How to Give Feedback to Your Boss, and When Not To

"I think some people on the team feel like you micromanage a bit."

Three things are wrong with that sentence and only one of them is the word micromanage. It is a verdict on who your manager is, so the only responses are agreeing and defending. It is sourced to people who never agreed to be sourced, so your manager spends the week working out who. And there is nothing in it they could do differently on Monday. That is how giving feedback to your boss usually sounds, and it is why nothing changes after it.

The other version is saying nothing for two years, then telling the truth in the exit interview where it is free and useless.

Worth being honest up front: telling your manager how they manage has the worst risk to reward ratio of any conversation at work. Sometimes it is still the right call.

The honest calculation, first

What you are risking. Your manager writes your review, assigns your work, and is the reference you need in two years. None of that changes because they said the door is open. Feedback from a report is remembered as an event, filed under your name, and even managers who take it well take it personally for a week.

What you might gain. Not a different manager. The realistic upside is one behavior changing on one thing that affects your work. Name that as the goal, because "she becomes a better leader" is not what one conversation delivers.

When the right answer is to say nothing

Not every manager can take it, and pretending otherwise is how people talk themselves into an expensive mistake. Say nothing, and start looking, if:

• You have watched them punish someone for it. Going cool toward a person who pushed back counts.
• The problem is character, not behavior. There is no usable sentence for someone being petty, or dishonest.
• You are already in a weak position. On a plan, in a disputed rating, a month from a layoff round. Feedback from someone whose standing is in question is not heard as feedback.
• Nothing would change if they agreed. Some of what your manager does is set a level above them.

That is not cowardice. Raising something is a real decision with a real price, and "be brave" is not a strategy. If the calculation says no, what you learned is how long you want to be here.

Read what they did with feedback, not what they say about it

Every manager says they want it. That sentence is evidence about their self-image. Look at the record instead.

• Has anyone given them critical feedback in front of you, and what happened afterwards?
• Corrected on a fact in a meeting, do they say "you're right, I had that wrong," or explain why they were nearly right?
• What happened to the last person who disagreed with them in public?

Two or three yeses means this is probably survivable, though never free. All nos is your answer.

Cheaper still, test it first. Give them one small piece of feedback about something low stakes and watch what comes back.

Ask permission first

The highest leverage move available, and it sounds like the softest thing here. Do not open with the feedback. Ask whether they want it, name the topic, and let them say yes.

That does three things. You stop being someone doing something to them and become someone they agreed to hear. They get a moment to prepare. And a vague deflection tells you where you stand at no cost.

Ask in a one on one. Never in a group, and never in writing first: writing gives them time to build a case, and leaves a document.

Wait for an actual yes. If what comes back is "sure, but make it quick," take the exit they handed you.

Make it about the work, not about them

Upward feedback fails when it becomes an assessment of a person. Nobody can act on "you are controlling," they can only argue with it. What they can act on is a specific thing that happened and what it cost.

One example, and a recent one. The urge is to prove the pattern with five, in case they call it a one off. But five examples is a case, and a case gets defended against. One is a problem you both look at.

The effect on the work, not on your feelings. "It made me feel undermined" is unfalsifiable, so it becomes an argument about intent. "I stopped sending drafts early, so you see things later" is a cost to the work, and the work is what they own.

Four moves, under a minute before you stop talking.

You have not accused them of anything. You described a sequence, a cost and a request, all checkable. It is harder to be defensive about a Tuesday than about a personality.

When they get defensive

Assume it. Defensiveness is the base case, not the failure case, and treating the first reaction as the verdict is how a survivable conversation goes bad.

Do not say it louder. Do not retreat to "forget it, it's not a big deal," which teaches them the feedback was never real. And do not stay to win: there is no winning here, and the only thing left to spend is goodwill.

Let them keep the frame, restate the request in one sentence, and end early.

Ending early is the move almost nobody makes, because leaving while they are still wrong feels like losing. You were not there to be agreed with.

Judge it on the next two weeks, not the next two minutes. Plenty of managers argue in the room and change quietly afterwards, and some take it warmly and change nothing.

Feedback that is actually a request

That script ends in a request. That is not a softening technique, it is often the entire thing that works.

Feedback is a verdict about the past: it needs your manager to accept they were wrong before anything moves. A request is about next week, and it can be granted without conceding anything. A manager who cannot do the first can often do the second.

So try converting it first.

• "You don't give me enough context" becomes "Before we start something, can you give me the why and who asked for it?"
• "You change priorities constantly" becomes "Can we set the top two on Monday, and if that changes midweek you tell me what comes off?"
• "You undermine me in front of the team" becomes "If you disagree with me in a review, can you flag it to me first?"

Nothing is lost. The behavior you wanted is the behavior you asked for. What is gone is the requirement that your manager agree they were failing, which was the expensive part. It is the same move that makes a workload conversation work: hand over something answerable rather than something to be litigated, which is how saying no to your boss runs too.

When the message should not come from you

Sometimes you are the wrong messenger, or the direct route costs more than it returns.

360 reviews

The best of the three, because it arrives framed as a process rather than as a person, and your manager is expected to sit and receive it. Write it the way you would say it, with one concrete example: generic comments get read as noise. And do not sandbag it, because a comment saying everything is going well becomes evidence next year.

Anonymous surveys

Right when the problem is a pattern several people see and none of you can raise safely. Wrong when you are the only person who could have written it, which is more often than people assume. Anonymity protects a name, not a sentence. Name a project that had two people on it and it is signed.

Skip-levels

The highest risk, and the one people reach for at the wrong moment. Use it for conduct that will not change, or when the direct route has failed. Do not use it to reopen a decision that went against you: if that is what this is, disagree and commit is the honest move, not a quiet appeal upstairs. Assume it gets back to your manager, and tell them you are going. Running a skip-level well is its own skill.

This does not work if your manager has already punished someone for doing it. Where the last person who said something honest lost the interesting projects, the answer is not a better sentence: it is a written record and a decision about how long you stay. Feedback upward is only safe in a place where it has been safe before.

Why this one is harder out loud

All of it is easy to agree with while reading. It is different at three on a Thursday, when the person who writes your review leans back, says "interesting, go on," and waits.

Most people do not fail on the content. They fail in two physical ways. The voice tilts up at the end, so the feedback arrives as a question and gets answered as one. And the silence after "then stop" gets filled with "but honestly, it's not a big deal."

Both are habits of the body, not of the plan, and the only thing that changes those is doing it under pressure. In your head, your manager thanks you for the courage. Out loud is where you find out whether the sentence survives.