Prompt and Applicable Context
Tell me about a time you received feedback that you initially disagreed with or found difficult to hear. Explain who gave it, the observable behavior it addressed, why you initially resisted it, how you tested it, what you changed, and how you later determined whether the change worked.
This behavioral question applies to engineering, data, product, operations, and management roles. Amazon's current senior software engineer interview guidance still uses behavioral questions to examine how candidates handled past successes and challenges, and recommends STAR, specific details, and data. Microsoft's interview guidance uses STAR(R), explicitly adding Reflection. Amazon's Leadership Principles place attentive listening, candid communication, and being vocally self-critical under Earn Trust. Together, these current official hiring materials support one conclusion: a strong answer needs an observable growth process, not merely “I took the feedback well.”
This article does not claim the question belongs to a particular company. The sample is fictional practice material and must not be presented as personal experience. Every number in it is placeholder data that must be replaced.
What the Interviewer Evaluates
The first signal is whether you can understand the feedback accurately. Feedback often arrives as “you are not collaborative enough” or “your communication is too forceful,” but those are evaluations, not actionable facts. A candidate should restate the concern and ask when it happened, which behavior was observed, what impact it had, and what the other person expected. Immediate rebuttal loses information; immediate confession can be performative.
The second signal is judgment. A feedback giver may have seen only part of the situation or expressed the issue imprecisely. A strong answer does not equate “accept feedback” with “accept every word.” It uses meeting records, work outcomes, observations from relevant people, or later events to decide whether the core signal is valid, which parts need refinement, and which advice should not be adopted.
The third signal is personal responsibility. Even when feedback is only partly right, identify the behavior within your control. Blaming the entire problem on the other person's style sounds defensive. Claiming fault that did not exist just to appear humble is equally unconvincing.
The fourth signal is whether the change is specific. Expect the interviewer to ask what you did differently in the next meeting, review, plan, or conversation. An observable change might be labeling which decisions remain open, inviting objections before stating your preference, changing the speaking order so others propose first, or scheduling an explicit follow-up.
Finally, the interviewer looks for closure. A successful project does not automatically prove the feedback was resolved. Stronger evidence is that the original giver or an affected partner observed a change, records from similar situations show different behavior, or decision quality, rework, or collaboration outcomes improved.
Questions to Clarify Before Answering
- Does the interviewer mean constructive feedback or any criticism? Prefer feedback about work behavior, judgment, or collaboration that has reached closure. Personal attacks, discrimination, or requests that violate professional boundaries do not need to be reframed as growth opportunities.
- Must the feedback have been correct? It need not have been entirely correct. A partly valid example can show judgment, but it needs a fair verification process. Do not choose obviously absurd advice merely to prove that you ultimately changed nothing.
- Did it come from a manager, peer, or direct report? Any can work. Feedback from a direct report or cross-functional partner can be especially useful because it shows how you listen across a power difference. Explain why the person was positioned to observe the behavior.
- Can you discuss performance feedback? Yes, while protecting sensitive personnel details. If the matter remains disputed, involves an active appeal, or has no outcome yet, a closed example is safer.
- How does seniority affect the answer? A junior candidate can emphasize clarification, practice, and timely help. A senior candidate should also explain how their behavior affected decision space, psychological safety, or cross-team efficiency.
- Must the result be quantified? No. Meeting records, later reviews, a follow-up from the owner, or an independent observation can all serve as evidence. Do not manufacture a percentage when no record exists.
30-Second Answer Framework
“During [situation], [feedback giver] told me [feedback about an observable behavior]. I initially disagreed because [my understanding at the time], but I first restated the impact they were concerned about and asked for specific examples. I then reviewed [independent evidence] and checked with [relevant observer]. I concluded that the core feedback was [fully/partly] valid: the more accurate issue was [diagnosis]. Starting with the next [similar situation], I changed [specific behavior] and checked it through [follow-up or verification]. The result was [real outcome], while I am still improving [remaining boundary].”
Step-by-Step Deep Answer
Step 1: Choose feedback that genuinely changed how you work
A useful story has four properties: the feedback addressed observable behavior; you were honestly surprised or disagreed at first; you had a chance to investigate rather than simply comply; and a describable behavior change followed. “My manager said I was doing well but should be more confident” is usually too weak. An unresolved personnel dispute or an event that cannot be anonymized safely is too risky.
Run a differentiation check. A failure question centers on the decision you got wrong and the repair. A fast-learning question centers on closing a capability gap. A disagreement question centers on challenging a proposed decision. This answer must spend its time on how you received, evaluated, and acted on feedback about yourself. If the story still works after removing the feedback giver, it may belong to another prompt.
Step 2: Translate an evaluation into behavior, impact, and expectation
Prepare four columns: original wording | observable behavior | impact | expected behavior. For example, “you did not genuinely ask for input” might mean: the design review invited comments, but the document presented a settled solution; it did not list open questions; partners believed an alternative could no longer affect the decision.
You cannot perform this translation alone on the giver's behalf. In the real conversation, first restate: “I hear that the problem is not the lack of speaking time; it is that people no longer had meaningful decision space when they received the material. Is that right?” Then ask for one or two specific situations. If the person cannot produce an example immediately, agree to revisit it. Do not turn clarification into cross-examination.
Step 3: Manage the first reaction without pretending it did not exist
Difficult feedback often triggers an urge to explain. You may honestly say that you were surprised, embarrassed, or unconvinced, but the next sentence must describe what prevented the reaction from taking over the judgment. A useful sequence is: pause, restate, request examples, confirm the expected behavior, and agree on follow-up.
“I accepted it completely on the spot” is not automatically a strength. A promise to change before understanding the facts is cheap agreement. Do not claim that you felt nothing merely to sound composed. The interviewer cares more about how you managed the reaction.
Step 4: Use evidence to accept fully, accept partly, or decline
Verification is not a search for allies. First write down what evidence would show the feedback is true, then inspect records: materials from several comparable meetings, comment timing, who proposed alternatives, when a decision was marked irreversible, and whether people raised objections only afterward. You may ask another person who actually observed the events, but use neutral wording such as, “At what point did you believe the proposal could still change?”
Reach an explicit conclusion:
- Accept fully: multiple independent examples point to the same behavior and impact.
- Accept partly: the impact is real, but the explanation of its cause or scope is too broad.
- Decline: the evidence does not support it, or the requested change conflicts with a more important safety, ethical, or responsibility boundary.
Even when declining the advice, acknowledge that the person exposed a perception or risk and explain how you will prevent the same misunderstanding. The decision rule is not “all feedback must be followed.”
Step 5: Turn the change into a small, observable experiment
Goals such as “be more open” or “communicate more” cannot be verified. Convert them into actions in the next comparable situation:
- Label sections of the pre-read as “decided,” “decision needed,” and “input requested.”
- State the criteria that would change the choice and the questions that remain open.
- Ask affected partners to state risks before you give your preference.
- Reserve time to compare alternatives and record why each was accepted or rejected.
- Follow up with the original giver after two or three comparable events.
The change must match the feedback. If the problem was that review happened too late, adding more one-on-ones does not restore decision space. If the problem was an intimidating tone, sending the document earlier is not sufficient.
Step 6: Close the loop with layered evidence
Present the result in three layers. First is behavioral evidence: did you follow the new practice? Second is other-person evidence: what did the original giver or another partner observe? Third is work evidence: did the team expose risk earlier, reduce rework, or reach a better option? Not every layer needs a number, but preserve the causal order.
Do not ask, “Am I doing well now?” A better follow-up is, “Across the last three reviews, where did the decision space still feel artificial?” That question is more likely to produce actionable evidence. If you have used the change only once, say that evidence is limited rather than declaring that your personality has changed.
Step 7: Use STAR(R) to protect the spoken emphasis
Situation establishes where the feedback arose. Task states your responsibility and why the feedback mattered. Action covers restatement, investigation, judgment, and behavioral change. Result gives behavioral, interpersonal, and work evidence. Reflection explains how you now detect the same blind spot earlier. Action and Reflection should receive most of the time.
During practice, have someone challenge you: “Did you only ask people who agreed with you?” “What if the giver says you still have not changed?” “Which part did you actually disagree with?” If the facts remain consistent under those questions, the story is less likely to be performative humility.
High-Quality Sample Answer
The following is a fictional example used only to demonstrate structure. Do not present the plot as personal experience. Every number is placeholder data that must be replaced.
“I led technical proposal reviews for a six-person cross-functional group. After one meeting, the product lead told me that although I asked whether anyone had comments, the materials and the way I spoke made the decision feel settled, so alternatives never entered the discussion. I initially disagreed because I always sent a pre-read and explicitly invited questions during the meeting.
I did not explain immediately. I first restated that her concern was not the absence of speaking time; it was that input arrived too late to affect the proposal. She gave one example: an engineer proposed a smaller rollout only after the meeting because they believed the meeting was confirming execution. I reviewed the previous four proposal documents and separately asked one engineer and one operations partner, “At what point did you believe the proposal could still change?” Three of the four documents were labeled ‘proposal,’ but none identified open questions, decision criteria, or rollback conditions. These are placeholder counts that must be replaced.
I partly accepted the feedback. I still believed I should bring a clear recommendation rather than hand the team a blank problem, but I confirmed that I had written the recommendation as if it were a final decision and always stated my preference first. I changed the pre-read template to separate ‘decided,’ ‘decision needed,’ and ‘input requested,’ listed the conditions that would overturn my recommendation, and reserved the first 10 minutes for affected partners to state risks before I spoke. Ten minutes is also placeholder data that must be replaced.
Across the next three reviews, two early suggestions changed the rollout scope and acceptance metric respectively. Three reviews and two suggestions are placeholder data that must be replaced. I followed up with the product lead who gave the original feedback. She said the decision boundaries were much clearer, but noted that I still summarized too quickly under time pressure. I kept that as the next improvement: another participant now summarizes the disagreement before I make the decision. The experience taught me that asking for input is not the same as creating meaningful influence. I need to verify that people can identify when and by what criteria the decision can change.”
When substituting your experience, discard the six-person group, four documents, 10 minutes, and three reviews. Preserve the evidence structure: original feedback, initial disagreement, specific example, independent verification, scope of acceptance, behavioral experiment, follow-up result, and remaining boundary.
Common Mistakes
- Turning the first reaction into a debate to win → The interviewer cannot see listening or recalibration → Restate the impact accurately before explaining how you tested it.
- Claiming immediate, complete acceptance → This may be appeasement and contains no judgment process → Show which facts moved you from doubt to full or partial acceptance.
- Choosing obviously absurd feedback → The story proves only that the other person was wrong → Choose a reasonable observation drawn from real behavior.
- Ending with “we had different personalities” → A changeable behavior becomes a fixed personality conflict → Name the specific situation, behavior, and impact.
- Asking only colleagues who support you → Verification becomes confirmation bias → Define disconfirming evidence first and ask people who actually observed the event.
- Changing only by “communicating more” → No observable behavior has changed → State what someone will see you do differently in the next event.
- Using a good project result as proof → The outcome may be unrelated to the behavioral change → Add a follow-up from the original giver or behavioral records.
- Portraying the giver as ignorant of the business → This exposes defensiveness and weakens credibility → Describe their observation boundary fairly, then identify the exact part you disputed.
- Claiming no remaining improvement → One adjustment is presented as a complete transformation → Name the limited evidence or the next behavior you still monitor.
Follow-Up Questions and Responses
Follow-up 1: Why did you disagree at first?
Name the concrete evidence or self-image you relied on, such as sending pre-reads, explicitly inviting comments, or achieving good prior project results. Then explain why that evidence was insufficient. Completing the ritual does not mean others had meaningful influence, and a good outcome does not prove a healthy collaboration process.
Follow-up 2: How did you avoid asking only people who supported you?
Define the behavior that would prove the feedback valid, inspect comparable event records, and ask people who observed the situation from different interest positions. Use neutral wording that does not reveal the answer you want. When evidence conflicts, preserve the conflict and narrow the conclusion rather than voting on the truth.
Follow-up 3: What if the feedback turned out to be invalid?
Separate the factual judgment from the relationship signal. You may decline the advice while still explaining why the person formed that perception and how you will make the decision boundary, responsibility, or standard clearer. Discriminatory, humiliating, or unsafe requests should be documented and escalated through the appropriate channel, not personally “absorbed.”
Follow-up 4: How do you know you actually changed?
Use at least two supporting layers: behavioral records from comparable situations, follow-up from the original giver, and work outcomes. If the change is new, state that you have only completed a template, practice, or one observation and cannot yet claim a durable habit.
Follow-up 5: What if the feedback came from your direct report?
A power difference raises the cost of speaking. Thank the person for taking the risk without demanding more sensitive details on the spot. Offer a later one-on-one or anonymous channel. Follow up after changing, but do not make the direct report responsible for reassuring you or proving that you improved.
Follow-up 6: How would you detect the same problem earlier now?
Turn the feedback into leading signals: decision documents with no open questions, the same people repeatedly objecting only after meetings, only the owner proposing options, or alternatives not appearing in the record. State when you check the signal, who can flag it, and which process changes when it appears.