Tell me about a time your feedback to a colleague did not land well.
answer
- the feedback and why it mattered
- how you knew it had not landed
- own the delivery, not the message
- the second attempt, specifically
- the habit you changed since
basics
~20 sProbes self-awareness and persistence. Answer with feedback that was right in substance but wrong in delivery, an honest diagnosis of what you got wrong, a second attempt you actually made, and what you changed in your habits afterward.
how to answer
5 beats- the feedback you gave and why it matteredSet it up briefly — the behavior, the stake, and enough context that the feedback sounds justified. Keep this short; the interviewer is not evaluating whether you were right, and a long defence of your position signals you have not moved on from it.
- how it landed, and how you found outName the reaction and the concrete signal that told you — silence, disengagement, agreement with no change. Say honestly how long it took you to notice. Do not soften the outcome; a misfire you describe as a minor misunderstanding gives the interviewer nothing to score.
- what you got wrongThis is the beat the question exists for, so give it real airtime. Name the specific mistake — venue, timing, wording, missing evidence, an audience — and do it without retracting the substance. Vague self-criticism reads as performance.
- the second attemptDescribe what you actually did differently and what you said. The strong version opens by owning your half rather than restating the point, then asks a question instead of delivering a verdict. Say what you learned in that conversation that you did not know before.
- where it ended and what you changed for goodClose with the outcome, honestly graded — fully repaired, partly, or not. Then name the habit you kept: a rule about venue, a practice of gathering evidence first, a check that the other person heard what you meant.
your answer
4 story prompts- Pick a time your feedback was ignored, rejected, or landed harder than you intended.
- Choose one where you went back a second time — a story ending at the failure has no signal.
- Name the specific thing you got wrong: venue, timing, wording, or missing evidence.
- Be ready to say what you still stand behind, so the repair is not a retraction.
draft and rehearse your own answer in a learn session
go deeper
This tests self-awareness and resilience rather than technique. Interviewers want to see whether you can separate a message that was right from a delivery that was wrong, take responsibility for the half that was yours, and go back a second time. A strong answer proves you repair rather than withdraw, and that the standard survived the misfire.
I had been mentoring a contributor on a mobile library I co-maintain, and I gave him a correct piece of feedback in the worst venue available. He had landed two changes to the image-cache eviction path without looking at the crash dashboard afterward, and the second one took crash-free sessions from 99.58% to 99.06% on the release track. I wrote it in the public pull-request thread — something like, we cannot keep merging changes and not watching what they do — and I used *we* while everyone reading knew exactly who I meant. He went quiet. No commits, no issue comments, five weeks. For the first two of those I told myself he was busy. Then I admitted that I had made a fair point unusable by choosing an audience for it. I sent him a direct message. I did not retract the substance, because the dashboard still needed watching, but I owned the venue: criticising him in front of the whole contributor list was my choice and my mistake, and the reason I cared was that I wanted him reviewing, not that I wanted him timid. Then I asked what would actually help. He said nobody had ever shown him how to read the crash dashboard, which had genuinely not occurred to me. We paired on it for an hour. He runs triage on alternating releases now, and crash-free has stayed above 99.7% since. My rule since then is simple: praise in the thread, corrections in a message.
The diagnosis is what makes this work — one named mistake, the venue, with the substance left standing. Admitting the two weeks of telling himself the silence was busyness is what makes the self-awareness credible. Retracting the point, or blaming the contributor's sensitivity, would sink this answer entirely.
I told a co-maintainer that his manual pre-release pass was holding the project back, and the way I said it cost me two months. What I said on the maintainers' call was that the checklist had become ceremony, and he heard, correctly, that I thought his caution was worthless. He did not argue. He simply stopped volunteering for release duty, which pushed it onto two people who were already stretched thin. What I got wrong was not the conclusion, it was that I had brought an opinion to somebody who had earned the right to data. So I went and got some. I read back through his last eleven pre-release passes: six of them had caught something real, including a lifecycle bug that would have taken crash-free sessions under 98% on every device running the older platform build. I went back to him privately and led with that — your pass catches things, here are the six, and calling it ceremony was wrong of me. Then I asked the question I should have opened with months earlier: which of these eleven checks needs a person, and which needs a script? He picked four for automation himself and kept three manual. The release interval went from nineteen days to nine, crash-free never dipped below 99.62% through the transition, and he took release duty back. What I actually carry from it is that if I am about to criticise someone's process, I owe them the work of understanding what it protects first.
Senior signal comes from the cost spilling onto other people and from a repair that required real work rather than an apology. Leading the second conversation with his evidence, not your point, is the move worth stealing. Grading the outcome honestly — he kept three checks manual — keeps it from reading as a tidy win.
A small story is fine — a review comment that stung, a suggestion that came out as a criticism. Show that you noticed the effect on the other person and went back, rather than deciding they were oversensitive.
Pick something with a real consequence: work slowed, a collaborator disengaged, a decision stalled. The centre of the answer is the diagnosis — venue, timing, wording or evidence — named precisely enough that the fix follows from it.
Your version should show the cost landing on more than the two of you, and a repair that did not require the other person to move first. Say what you changed permanently, and be willing to name a repair that only partly worked.
The interesting version is feedback that failed because of position, not phrasing — your standing made a suggestion land as a directive, or vice versa. Show how you now create the conditions for hard feedback rather than relying on getting each delivery right.
saying these in an interview costs you the question
- Concluding the other person was simply too defensive to hear it
- Retracting the substance to make the discomfort go away
- Stopping at the failure with no second attempt
- Blaming the venue or the timing without naming your own choice in it
- A diagnosis so vague it could apply to any conversation
- What told you it hadn't landed?Be concrete about the signal, because vague ones suggest you noticed late. Behavior changes are the credible evidence: they went quiet, stopped asking you to review, agreed too quickly and did nothing, routed around you. Include how long it took you to notice and be honest if that was longer than it should have been — the noticing is part of what is being scored.
- Was any part of their reaction fair?Answer yes and mean it, or you look defensive. Separate the two halves out loud: the substance you still stand behind, and the part of their reaction that was a reasonable response to something you did. Candidates who cannot concede anything here read as people who will keep making the same delivery mistake with better justifications.
- How long did you wait before trying again?Give the real interval and what you did with it. A short gap that let you gather evidence or think about venue is strong; weeks of silence you spent hoping it would resolve is weak unless you say so and name why. Avoid implying you had a plan the whole time when you were actually avoiding it.