sandbag next quarter. Labelled clearly and judged on evidence, they let teams reach without fear, which is the entire point of setting a difficult goal.
The strategic reading
At the highest level, Key Results are how you verify a strategy is working. If your theory of advantage is real, customer behaviour should move in a direction it predicts, and your Key Results are where that shows up. Read this way, a Key Result is not just a progress bar; it is evidence of advantage, or evidence that the bet was wrong, which is exactly the feedback a strategy needs to improve.
Task-based (avoid): "Ship the referral feature." Outcome (use): "Increase referred sign-ups from 5% to 20% of new users." The first is done the day you ship; the second is only done if the feature actually worked.
How many Key Results per Objective? +
One to four, two on average. More than four usually means the Objective is doing two jobs and should be split, or that some Key Results are really tasks.
What is a baseline Key Result? +
One used when you cannot yet set a credible target: you measure to establish the number this quarter, then set a real from-X-to-Y target next quarter. It is honest, not a cop-out.
Can a Key Result be a yes/no milestone? +
Rarely, and only when a genuine binary outcome matters (a certification achieved, a market entered). Most milestones are tasks in disguise; prefer a metric that shows the milestone created value.
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.
Most OKRs fail on the Key Results: task lists in disguise. We coach teams to write outcomes that prove progress, and the free AI OKR Coach flags the task-based ones on sight.