Home/Blog/OKRs/How to Write Key Results: Evidence the W…
OKRs

How to Write Key Results: Evidence the World Changed

How to write key results that prove a change happened, not that work was done: start from the outcome, then pick the number, and run two tests on each draft.

Cover of the FIKR Space guide: How to Write Key Results: Evidence the World Changed

To write key results, describe the change you want in one plain sentence first, then choose the number that would show it happened, and make sure the result is a measurable state rather than a piece of work. A good key result is evidence that the world changed.

Most bad key results are not badly worded. They are the wrong kind of thing. This article walks through the order of work that avoids that, the two tests that catch the rest, and the habits that keep a key result honest once the quarter is under way.

How to write key results: start from the outcome, not the metric

The sequence matters more than the phrasing. Begin by describing, in a sentence of plain language, what would be genuinely different in three months if this went well. Only after that do you ask which number would demonstrate it.

Teams that do it the other way round, starting from a metric because one is available or easy to measure, reliably end up with goals that are precise and pointless. Availability is not relevance. The most measurable thing in your business is rarely the most important one, and a dashboard that already exists is a poor guide to what you should be trying to change.

So the first draft of a key result is not a number at all. It is a sentence about a different state of affairs. The number comes second, and its job is to be the smallest convincing proof of the sentence.

A key result is a state, an objective is a direction

It helps to keep the two layers apart. An objective describes where you are trying to get to and why it matters, in language a new joiner could follow without context. It is qualitative and memorable, and slightly uncomfortable to commit to.

A key result sits underneath it and says how you will know you are getting there. It must be a measurable state. It must not be an activity. If your objective could be completed by ticking items off a list, it is a work plan in the clothes of a goal, and the same suspicion applies to any key result you write beneath it.

The distinction is the whole method, and it is the part most teams get wrong. The broader reasons it matters are in what are OKRs.

Free guide, 32 pages: The OKR Playbook. Send me the guide.

Activity versus state: a worked contrast

The guide gives a clean contrast. Publish twelve articles is an activity. You can do it perfectly and change nothing. Grow organic signups from forty to one hundred and twenty a month is a changed state. The twelve articles might be how you attempt it, and if they are not working, you can do something else without failing the goal.

That last point is the practical advantage. A key result written as a state leaves room for a better route than the one you imagined at the start. A key result written as an activity locks in the route, so the team can finish the route perfectly and arrive nowhere.

Written as an activityWritten as a stateWhat the state frees you to do
Publish twelve articlesGrow organic signups from forty to one hundred and twenty a monthSwap the articles for another method if they are not moving the number
Launch the new onboarding flowA measurable change in what new users actually do, with a starting value and a targetJudge the flow by its effect, not by its release date

The second row is built from the guide's disguised task example: a launch can be delivered in full while changing nothing, so the state you actually care about has to be named instead.

The two tests every draft must pass

Run both of these on every draft before it is agreed.

  1. The so-what test. If we achieved exactly this and nothing else, would the business be meaningfully different?
  2. The would-we-know test. Is there a specific number, with a baseline and a target, that we can read at the end of the quarter without an argument about interpretation?

The first test catches activity dressed up as progress. The second catches vagueness. A draft that passes the first but fails the second usually just needs a number and a starting point. A draft that fails the first needs to be rewritten from the outcome sentence.

The second test depends on a starting value, which is why it has its own article: how to set a key result baseline.

Four ways a key result goes wrong while it is still on the whiteboard

The guide names four patterns that are easy to spot early and expensive to find late.

  • The disguised task. A piece of work with a deadline attached. It can be delivered in full while changing nothing.
  • The baseline-free metric. A number with no starting point, so nobody can tell whether it is ambitious or already achieved.
  • The sandbagged target. A figure the team is confident of hitting. This is a rational reaction to consequences, so check your incentives rather than the team.
  • The unownable key result. One that no person or team can meaningfully influence. It produces reporting rather than effort, and it quietly teaches everyone that the goals are decorative.

How these four connect to the wider list of failures is the subject of OKR mistakes founders make.

What to do when your draft is a task

Most first drafts are tasks, and that is normal. The fix is not to polish the wording but to ask a question of the task: if we did this perfectly, what would be different afterwards? The answer to that question is the state you should be writing down.

Take the guide's own disguised task, launching a new onboarding flow. Delivered in full, it proves only that something shipped. So you keep the launch in your plan as a way of attempting the goal, and you write the key result about the thing the launch was supposed to alter. If the launch does not move that, you are free to try something else, because the goal never depended on the launch.

This is also why a good set of key results feels slightly less certain than a list of tasks. A task list tells you what will be done. A set of states tells you what you are betting will change, and a bet is supposed to carry some risk.

FIKR OS, strategy meets execution: OKRs, roadmaps and kanbans in one system. See FIKR OS.

Label how ambitious each one is

Once a key result is a state with a number, say what kind of goal it belongs to. The guide separates committed goals, which you expect to deliver in full and where anything less is a real problem, from aspirational ones, which are set beyond comfortable reach and where achieving most of it counts as a success.

Both are legitimate. Mixing them without saying which is which is not, because nobody can tell whether a partial result is good news or bad. Put the label next to the number when you write it, and agree what a good score looks like for each type before the quarter starts.

Make it measurable in practice, not only in theory

A key result can pass both tests on paper and still be unusable on a Monday. The checklist in the guide asks four practical questions about measurability.

  • Does every key result have a baseline as well as a target?
  • Do you know today's value for every metric, or is finding it itself a key result?
  • Is the measurement source agreed and accessible before the quarter starts?
  • Would two people report the same number from the same data?

The last question is the one teams skip. If two people could read the same source and arrive at different figures, the end of the quarter will be spent on an argument about interpretation, which is exactly what the would-we-know test was meant to prevent.

Check that someone can actually move it

A well-formed key result still fails if nobody can influence it. Two more checks from the guide cover that.

  1. Give every key result a single named owner, a person rather than a department or a committee.
  2. Confirm that each owner can actually influence their number, and that no key result exists only to be reported.

The owner is not the only person who works on it. They are the person who knows its state, raises it when it is stuck and is accountable for saying so early. Ownership is about who speaks up, not about who does all the work. How that works across several teams is covered in how to align OKRs across teams.

Free guide, 32 pages: The OKR Playbook. Send me the guide.

A short routine for drafting

If you want something to follow, this order works for each key result.

  1. Write one sentence on what would be different in three months.
  2. Choose the number that would show it.
  3. Find today's value. If you cannot, make finding it the first key result.
  4. Set the target against that value.
  5. Run the so-what and would-we-know tests.
  6. Name one owner and confirm they agree out loud.

Write them slowly once. The writing is where good intentions turn into either a usable instrument or a document nobody reopens.

FIKR OS, strategy meets execution: OKRs, roadmaps and kanbans in one system. See FIKR OS.

Frequently asked questions

Can a key result ever be a piece of work?

The guide's position is that it should describe a changed state. A deadline-driven task can be delivered in full while changing nothing, so it belongs in the plan beneath the goal, not in the goal.

What if the most important outcome is hard to measure?

Start from the outcome anyway and then look for the number that would demonstrate it. The most measurable thing is rarely the most important, so do not let availability choose for you.

How many key results should an objective have?

No more than three. A longer list stops forcing choices, which is covered in how many OKRs a startup should have.

Who should write the key results?

Teams propose the key results they believe will move the company objectives, and leadership reconciles them in a conversation where the teams can genuinely push back.

The full method, with the checklists, is in The OKR Playbook, a free 32-page playbook.

LG

Luis Goncalves

// Founder & CEO at FIKR Space

Three-time founder. Built and exited Evolution4All before this. Now building FIKR Space — the operating infrastructure underneath every innovation ecosystem (startups, accelerators, governments, investors). Lisbon-based, works global.

Claim your founding member spot.

Every founding member is hand-reviewed. Tell us who you are and how you want to co-create — we get back inside 48 hours.

Role
Job Title
Country
Cookies

We use cookies.

Essential ones keep the site working. The rest only run if you say so. No dark patterns, no pre-ticked boxes.Privacy policy

Co-create with us. It's about building it together.Co-Create With Us