Should OKRs Be Tied to Compensation? The Real Damage Is How You Treat Overperformers

In a Q3 planning meeting a few weeks ago, one of my directors waited until the room had mostly cleared out, then asked, quietly, whether her team's target could come in about 15% lower than what we both knew they could hit. She wasn't being lazy. She'd learned the lesson two cycles earlier, the hard way, when her team blew past an aggressive number at 130% and found the follow-up conversation about her bonus harder, not easier, because of it.

Everyone blames comp-linked OKRs for sandbagging, and that's true as far as it goes. But the policy on paper isn't what actually teaches a team to protect its numbers. What teaches them is watching one team overperform and get punished for it — a manager quietly raising next quarter's bar because 130% "proved" the target was too easy, instead of proving the team executed well. Every other team watches that happen once and adjusts accordingly. The fix isn't better goal-writing. It's refusing to let overperformance become the informal penalty it usually is.

Here's what that mistake actually costs, and what I'd tell any CEO or department head about to make it — starting with the part almost nobody names.

The team that overperforms is the one most likely to get punished for it — protect them first

This is the part nobody warns you about going into a planning cycle. The moment a team hits 130% of a genuinely ambitious target, the instinctive leadership response is often to raise their bar sharply the following quarter, as if the overperformance proved the target was too easy rather than proving the team executed well. That instinct teaches sandbagging faster than any comp policy written down anywhere.

We caught ourselves doing exactly this with a team that beat a lead-gen target by 30 points one quarter. The reflexive move was to set next quarter's number 30 points higher to "match" them. What stopped us: asking whether we were rewarding the win or quietly penalizing it. This is close to the blind spot Radar was built to close — surfacing a team's actual trajectory on day three instead of day thirty, so a manager has real signal about why a team overperformed before reflexively ratcheting up next quarter's number as an informal punishment for being the one team that didn't sandbag. Once your best team learns that honesty gets rewarded with a harder target and a quieter bonus conversation, you've taught your whole organization how to lie more skillfully, and that lesson spreads faster than any planning memo written to prevent it.

The bonus conversation and the goal-setting conversation cannot happen in the same room

Before we separated the two, our quarterly planning meetings were, on paper, about setting ambitious targets. In practice, they were negotiations. A VP would propose a number, someone would ask what it meant for the team's bonus pool if they missed it by 10%, and the number would get quietly revised downward before anyone wrote it in the doc. There's a specific kind of absurdity in a room full of smart operators doing math on how to look ambitious without actually being ambitious.

The fix wasn't asking people to be braver. It was removing the reason to negotiate in the first place. Once comp moved to a separate review entirely, the same VP proposed a number 20% higher the very next cycle — not because she'd had a change of heart, but because the number no longer had to double as a safety net.

Sandbagging isn't a character flaw. It's a rational response to how the system pays out

I used to think of target-lowering as a motivation problem, something you fixed with a better pep talk before planning season. It isn't. Locke and Latham's goal-setting research names the actual mechanism: when a reward is contingent on hitting a specific, difficult target, and someone senses they might come up just short, their sense of self-efficacy drops on the near-miss, and their performance drops right along with it. Once a team learns that missing a stretch target by a little costs them a bonus, the rational move isn't trying harder next time. It's aiming lower, so the near-miss never happens again.

That's not cynicism about your people. It's what any reasonably smart operator does when the incentive structure rewards a lower target more reliably than it rewards an honest one. Blaming the individual for sandbagging is blaming the symptom. The system taught them to do it, usually in one bad cycle, and they only need to learn it once.

Separating the scoreboard from the paycheck doesn't make goals less serious. It's what keeps them honest

This is the part leadership teams get backwards. The worry is always that decoupling OKRs from comp will make people take goals less seriously. In practice, the opposite happens: goals get more ambitious, not less, because nobody's protecting their number anymore.

Once we separated the two, the real alignment work started happening somewhere else entirely — in how visible the goal cascade was, not in how much money rode on it. A CEO changes a target midstream now, and Goals & Alignment updates every downstream team's view of it the same day, automatically, with no re-communication tour and no ambiguity about whether the old number still counts. Nobody has to wonder if the new number is real, because the system shows it whether or not it affects anyone's bonus. Visibility ended up doing the trust-building work that comp incentives had been mistakenly assigned to do.

If you don't say the separation out loud, your team will assume the old rules still apply

Decoupling comp from OKRs in a policy doc does nothing if nobody hears it said in a room that matters. Teams default to whatever rules they've previously been burned by, and if your org has tied a bonus to a KR any time in the last two years, that memory doesn't disappear because someone quietly updated a wiki page.

Say it out loud, every single cycle, starting in the kickoff meeting: OKR scores measure ambition and progress, comp gets evaluated separately, on separate evidence. We made that separation literal in our own weekly operating review, where the OKR number and the story behind it show up as two distinct inputs rather than one performance verdict compressed into a single slide — which is exactly the structure Reviews is designed to hold. A team can report the real number in that room without it doubling as a verdict on the person who owns it.

I still think about that Q3 conversation with my director. I told her the target stays where it actually should be, not 15% lower, and that her bonus conversation this cycle would be about her judgment and execution, not about whether she'd hit an arbitrary number that had nothing to do with either. She looked surprised. That reaction told me more about what our old system had taught her than anything else could have.

If you're setting comp bands or planning your next OKR cycle right now, do one concrete thing before the kickoff meeting: write the separation down in plain language, and say it out loud to the room, not just in a doc nobody reads. Say it every cycle, until nobody has to guess whether the old rules still apply — and until the team that overperforms stops bracing for the follow-up conversation.

Try it free at rhythms.ai.

Frequently Asked Questions

Should OKR scores affect my team's bonuses?

No. The moment a bonus depends on hitting an OKR number, teams will rationally set lower targets to protect their comp, which defeats the purpose of OKRs as a stretch-goal system. Evaluate comp on a separate set of criteria — execution quality, judgment, impact — not the raw OKR score.

Why does my team keep setting OKR targets lower than what they can actually hit?

This is almost always a learned response, not a motivation problem. Someone on the team was likely burned in a prior cycle when hitting an aggressive number too well affected their bonus or review. Sandbagging becomes a rational habit the first time a team learns that lesson.

How do I separate goal-setting from performance evaluation without confusing my team?

Be explicit about it in writing and say it out loud in the room where planning happens: OKR scores measure ambition and progress, and they're discussed separately from the criteria used in comp and performance conversations. The separation has to be stated outright every cycle, or teams will assume the old comp-linked rules still apply by default.

What should OKRs be used for if not compensation decisions?

OKRs work best as a forum for surfacing what's actually true — what's on track, what's stuck, and why — so leadership can make better decisions faster. That only works if teams feel safe reporting an honest number, which comp linkage actively undermines.