For teams already running OKRs
Why do OKRs fail?
The short answer
OKRs usually fail for one of three reasons: the key results are tasks rather than outcomes, nobody checks whether the team has room to carry the work, or nothing keeps the goals alive between quarterly meetings. The framework is rarely the problem. The practice around it almost always is.
Key takeaways
- If you can tick it off, it isn’t a key result. Key results are numbers that move.
- Every goal needs a removal line: what comes off someone’s plate, and who agreed.
- Goals die between quarterly meetings. A ninety-second weekly check-in keeps them alive.
- The popular “70% of OKRs fail” figure has no reliable source. Diagnose your own cause instead.
What percentage of OKRs fail?
Nobody knows, and it is worth being suspicious of anyone who says they do. The figures that circulate (60%, 70%, even 90%) come from software vendors’ own surveys, or they recycle an older claim that 70% of all change programmes fail.
That older claim is usually credited to Harvard Business Review. In 2011 the researcher Mark Hughes traced it through its five best-known sources, one of them in HBR, and found no valid empirical evidence behind any of them (Journal of Change Management). A number repeated often enough starts to sound like research.
What we can offer instead is what we have seen. Our case log holds 47 engagements since 2015, 27 of them with family- or founder-led companies. In 15 of the 47, the thing we left behind was the same: a short list of priorities, an owner for each, and a blueprint for putting them into practice.
That is the pattern this article is about. A goal on its own changes nothing. Someone has to own it, and something has to carry it through an ordinary week. So the useful question isn’t how often OKRs fail. It is which of three failures is happening in yours.
How OKRs die, week by week
The pattern repeats with a consistency that stops being funny after the fortieth room. A leadership team spends a day writing goals, and everybody leaves energised.
Around week three the first update lands, and most of it describes activity. By week five a busy month has arrived and the goals have stopped being mentioned. Nobody cancelled them. There was no meeting where they were called off.
By week eight, “how is the quarter going?” gets the answer “fine, I think.” At the week-twelve review nobody can say with confidence whether the goals were met, and the honest answer is that they stopped mattering somewhere around week five.
In our experience, OKR programmes break in one of three places. They are worth taking one at a time, because each has a different fix, and fixing the wrong one wastes a quarter.
Root cause 1: The key results don’t measure real change
If you can tick it off, it isn’t a key result. A key result is a number that shows whether something changed. A task is something you do. Most failing OKR sets are task lists with better formatting.
A definition first, in case the vocabulary is new: an objective is a sentence saying what you are trying to change this quarter. Key results are the two or three numbers underneath it that prove whether it moved.
“Launch the new website” is a task. You can finish it on a Tuesday and be no better off. “Increase signup conversion from 3.2% to 5%” is a key result: it either moved or it didn’t, and the number says which.
In most leadership goal documents I read, the majority of the lines are tasks. Not some. Most. The people who wrote them are experienced operators, and they would spot the same mistake instantly in someone else’s document.
Four rewrites
| Team | Written as a task | Rewritten as a key result |
|---|---|---|
| Marketing | Launch the new website | Signup conversion from 3.2% to 5% |
| Sales | Launch a partner referral programme | Deals sourced through partners from 2 to 10 a quarter |
| Operations | Roll out order-tracking software | Orders delivered on the promised date from 78% to 92% |
| People | Train every team lead in coaching | First-year attrition from 24% to 15% |
The tasks don’t disappear. They become the work underneath: the bets you make to move the number. If a bet doesn’t move it, you change the bet, not the key result.
The same cause has three quieter forms. The KPI dashboard copied in as key results (“revenue: ₹12 crore”) describes a number you watch, not one you are changing. Business-as-usual dressed up as an objective describes running, not changing. And targets tied to bonuses get set to be safe rather than true, because nobody commits to a hundred when they will be rated on it.
Root cause 2: Nobody asked the people carrying it
Most OKR programmes add goals to teams that are already full. Every template asks what you want to achieve. Almost none asks who will do the work, and what comes off their plate so they can.
To be fair to the canon, focus is not a new idea: Andy Grove and John Doerr both wrote at length about saying no. But no OKR tool or practice actually enforces it. Every platform will let you commit to twelve objectives and then add a thirteenth. None of them asks what was removed.
Over-capacity rarely looks like failure, which is why it goes unnoticed. It looks like a quarter where everything was delivered and two good people resigned. It looks like quality drifting in a way nobody can attribute, or a head of function who used to raise problems early and no longer does.
Growth Without Burnout starts by asking who will carry the goal. Capacity belongs inside the goal, not beside it.
The removal line
The fix is one sentence per goal: what comes off someone’s plate, and who agreed to it. It is not a time-tracking exercise. Hours-based versions are invented precision, and teams see through them. One line, in plain words, with a name against it.
Example
The four outlet managers carry this goal. They cannot take custom orders and run a busy café floor at the same time. What comes off: they stop re-typing enquiries into the order book. Agreed by the founder.
Asking also changes how the goal is held. People who helped choose the strategy behind a measure are less likely to chase the number at the expense of what it stands for (Choi, Hecht & Tayler, 2013). Goals written for a team and handed down skip that step, which is why cascaded OKRs so often get gamed.
From our case log
Sometimes nothing can come off. A private-equity-backed consumer wellness business asked us what systems it needed to support its growth ambitions. The leadership team agreed on those systems, but part of the answer wasn’t a system at all: the business hired a new COO.
When nothing can come off, the honest options are to hire, to automate the work, or to set fewer goals. All three beat pretending.
Root cause 3: Nothing happens between the quarterly meetings
A goal that is looked at once a quarter is not being managed. It is being remembered. The literature calls this set and forget, and it is the cheapest failure to fix: a check-in of about ninety seconds, every week.
The check-in has two parts. For each key result: the current number, and the owner’s honest read on whether it will still land. Most companies skip the second part, and it is the valuable half. The number is history. The owner’s read is a forecast, and it moves first.
A key result can sit comfortably on plan in week four while the person carrying it already knows it won’t land. They know about the supplier, the hire that fell through, or that the last three weeks came from a one-off. The number won’t show any of that for another month. So the read turns amber while the number still looks green, and founders who watch only the number find out in month three what they could have known in week five.
Make it survive a bad week
- Make it absurdly short. Anything over five minutes gets cancelled when the week goes wrong, which is exactly the week you needed it.
- Give it a fixed slot. Same morning, every week. No agenda, no invitation, no decision about whether to hold it.
- Keep the founder in it. Goal practices survive when the founder visibly does the ritual, and die when it is delegated in month two.
- Let the calendar remember. A recurring entry keeps working when everything else is on fire.
We run this as a weekly check-in, one of our OKR Rituals: a score from 1 to 10 for each key result and one sentence on what changed. It is one of the simplest Workplace Rituals: a small, repeated practice that keeps a goal alive without wearing people out. A missed week is not the failure. Deciding that a missed week means the practice is over is.
The 10 usual reasons OKRs fail, traced to three causes
Most lists give you ten reasons; some give forty. They are real, but they are symptoms, and each traces back to one of the three causes above. Fix the cause and the symptom goes with it.
| Commonly cited reason | Root cause | The fix |
|---|---|---|
| Key results measure activity, not outcomes | 1. No real change | Every key result is a number that moves |
| The KPI dashboard copied in as OKRs | 1. No real change | Watch KPIs; set OKRs only for what you are changing |
| Business-as-usual dressed up as objectives | 1. No real change | Ask of every line: is this changing, or running? |
| OKRs tied to bonuses and ratings | 1. No real change | Keep goals and appraisals as separate conversations |
| Too many objectives | 2. Nobody asked | Two or three for the company; count the load per person |
| Goals cascaded from the top | 2. Nobody asked | Write them with the people who carry them |
| No single owner | 2. Nobody asked | One named person per key result, never a department |
| No check-ins between reviews | 3. Nothing in between | A ninety-second weekly check-in |
| Leaders stop showing up after month two | 3. Nothing in between | The founder does the ritual, visibly |
| A tool everyone updates and nobody discusses | 3. Nothing in between | Decide when the conversation happens, not just where |
Which one is killing your OKRs? A five-minute check
Check the three causes in order. A weekly check-in on key results that are really tasks only tracks the wrong thing more often, and a goal nobody has room for fails however closely you watch it.
Answer yes or no to each question. Two or more yeses in a group tell you where to start.
Real change
- Could you tick off any of your key results like a to-do item?
- Is any key result a number you already report every month anyway?
- Would the business look much the same if every key result were hit?
Capacity
- Does any key result belong to a team or department rather than a named person?
- Did this quarter’s goals go on without anything coming off anyone’s plate?
- Were the goals written for the people carrying them, rather than with them?
Rhythm
- Has a week gone by without anyone looking at the key results?
- Would the owners struggle to say today whether each one will land?
- Is the quarterly review the first place a missed goal gets discussed?
If you would rather have it read properly, scored and with evidence, that is what the OKR Audit is for.
How to restart OKRs that have already stalled
Don’t wait for next quarter. A half-quarter restart teaches you more than a perfect plan three months from now, and it takes about two weeks.
- Cut to one objective: the one the business would miss most if it quietly died.
- Rewrite its key results as numbers with a starting figure and a target. If a starting figure doesn’t exist yet, measuring it is the first week’s work, and it counts.
- Write the removal line: what comes off whose plate, and who agreed.
- Name one owner per key result: a person, not a department.
- Book the weekly check-in for next Monday and every Monday after: ninety seconds, three numbers, one honest read each.
From our case log
Stalled work is rarely rescued by rewriting the goal. A family-run, publicly listed manufacturer asked us why its rebrand wasn’t moving; it had been held up for fourteen months. The work was a diagnosis of the rebranding process.
Within three months of that diagnosis the rebrand was in implementation, and the new identity launched. When something important stalls, look at the practice around it before you touch the goal.
Frequently asked questions
Are OKRs worth it for a company under 100 people?
Yes, and arguably more than for a large one: a small company feels a wasted quarter immediately. What has to change is the volume. Two or three objectives for the whole company, not eighteen, and none of the cascading apparatus built for ten thousand people.
How long before OKRs start working?
Expect the first cycle to be rough and the third to be useful. Nearly every first-timer writes some tasks by accident and picks a number they cannot yet measure. That is normal, not failure.
Is it the OKR software’s fault?
Rarely. Software makes updating easy, but no tool decides when the conversation happens, who is in it, or what you do about what it says. That is a practice question, not a product one.
Should OKRs be tied to bonuses?
No. The moment a goal decides someone’s rating, they commit to what they are sure of hitting rather than what matters. Keep goal-setting and performance reviews as separate conversations, and grade committed and stretch goals separately.
How many OKRs should a team have?
Two or three objectives for the whole company, with about three key results under each. Count the load on each person, not the goals per level: nobody should own more than three key results across every level combined.
What is the difference between an OKR and a KPI?
A KPI is a number you watch continuously to know the business is healthy. An OKR is a number you are deliberately trying to move this quarter. The same metric can be both at different times: a KPI all year, and an OKR for the quarter you decide to change it.
What is the single highest-value change?
Add a weekly check-in of ninety seconds. Most goal systems die between the quarterly meetings, and that is the cheapest possible intervention.
Sources
- Hughes, M. (2011). Do 70 per cent of all organizational change initiatives really fail? Journal of Change Management, 11(4), 451–464.
- Choi, J. W., Hecht, G. W. & Tayler, W. B. (2013). Strategy selection, surrogation, and strategic performance measurement systems. Journal of Accounting Research, 51(1), 105–133.
- Grove, A. S. (1983). High Output Management. Random House.
- Doerr, J. (2018). Measure What Matters. Portfolio.
