OKRs vs MBOs
Management by Objectives, from Drucker in 1954, set objectives jointly between manager and employee, reviewed annually and usually tied to appraisal and pay. OKRs kept the joint objective-setting and changed three things: a quarterly cycle, company-wide transparency, and a deliberate separation between goal attainment and compensation.
| MBOs | OKRs | |
|---|---|---|
| Origin | Drucker, 1954 | Grove at Intel, then Doerr at Google |
| Cycle | Annual | Quarterly, checked weekly |
| Visibility | Manager and employee | Usually company-wide |
| Unit | The individual | The team |
| Tied to pay | Almost always | Deliberately not |
| What "good" means | Achieved the number | Moved the outcome, judged on evidence |
| Main failure mode | Sandbagged targets | Too many goals, written as tasks |
The one difference that matters most
Take away the cadence and the transparency and you still have two similar methods. Take away the link to pay and you have changed the physics.
The moment somebody’s bonus depends on a number they helped set, they will set a number they are confident they can hit. That isn't dishonesty, it's arithmetic, and it's the most predictable behaviour in management.
MBOs wire the number to the reward and then ask for ambition. OKRs separate them so that a stretch target can be missed honestly and still be the right thing to have attempted.
What MBOs got right
Joint objective-setting. Drucker’s core insight was that people commit to goals they helped write, and every functioning goal system since has kept it.
Clarity of ownership, too. MBOs are unambiguous about who is accountable, which is something OKR rollouts often blur in the name of team ownership until nobody is quite responsible for anything.
Where OKRs are genuinely better
Speed of correction. An annual objective is discovered to be wrong in month nine. A quarterly one with a weekly confidence check is discovered to be wrong in week three, while you can still do something about it.
And transparency. When everyone can see everyone’s goals, duplication surfaces, dependencies surface, and the argument about priorities happens in the open rather than in nine separate one-to-ones.
If you're moving from MBOs to OKRs
Change the cadence first and leave everything else alone for a quarter. Quarterly objectives with a weekly check will teach you more about your business than any amount of framework redesign.
Then unwire pay from the team goal number. This is the hard one politically and the one that decides whether the change is real. Keep a bonus, absolutely, but judge the individual half rather than computing it from a goal percentage.
Only then move from individual objectives to team ones. Doing it in the other order tends to produce goals with no owner at all.
Are OKRs just MBOs with a new name? +
No, though the family resemblance is real. The substantive changes are a quarterly rather than annual cycle, company-wide transparency, a team rather than individual unit, and the deliberate separation of goal attainment from compensation.
Why do OKRs separate goals from pay? +
Because the moment pay depends on a number somebody helped set, they will set a number they are sure they can hit. You get reliable forecasts and no ambition. Separating them lets a stretch target be missed honestly.
Can you still pay a bonus if you use OKRs? +
Yes, and most companies should. The rule is that no bonus attaches to a team’s goal number. Shared company goals can carry a number; the individual half should be a judgement made by somebody who was there.