10. Correct with Compassion, Reward with Fairness¶
Development assumes people are reaching for the standard. This chapter is the other case — when people miss the standard, how correction stays fair without becoming punishment.
The resume landed on my desk and held my attention longer than it should have.
A Sales Manager at a large state-owned company, applying for a junior Product Manager role. The salary on offer was a fraction of what he was earning. The seniority drop was real. None of the numbers made sense on paper.
When I asked him why, the answer disarmed me.
"I chased the title. I am tired of the politics and the constant bowing to people. I just want to build things."
I respected that. Or I thought I did.
I hired him. I gave him a project.
Early on, he made a few sharp calls. Intuitive ones. The kind that suggest a person can see around corners. I noticed and I showed it. Extra leave when he asked. Active mentorship. Inside the room and outside it, he was openly the favorite of the department.
I thought I was reinforcing excellence. I was rewarding a honeymoon phase.
The mask slipped on the first medium-sized project. A five-week sprint. By the end of week one, nothing had moved.
I asked a Senior PM to support him. Her report was direct. He was skipping grooming sessions. When he showed up, he dragged the discussion in circles and pushed every hard call back to her. I checked the session recordings. The developers were frustrated. The PM I thought I had hired was ghosting the team he was supposed to lead.
I sat him down. I expected ownership. What I got was a defense I had not seen coming.
He blamed me.
"You always seemed satisfied. You talk to me every morning in a good mood. If I were doing something wrong, I assumed you would be angry. Since you were nice, I figured I was doing fine."
It took me a while to absorb that line.
He was not lying. He was telling me, with surprising honesty, the only signal he had been calibrating to. The boss's mood. Not the work's impact. Not the team's output. Not the deliverable.
The boss's mood.
He was not working to the requirements of the job. He was working to my mood. And my mood had been telling him everything was fine.
After the first warning sign, I did not treat the conclusion as final. I made the expectation explicit: he needed to show ownership of the delivery flow, clearer follow-through with stakeholders and developers, and better discipline in reporting the real product status instead of waiting for someone else to interpret it.
The Senior PM support was not meant to rescue the role for him. It was meant to separate two questions: was he failing because the system had not given him enough context, or was he failing because the role required judgment he was not showing yet?
So the probation extension had a window. The measure was not whether he looked busy. It was whether he could own the work with less correction, surface problems earlier, and create enough confidence that the team did not have to manage around him.
That did not happen.
I extended his probation. Moved him to a different team for a fresh start. The pattern held. The sessions were still skipped. The output never came. At the end of the extended period, I let him go.
I am not telling this story to make him the villain.
He was not lazy. He was not weak. He had a leader who had given him a prize before the race was finished, and once the prize was in his hand, there was no reason to keep running. The information failure was mine. The reward had outrun the evidence, and the only person in the room with the authority to fix that was me.
What bad leadership usually does¶
There are two defaults here. They look opposite. They share the same root.
The first is correction in public, aimed at the person.
Most leaders confuse "giving hard feedback" with "not holding back." They say what they think, in front of whoever is in the room, and call it directness. The result is not a corrected behavior. It is a defended person. After enough rounds of this, the person stops hearing the feedback at all. They hear a verdict on their character, and they put on armor before the sentence is finished.
The second is generosity as strategy.
Many leaders believe rewarding early builds loyalty and raises the ceiling. The instinct is not wrong. The execution breaks when the reward is not anchored to anything concrete. "You are doing great" without naming what great looks like is not feedback. It is ambient approval. The person calibrates to mood, because mood is the only signal on offer.
Both defaults skip the same step. Both refuse to name impact.
Public correction skips impact and goes straight to character. Ambient reward skips impact and goes straight to warmth. Either way, the person never finds out what their behavior actually caused.
The principle¶
Correct with compassion. Reward with fairness.
Be kind, but not weak. Be fair, but not indulgent. Aim correction at behavior, not identity. Aim reward at impact, not energy.
Why it matters¶
When correction targets identity, the person learns to defend themselves, not improve. Years of that and the armor becomes the personality. By the time a better leader arrives, the damage is already structural.
When reward is not anchored to impact, the person learns to read the leader's face instead of the work. The work drifts. The face stays friendly. By the time the truth surfaces, the gap is too wide to coach across in one conversation.
In both cases, the leader thinks they have been giving feedback. The team has been receiving something else.
What better leadership looks like¶
The other story I carry from this principle is shorter, and it is on the correction side.
There was a developer on a team I took over. He had a reputation. Not a bad person — someone who had spent years being corrected badly. His previous manager had a style: critique in meetings, in front of the team. Public. Repeated. Pointed. Not cruel in intent, but cruel in effect.
I do not say that to make the previous manager a villain either. The pattern was the problem, not the man. Public correction aimed at character is a structural error, not a personality flaw. He was running the same default a lot of leaders run.
The developer had learned to associate feedback with shame. He became defensive before the correction was finished. He had been trained to protect his identity, not improve his work.
My first correction landed wrong too. The same armor snapped on. I did not push through it. I came back later, privately, and tried a different shape.
Here is what I saw. Here is what it caused. Here is what I need instead.
Not: here is what kind of person you are.
It took time. The defensiveness did not vanish. It was a learned pattern from years of damage, not weeks. But the direction changed. He started hearing feedback as information about his work, not as a verdict on his worth.
The lesson I took: the correction had been aimed at the wrong target for years. Once the target shifted from who he was to what he did, he could receive it.
That is what better looks like on the correction side. Private when personal. Specific when behavioral. Anchored to impact. Followed by a path to improve. No verdict on the person.
On the reward side, better looks like the same discipline pointed the other direction. Name what they did. Name what it caused. Then, if the impact deserves it, reward it. The warmth is fine. The warmth is welcome. The warmth just cannot be the whole signal.
The tool — Feedback Format¶
The toolkit carries a format for this. It is the named tool for this chapter.
Feedback Format
Use this for any correction or reward conversation that matters. Walk it in order. Do not skip the second field.
- Observation. What specifically did you see? Behavior, not interpretation.
- Impact. What did that behavior cause? Concrete. Named. Visible.
- Expectation. What does the standard look like in this situation?
- Root cause. Skill, clarity, workload, system, ownership, or something else?
- Next step. What changes, by when, owned by whom?
- Support. What do they need from you to make the change real?
- Follow-up. When will you check in on whether the change held?
The load-bearing field is Impact.
Without Impact, correction collapses into a verdict on the person. Without Impact, reward collapses into ambient approval. Impact is what keeps the conversation about the work instead of about the person's character or the leader's mood.
What would Impact have changed in the two stories?
For the developer: his previous manager could have separated behavior from identity simply by passing every correction through Impact first. This decision in the review caused this downstream rework is a different sentence than you are careless. The first is a problem to solve. The second is a person to defend.
For the PM: I should have named impact early and named it clearly. Your call on this requirement saved the team a week of building the wrong thing — that is what good looks like here. That sentence carries information. You are doing great does not. If he had been told concretely what performance looked like, my mood would have stopped being his primary signal, because there would have been a better one.
The tool fixes both failure modes by forcing the conversation to pass through a concrete consequence before it moves to anything else.
A note on the ethical fences¶
Correct with compassion does not mean correct gently so people feel good. The correction in the developer story was not soft. It was private, specific, and direct. It was compassionate because it aimed at the behavior, not the person — not because it avoided discomfort. Hard feedback can be compassionate. Public humiliation is not directness.
Reward with fairness does not mean withhold reward to maintain leverage. The lesson from the PM story is not "be stingy with praise." The problem was not that I was too generous. The problem was that I was generous without information. The fix is to name impact, not to reduce warmth.
The PM was let go. That is the consequence in the story, and I am not going to soften it. But the chapter is harder on me than on him. He failed inside conditions I created. I rewarded him before there was anything real to reward, and once the reward was received, the standard was already gone. That is not a story about a weak hire. That is a story about a leader who never made impact visible enough to function as a standard.
The point is not control. The point is responsibility.
Simple rules¶
- Correct behavior, not identity. The target is what they did, not who they are.
- Private when the correction is personal. Specific when the correction is behavioral.
- Name impact before anything else. If you cannot name what the behavior caused, you are not ready to give the feedback yet.
- Reward impact, not energy. Warmth is welcome, but it cannot be the whole signal.
- Do not give the prize before the race is finished. Early reward without anchor becomes a free pass.
- Every correction needs a path to improve. Every reward needs a named reason.
- Read the room you have built. If the person is calibrating to your mood, your mood has been doing the talking.
Reflection questions¶
- For the last correction I gave — was the target the behavior or the person?
- For the last reward I gave — can I name the specific impact it was for?
- Am I confusing my warmth with their feedback?
- Is this person calibrating to the work, or to my face?
- Did I give the prize before the race was finished?
- When I correct, is there a path to improve, or only a verdict?
- If someone replayed my last feedback conversation, would they hear an Observation and an Impact, or would they hear a mood?
Reminder¶
Correct with respect. Reward with fairness. Never manipulate people into giving more.
If correction lands on identity, the person learns to defend themselves, not improve. If reward lands on mood, the person learns to read your face, not the work.
Aim at behavior. Anchor on impact.