Managers: You Are Setting OKRs You Are Supposed to Miss. Nobody Told Your Performance Review.
I found this mistake on my own website last week.
I had written that the OKR convention of aiming for 60 to 70 percent attainment came out of research. I named a person and implied a study. It reads well. It is also not true, and I had to go back and rewrite it.
The convention is real. Set your key results ambitiously enough that hitting all of them means you did not aim high enough, and treat somewhere around two thirds as a good quarter. That practice came out of Intel, went to Google, and got popularized from there. What it did not come out of is a study. Nobody ran a trial where one group aimed for 70 percent and the other aimed for 100 and the first group won. It is a practitioner norm that hardened into a citation, and it now circulates as though somebody proved it.
I am not writing this to be pedantic about sourcing. I am writing it because the misattribution hides the part that actually matters, and the part that actually matters will decide whether OKRs work on your team or quietly make things worse.
The condition nobody states out loud
The 70 percent norm has a precondition. It only functions if it is genuinely safe to miss.
Read that again, because it is doing more work than it looks. The whole design of an ambitious key result assumes that falling short is an expected outcome, not a bad one. That is the trade. You get bigger targets and more honest forecasting, and in exchange, missing is normal and nobody is punished for it.
So what happens when the second half of that trade does not exist?
Because in most companies it does not. Your compensation review does not have a field for “aimed high, landed at 68 percent, exactly as designed.” Your skip level does not distinguish between a stretch goal that landed at two thirds and a commitment that was blown. The promotion conversation certainly does not. In a lot of organizations, and in almost every small one I have worked with, a number under 100 percent shows up in exactly one way: as a miss.
Now put a manager inside that. You have told your team to set ambitious key results. Your team is not stupid. They can see that the same review cycle which nominally rewards ambition also records shortfalls as shortfalls. Within one quarter, sometimes within one conversation, everyone learns the actual rule.
And the actual rule is: set key results you can definitely hit.
What you get instead
This is the failure mode, and it is worse than not using OKRs at all.
You now have a team writing goals they have privately already achieved, or are confident they will. The numbers get sandbagged. The key results get worded so that ambiguity works in the author’s favor. Nobody says any of this out loud, because the language of ambition is still mandatory in the document.
You have rebuilt ordinary targets, added a quarterly ceremony on top, and taught your team that the stated system and the real system are different. That last part is the expensive one. It generalizes. A team that learns to game one process will apply the lesson to the next one you introduce.
Get George’s stories in your inbox
Join Medium for free to get updates from this writer.
Meanwhile you, the manager, are looking at a spreadsheet of green and concluding that goal setting is going well.
The two honest options
There are only two coherent positions here, and you have to actually pick one.
Option one: make missing safe, and prove it. Not with a statement of intent at the kickoff meeting. With a visible instance. The first time someone on your team lands at 65 percent on an ambitious key result, you say so in front of the team, you say it was the right target, and nothing bad happens to that person in any subsequent conversation about their performance. One demonstrated case does more than a year of encouragement. If you cannot promise that, because your own manager will read the number differently, then you do not control this condition and you should not pretend you do.
Option two: drop the ambition norm and be straight about it. Set goals you intend to hit. Call them commitments. Track them properly. This is not a lesser form of management, and it is a much better fit for a small team where one missed quarter has real consequences for cash or for a customer. You lose the stretch effect. You gain a system where the words on the page and the expectations in the room are the same thing, which is worth more than most managers think.
What you cannot do is take the vocabulary of option one while operating under the incentives of option two. That is the configuration that produces cynicism, and it is by far the most common one I see.
Where this leaves the frameworks
None of this is an argument against OKRs. It is an argument against importing a norm without checking whether its precondition holds in your company.
It is also a reminder that OKRs, KPIs and SMART goals are not three brands of the same thing competing for your loyalty. They answer different questions. A KPI is what must not get worse and has no end date. An OKR is what you are deliberately trying to change this quarter. SMART is a standard for how a goal is written, not a framework for deciding what to pursue. Most of the framework argument you have sat through was people answering different questions at each other. I laid the whole split out in OKRs vs KPIs vs SMART goals, including which one to start with depending on your situation, and the ambition convention is discussed there with the caveats intact.
If you want to know whether your current goals are already set up to fail before you change anything, we built a fifteen question assessment that covers exactly this: how many goals you are carrying, whether each one has a single owner, and whether anything is measurable enough to argue about.
And if the practical problem is that your goals live in a document nobody opens between quarters, the tooling question is smaller than most vendors want it to be. You need to set direction, break it down so people can see how their work connects, and see status without chasing anyone. Celorly does those three things and is aimed at companies up to 49 people, with a free solo plan if you want to run your own goals through it before putting it in front of your team.
The part I keep coming back to
I had the wrong sentence on my own site for months. Nobody emailed me about it. I found it while checking something else entirely.
That is how these things spread. A practice becomes a convention, the convention gets quoted with more authority than it earned, and then it gets imported into companies whose incentive structures make it actively harmful. Nobody involved is being dishonest. Everyone is repeating something reasonable.
So before your next planning cycle, ask the one question the convention depends on and never states.
If someone on this team lands at 65 percent on a goal we agreed was ambitious, what actually happens to them?
If you do not like your own answer, do not set stretch goals this quarter. Set real ones.
