How Many OKRs Should a Startup Have? Fewer Than You Want
How many OKRs should a startup have? No more than three objectives with three key results each, and usually less, plus a written list of what you are not doing.
How many? A startup should have no more than three objectives, each with no more than three key results, and most groups would do better with fewer. If choosing what to leave off the list did not feel uncomfortable, you made an inventory rather than a set of priorities.
The guide treats the size of the set as the single strongest predictor of whether the method works. This article explains why a short list works, what the discomfort is for, and how to write down the things you are deliberately not doing, which is the part almost everybody skips.
How many OKRs should a startup have? The ceiling and the reason
The ceiling is three objectives with three key results each, for a company or for a single team. That makes nine key results at the very most, and the guide is explicit that most groups would be better served by less.
The reason is not that more is impossible to track. It is that more cannot do the job a goal exists to do. A goal is a decision, made in advance, about what wins when two things compete. A set that contains everything worth doing wins nothing, because every item on the list can claim to be the priority, so the list cannot settle a single conflict when the conflict arrives.
That is why the guide calls this feeling wrong in the right way. Everything on the longer list is genuinely worth doing. That is exactly the problem.
The discomfort is the prioritisation
The guide makes a point that is easy to read past: the discomfort of leaving good things off the list is the prioritisation. It is not a side effect to be minimised. It is the evidence that a choice was made.
You can use that as a test. When you finish drafting, ask whether leaving things out hurt. If the answer is no, you did not choose between anything. You copied your existing intentions onto a page. The checklist for the start of a cycle includes exactly this box: leaving things off the list was uncomfortable.
| What the set looks like | What it tells you |
|---|---|
| Everything worth doing is on it | No choice was made, so no conflict can be settled by it |
| Some good ideas were painfully left off | A real choice was made and the set has something to say |
| Nothing on it ever causes a no | It describes work that was going to happen anyway |
The third row connects back to the central claim of the whole method, which is set out in what are OKRs.
What a startup has to decide before counting
Before you can say how many, you need to know what counts as one. The guide separates an objective from a key result, and the count applies at each level.
- Objectives are directions. At most three at company level.
- Key results are evidence of change. At most three under each objective.
- Tasks are neither. They belong in your plan and do not count towards the set, which is a quiet relief for anyone worried that a small number means a small amount of work.
That last point matters. A short list of goals does not mean a short list of activity. It means a short list of things you are betting will change. Plenty of work happens beneath a goal, and the number of key results says nothing about how busy the team is. Writing each result as a changed state, rather than as work, is covered in how to write key results.
Write down what you are explicitly not doing
The guide describes a short list of the significant things deliberately left out of this quarter as one of the highest-value paragraphs in the whole document. Almost nobody writes it.
It does three jobs.
- It turns the goals from a statement of intent into a record of a decision.
- It prevents the same debate from returning every month.
- It gives everyone a legitimate way to raise something later without reopening the whole plan.
That third job is the most useful in a small company where requests arrive constantly. Instead of an argument, a person can say that the thing was on the not-now list and ask whether anything has changed. The question is short, it is fair, and it gets an answer on the merits.
How the not-now list protects a small team
In a startup, the person asking for something is often the founder, a customer-facing colleague or someone else with real authority. Saying no on authority alone spends goodwill, and the guide notes that it teaches people to route around you.
A written list of what you chose not to do changes that exchange. You are not refusing a person. You are pointing to a decision the whole team can read, and inviting them to say what has changed since. The reasoning behind that approach is also described in the guide's account of saying no with a reason instead of with your authority.
A count check you can run in ten minutes
If you already have a draft set, you can run a quick check against the guide's own checklist.
- Are there no more than three objectives at company level?
- Does each objective have no more than three key results?
- Was leaving things off the list uncomfortable?
- Is there a written list of significant things you are explicitly not doing this quarter?
If any answer is no, cut before you add. The usual instinct is to find a way to keep everything by merging items or by writing broader objectives. Resist that, because a broad objective is a way of hiding a missing decision. The aim is a set small enough that a single owner can name each result and say how it is going. Single ownership is covered in how to align OKRs across teams.
Why a short list makes the weekly check possible
There is a practical reason as well as a philosophical one. The guide asks for a weekly check of fifteen minutes in which each key result gets its current number, the owner's confidence and what is in the way. A set of three objectives and no more than nine key results fits inside that time. A sprawling set does not, and the meeting either overruns or turns into a skim that nobody values.
So the size of the set and the rhythm that runs it are linked. Keep the set small and the check stays short and regular, which is where its value comes from. The format of that meeting is in how to run a weekly OKR check in.
Why a longer list feels safer
A long set of goals feels like insurance. If something important is on the page, nobody can say it was forgotten. If it is off the page, someone might feel overlooked, and you will have to explain why.
That feeling is understandable and it is exactly what the discipline asks you to override. The cost of the long list is paid later and in smaller instalments: a conflict in the middle of the quarter that the list cannot settle, a weekly meeting that skims, an owner who can name nine things but drive none. The cost of the short list is paid once, at the drafting table, in the form of discomfort. Most people, shown both bills, prefer to pay the second.
There is also a quieter benefit. When the set is short, each result is visible to everyone, so people can see what the company is actually trying to change. A person who knows the three things that matter this quarter makes better independent decisions and asks for less permission, which is the alignment effect the guide describes.
Company sets and team sets
The ceiling applies at the company level and to a team on its own. A team that looks at the company objectives should pick what it is best placed to contribute, and the guide says plainly that the honest answer is often that it contributes to only one of them. That keeps each team's set short too, without any need to cover every company objective.
So the total number of results across an organisation is not three times three times the number of teams. Teams do not mirror each other's lists. They choose, and a team that serves one objective well is doing its job.
What if everything really is important?
Founders often push back here, and the guide's answer is consistent. Everything on the longer list being worth doing is not an argument for including it. It is the definition of the problem. Capacity is limited, so choices are going to be made anyway. The only question is whether you make them deliberately now or let recency and volume make them later, one interruption at a time.
If some item is truly essential, such as a deadline you cannot miss, the guide has a place for it: label it committed, so that the team knows anything less than full delivery is a real problem. That is a way of being clear about importance, not a licence to add more goals.
Frequently asked questions
Is three objectives a target or a limit?
It is a ceiling. The guide says most groups would be better served by less, so treat three as the most you should ever carry, not the number to aim for.
Do small teams need fewer than a larger company?
The guide does not give a separate figure for small teams. The principle is the same at any size: the set has to be short enough that leaving things off hurts.
Where do the things we are not doing go?
On a short written list beside the goals, covering the significant items deliberately left out this quarter. It lets anyone ask whether something has changed without reopening the plan.
Can a team have its own set as well as the company set?
Yes, and the same ceiling applies. How a team writes its own results in service of the company objectives is covered in how to align OKRs across teams.
The full method, with the checklists, is in The OKR Playbook, a free 32-page playbook.