How to write OKRs that measure outcomes, not tasks
Seven steps to OKRs where key results measure change for users and the work sits in a separate list of initiatives.

Many quarterly OKRs read like a to-do list with percentages attached. This guide is for product managers writing their team's OKRs for next quarter. At the end you will have objectives with key results that measure change for users, and a separate list of the work that should get you there.
In brief
- An objective says where you want to be. It is qualitative, clear and time-bound.
- Each objective gets two to four key results. Each one is a measurable outcome, written from a baseline to a target.
- Tasks and launches are initiatives. They go in their own list.
- The test: if you could tick it done without anything changing for users, it is a task.

Three words, used strictly
Most bad OKRs come from mixing up three things. Keep them apart and the rest of this guide is easy.
| Term | What it is |
|---|---|
| Objective | Qualitative, clear, time-bound. It says where you want to be. |
| Key result | A measurable outcome. Two to four per objective, each written from a baseline to a target. |
| Initiative | A task or a launch. The work you do to move the key results. |
If you could tick it done without anything changing for users, it is a task.
The test for every key resultStart from the problem, not the plan
Before you write a single line, write down the problem in plain words. Who is stuck, and where? What is worse for them than it should be?
If you start from the plan, the plan turns into the objective. You end up with "launch the new thing", and the quarter is judged on whether you launched it. Starting from the problem keeps the question open: what should be different for users when the quarter ends?
Use whatever evidence you already have, such as support conversations, interviews or usage data. One or two sentences is enough.
Write the objective
The objective describes a place, not an activity. It is qualitative, so it carries no number. It is clear, so a colleague can read it and say what would be different for users. It is time-bound, so it belongs to this quarter.
Watch the verbs. "Launch", "ship" and "build" describe work. "Reach", "stop losing" and "finish" describe a state users are in. Take the problem from step 1 and turn it around: where do you want users to be?
For each key result, ask "how would we know?"
Read the objective back and ask: if we got there, how would we know? Every honest answer is a candidate key result. Write down more than you need.
Some answers will be about what you did: "we would have shipped it", "the articles would be live". Those are not evidence that anything changed for users. Set them aside for step 6. Keep the answers that describe what users do, or how much they struggle, and cut down to two to four per objective.
Write the baseline and the target
Write each key result as a movement: from where it is now to where you want it to be. "From 35% to 55%" tells the team more than "improve activation" ever could.
The baseline matters as much as the target. Without it, the target is a guess, and at the end of the quarter nobody can say how far you moved. If you do not know the baseline, finding it is your first piece of work. Put that on the initiatives list.
Pick a target you would be glad to see, and be able to explain why that number and not another.

Check the team can move it within the quarter
A good measure is no use if your team cannot affect it before the quarter closes. For each key result, ask two things. Is there work this team could do that would plausibly move it? Would the number show the change in time?
If the answer to either is no, look for a measure closer to the behaviour you are trying to change. You are not lowering the bar. You are choosing a number the team can actually act on.
List the initiatives separately
Now the work gets its place. Everything you set aside in step 3 goes here, along with any measuring you need to do. Tasks and launches are initiatives, and they are useful. They are just not results.
Next to each initiative, name the key result it is meant to move. If an initiative moves none of them, ask why it is on the list. If a key result has no initiative behind it, you have a number and no plan. You can keep the objective, key results and initiatives together in PikoPM's OKRs tool.
Score each key result at the end of the quarter
When the quarter ends, score the key results, not the initiatives. Look at the baseline, the target and where the number landed.
Shipping something and seeing no movement is a result too. It tells you the plan was wrong, or the problem was. A simple way to score is the share of the distance from baseline to target. In the example below, if KR1 ends the quarter at 50%, the team covered 15 of the 20 points between 35% and 55%: a score of 0.75.
A worked example
This example is made up for the guide. It is not from a real team.
Here is the same quarter, written first as a plan and then as an outcome.
Bad
Objective: Launch new onboarding.
- Ship onboarding v2 by March
- Write five help articles
Both can be done while nothing changes for users. Both fail the test.
Good
Objective: New teams reach their first win in week one.
- KR1: share of new teams that create a project in their first 7 days, from 35% to 55%
- KR2: median time to first project, from 3 days to 1 day
- KR3: onboarding support tickets per 100 sign-ups, from 12 to 6
Initiatives: onboarding v2, the five help articles.
Look at what changed. The objective is now a place: new teams reaching a first win. Each key result has a baseline and a target, and each can only move if new users behave differently or struggle less. None of them can be ticked done by finishing a piece of work.
The old key results did not disappear. Onboarding v2 and the five help articles are now initiatives. They are the team's bets on how to move the three numbers. If onboarding v2 ships and the numbers stay put, the team learns something real, and it can try a different bet next.
Before you commit
- The objective describes where users will be, not what the team will do.
- Each objective has two to four key results.
- Every key result runs from a baseline to a target.
- No key result can be ticked done without something changing for users.
- The team can move each number within the quarter.
- Tasks and launches sit in a separate initiatives list, each tied to a key result.
- You know how you will score each key result when the quarter ends.


