Manage your team.
How to write key results that measure outcomes, not tasks
How to write key results with a clear finish line: outputs versus outcomes, the from X to Y formula and examples by team. From the Manager Kit guide.
You have an objective you believe in. Now you need the key results, and the first draft looks like a project plan: launch this, build that, hold ten meetings. Each item is real work. None of them tells you whether the objective was met.
This guide shows how to write key results that measure the result instead of the activity. It covers the difference between outputs and outcomes, a simple formula for a clear finish line, examples for eight kinds of teams, and a short check to run before you commit.
What makes a good key result?
A good key result does three things:
- It measures a change, not effort. It counts what customers, users or the business do differently, not how busy the team was.
- It has a finish line everyone reads the same way. A start value and a target, so “done” is not a matter of opinion.
- Someone can move it this quarter. It has one owner, and that person can influence the number within about thirteen weeks.
If a key result fails any of the three, it will either be gamed, argued about or ignored.
Outputs versus outcomes
The most common weak key result is an output dressed up as a result. Here are three pairs from the Manager Kit guide:
| Weak | Better | Why |
|---|---|---|
| Launch the new onboarding flow | Raise week-one activation from 38% to 55% | The first is a task. The second is the change the task should cause. |
| Improve customer satisfaction | Lift CSAT from 4.1 to 4.5 | Without “from X to Y”, everyone imagines a different finish line. |
| Hold 10 sales meetings | Grow qualified pipeline from $120k to $300k | Count what customers do, not how busy you were. |
The weak versions are not wrong to do. The onboarding flow may well be the way to lift activation. It just belongs in the plan of work under the key result, not in the key result itself.
The “from X to Y” formula
Almost every strong key result fits one sentence:
Verb + metric + from start value + to target value (+ by when, if not the quarter’s end).
Take a vague goal, “customers should love our support”, and run it through four steps:
- Name the change. What would be different if customers loved our support? They would get answers faster, rate us higher and stay longer.
- Find the metric you can actually measure. First response time, CSAT, monthly churn.
- Write today’s value. First response is 9 hours, CSAT is 4.1, churn is 3.1% a month.
- Pick a target that stretches. Not what you will surely reach. What would make the quarter a clear success.
The result:
- Reduce first response time from 9h to 2h.
- Lift CSAT from 4.1 to 4.5.
- Cut monthly churn from 3.1% to 2%.
Three key results, each with one owner, all pointing at the same objective: customers stay because they want to.
Three types of key results
Not every key result is a count. There are three useful shapes:
- Numbers: users, hours, deals, tickets. “Cut P1 incidents from 6 to 2 a quarter.”
- Percentages: rates and shares. “Raise uptime from 99.5% to 99.9%.”
- Milestones: done or not done. “Automated vendor payments go live.”
Milestones are the easiest to write and the hardest to learn from, because they show nothing until the very end. Use them when there is no sensible number yet, keep them to a minority, and make sure each objective has at least one key result you can measure every week.
Key result examples by team
The example library in Manager Kit has 36 objectives across twelve areas, each with three key results. A few to borrow:
- Product: Cut median time to create a report from 12 to 6 min.
- Engineering: Raise deploys from 3 to 10 a week.
- Sales: Raise demo-to-close rate from 15% to 22%.
- Marketing: Grow demo requests from 60 to 110 a month.
- Customer success: Cut time to first value from 14 to 5 days.
- People: Cut time-to-hire from 52 to 35 days.
- Operations: Raise on-time payments from 82% to 98%.
- Leadership: Hold 1:1s with every report every two weeks, from 50% to 100%.
Treat these as shapes, not targets. Your starting values and your stretch will be different.
A quick check before you commit
Run each key result through these questions:
- Does it start with a verb like launch, build or hold? Then it is probably a task.
- Is there a start and a target in the title? If not, add “from X to Y”.
- Could the team hit it without anything changing for a customer? Then it measures effort.
- Does it repeat the objective in other words? A key result is the proof, not a repeat.
- Is the start the same as the target? Then progress can never move.
- Who updates it every week? If the answer is “the team”, pick one person.
Getting the wording right is half the job. The other half is checking in on the numbers every week, which is covered in OKR scoring and the weekly check-in. If you are new to OKRs altogether, start with what is an OKR.
How Manager Kit helps you write key results
In Manager Kit, each key result has a type (number, percentage or milestone), a start, a target, a unit and one owner. A built-in quality check warns when a milestone reads like a task, when there is no “from X to Y” in the title, when start and target are the same, or when a key result repeats the objective’s wording. The OKRs → Guide tab has the weak and better pairs above and the full example library to borrow from. There is no AI inside the kit: the guide gives you prompts to draft or improve key results, you paste them into the AI tool you already use, and you paste the answer back to add it.
Questions
What is the difference between an output and an outcome?
An output is something you produce, such as a launched page or a held meeting. An outcome is the change it should cause, such as more trials turning into paying customers. Good key results measure outcomes.
Can a key result be a milestone?
Sometimes. A milestone such as "automated vendor payments go live" is fine when there is no sensible number yet. Keep it to a minority, and pair it with at least one key result you can measure every week.
How many key results should each objective have?
Two to five, with three as a good default. One key result is easy to game; more than five spreads attention too thin to track.