Question and Suitable Scenarios
Tell me about a time you managed a team member whose performance was below expectations. Explain what the role required, which sustained and observable work results missed that standard, how you verified the facts and examined your management responsibility, how you discussed causes, support, and expectations, and what evidence drove the eventual decision.
This is a senior behavioral question for engineering managers, development managers, technical leads with formal people responsibility, and comparable management roles. It requires evidence across people development, resource allocation, delivery protection, and a difficult decision, so it best fits somebody formally accountable for a direct report's role outcomes, support, and follow-up. The answer should center on one real experience and its evidence rather than guessing a company's fixed wording.
The competency boundary is sustained performance accountability. A difficult-feedback answer centers on one consequential conversation and behavioral closure. A mentoring answer centers on independent transfer of a capability the person is developing. A delegation answer centers on ownership, authority, and guardrails. Feedback, coaching, and delegation may all appear here, but the manager must still own whether role standards were clear, support was sufficient, the team and customers were protected, and a fair step followed if improvement did not occur.
Prefer a real case in which you held formal management responsibility. If you were a peer or project lead, you can describe identifying risk, giving feedback, offering help, and escalating to the formal manager. You cannot claim to have set a performance rating, improvement plan, or employment decision that you did not own. Documentation, accommodations, employee representation, formal warnings, and termination requirements vary by employer and jurisdiction. State that you followed local policy and consulted HR or employee relations when appropriate; this article is not a legal procedure.
The demonstration later is fictional practice material and must not be presented as personal experience. Six weeks, eight milestones, five missed milestones, four weeks, seven met milestones, four review rounds, two review rounds, zero customer incidents, and every other quantity and result are placeholder data that must be replaced.
What the Interviewer Evaluates
First, can you replace a label with evidence? “Underperforming” and “bad attitude” are not operational facts. A strong answer names the role standard, observation window, task complexity, and comparable record: which committed design milestones were repeatedly missed, how quality was defined, and what impact reached the team or customer. One mistake, one unfamiliar system, or one reasonable technical disagreement usually does not establish a sustained performance gap.
Second, did you inspect management and system contributions? Ambiguous goals, changing priorities, work beyond capacity, weak onboarding, missing access, or your tolerance of inconsistent standards can create or amplify a gap. Owning those factors does not erase individual accountability. It shows that you corrected conditions within your control before judging performance against a clear, feasible standard.
Third, can you communicate directly and respectfully? The person should not first hear about the problem in a rating or consequential decision. You need to discuss observations, impact, and expectations privately, invite additional facts, and confirm shared understanding. Asking what prevented an agreed result is more professional than guessing about laziness, motivation, health, or family circumstances.
Fourth, is the improvement path specific and fair? It should define work outcomes, evidence, reasonable time, support and resources, review rhythm, and what happens if progress remains insufficient. Targets cannot be designed to fail, and “be more proactive” or “improve quality” cannot be the entire plan. If the role, level, or workload changes, the standard may need recalibration too.
Fifth, how do you protect others? Moving every rescue indefinitely to high performers makes the team pay for a management problem. Removing all responsibility immediately prevents the person from demonstrating improvement. A mature answer segments customer risk, creates temporary guardrails for critical paths, preserves meaningful and assessable ownership, and explains work allocation without exposing private performance information.
Sixth, can you handle different outcomes? Improvement, partial improvement, role or task mismatch, and sustained non-improvement require different decisions. A strong answer does not make “I saved someone” the only acceptable ending or declare success after one good week. It names the observation period, comparable next task, decision participants, and policy-based escalation path.
Finally, interviewers test reflection and attribution. The person owns their choices and work. The manager contributes clear conditions, support, risk protection, and a timely decision. Name where you intervened too late, left a standard ambiguous, or checked so frequently that you took over, then specify the earlier signal that now changes your behavior.
Questions to Clarify Before Answering
- Were you the formal manager? State your authority over evaluation, assignment, and escalation. Without formal authority, frame the story as discovery, feedback, support, and handoff.
- Below which standard? Name the role and level expectation, communicated outcome, quality or timing threshold, and available resources. Do not use the team's strongest person as the default standard.
- Was this capability, environment, conduct, or serious misconduct? Knowledge, judgment, or resource barriers may call for training or work changes. Refusing a reasonable and understood requirement may be conduct. Harassment, safety, or integrity issues often require a separate process and should not be disguised as routine coaching.
- How much comparable evidence existed? Separate direct observation, work records, quality results, and relevant feedback. Verify secondhand opinions; an anonymous complaint is not an automatic conclusion.
- What was your management contribution? Check clarity, priority stability, capacity, context, and whether you delayed feedback.
- Was support or an accommodation channel relevant? Ask about work barriers and needed support without probing for a private diagnosis. Involve the responsible professional under company policy for health, disability, discrimination, leave, or another protected matter.
- How would you protect the team and customer? Define high-risk work that needs review, responsibility the person retains, and how extra work is time-bounded and reassigned.
- What does improvement or non-improvement trigger? Define evidence, review dates, and possible consequences before evaluating the result so the goalposts do not move afterward.
- Can the story be anonymized safely? Remove names, ratings, health information, and irrelevant identity details. Keep only work facts needed to understand your judgment.
The 30-Second Answer Framework
In [team and role context], over [observation window], I saw [repeated gap against a clear standard],
which caused [team, delivery, or customer impact]. I checked [direct records and comparable work] and
also found that I contributed through [goal clarity, capacity, support, or feedback timing]. I privately
shared the observation and impact, invited context, and separated [capability, judgment, capacity,
process, or conduct factors]. We defined [work outcome, evidence, support, duration, checkpoints, and
the consequence of insufficient improvement]. I protected critical delivery with [temporary guardrail]
without transferring rescue work to the team indefinitely. At [review point], [performance evidence]
and [team or customer evidence] showed [improvement, partial improvement, or non-improvement], so I
[restored ownership, adjusted the plan or role, or escalated under policy]. I learned that I had
[specific management mistake], so I now act when [earlier signal] appears.Use STAR for the timeline and add Reflection. Situation establishes the team, role, and standard. Task explains your dual responsibility for fair people management and delivery. Action should take most of the answer: evidence, conversation, support, guardrails, and review. Result accounts for both individual performance and team impact. Reflection states what you now do earlier.
Do not turn the answer into a management philosophy. The interviewer needs one real case, the pivotal things you said and did, new information the person supplied, how the plan changed, and why result evidence supported the final decision.
Step-by-Step Deep Answer
Step 1: Choose a story with formal responsibility and a sustained gap
A useful story includes a clear role standard, comparable evidence across more than one event or a reasonable period, management action you personally owned, context the person could add, a supported improvement path, and a decision after improvement or non-improvement.
Avoid a single ordinary mistake and an unresolved case whose basic facts cannot be shared. If your entire action was “I delivered difficult feedback and they accepted it,” use the feedback question. If the person chose a development goal and the ending is increased independent capability, mentoring may be a better fit. This question requires accountability for a sustained standard, burden on other team members, and the next branch.
Step 2: Establish a factual baseline before using a performance judgment
Separate five layers: role and level standard, communicated commitment, direct observation or work record, delivery and collaboration impact, and unverified explanation. Ask whether a reasonable reviewer familiar with the role would see the same gap from the same anonymized record.
Check comparability. Someone working in a new system, a high-risk legacy module, or a heavier interrupt rotation cannot be compared through raw output counts alone. Defect totals also need scope, severity, and discovery stage. Use the minimum evidence needed to explain the judgment and do not expose internal personnel data in the interview.
List the conditions you created or permitted. Did you assign conflicting goals, use “as soon as possible” as a deadline, fail to define done, appear only when work failed, or wait while hoping the problem would resolve itself? Correct those conditions and retain the clear result the person still owns.
Step 3: Hold a private conversation that is clear and answerable
Open with purpose, observation, and impact: “I want to discuss delivery across the last three design milestones. We agreed on a reviewable version by Wednesday; twice, the critical risk analysis was still missing on Friday, and the team had to add it before release. I need to understand what I missed and make the role standard clear.” Every count and day in such wording must come from your real case.
Invite responses to facts, causes, and feasible support separately. Possible causes include unclear goals, missing foundations, undeveloped judgment, too much parallel work, tool or access barriers, collaboration problems, or a different understanding of the role. Do not automatically dismiss an explanation as an excuse or accept it as proven. Select testable factors and corresponding actions.
If the person shows that you changed priorities repeatedly, verify and acknowledge it. If the core standard remains unmet, remain direct: “I will remove conflicting work and supply the missing context. With those conditions corrected, this role still requires a reviewable result by the agreed date.” Fairness includes both listening and explicit accountability.
Step 4: Confirm a workable improvement agreement
The plan answers seven questions: What outcome is required? What proves it? Which work is in scope? Which support is available? When do you check? When do you evaluate? What happens if the outcome remains below standard? Ask the person to restate their understanding and challenge feasibility; the manager remains accountable for the final role standard.
Turn vague goals into observable results:
- Replace “communicate proactively” with updating the decision record and notifying the owner within one working day after an agreed risk threshold is crossed.
- Replace “improve design quality” with covering requirements, alternatives, principal risks, rollback, and open decisions before review.
- Replace “deliver on time” with producing a named artifact at each checkpoint and bringing scope, resource, or date options when the forecast changes.
Support should match a verified barrier: training, demonstration, paired review, removal of conflicting work, access, decision-right clarity, or a better-bounded assignment. A library of generic courses is not diagnosis. Quietly lowering the role standard and calling the issue solved is not improvement.
Step 5: Protect the team and customer with time-bounded guardrails
Segment work by impact and reversibility. Customer funds, privacy, safety, data integrity, or irreversible releases may need another reviewer or a temporary change to final approval. Low-risk, reversible work should remain with the person so that real evidence can emerge.
Extra burden needs an owner and an exit condition. If a senior peer reviews designs temporarily, say for how long, what they review, which coordination cost you absorb, and when normal allocation resumes. Tell the team only what it needs to execute: “Critical settlement changes have a second reviewer this month.” Do not announce a performance plan or recruit peers to monitor the person.
Step 6: Monitor outcomes without turning monitoring into takeover
Use the same evidence at each check-in: completed outcomes, missed items, barriers, support used, next commitment, and decisions needed. Record important agreements, share them, and let the person correct the record. Timely daily feedback and formal reviews should align; a final evaluation must not introduce a surprise standard.
Do not ask for status every hour or complete the key judgment yourself. Match frequency to risk and plan duration, then reduce it with stable improvement. If the manager repeatedly rescues the work, successful delivery cannot establish individual performance. Track what the person completed independently and how much you intervened.
Step 7: Select the next branch from predeclared evidence
Meets the standard: Confirm stability across comparable tasks, remove temporary guardrails gradually, restore responsibility, and name behaviors to sustain. One good result is not enough to end observation immediately.
Partially improves: Decide whether the remaining gap can reasonably close with time and support and whether work conditions or standards changed. Extend or adjust scope only with a stated basis and new endpoint; do not extend indefinitely to avoid a decision.
Role or task mismatch: If the person consistently produces needed results in another responsibility and a genuine role exists, discuss a role or duty change. Do not invent a position to hide the problem or unilaterally change employment terms.
Does not improve: Bring the facts, support provided, and review record into the applicable HR or employee-relations process. Steps vary by jurisdiction and company. The interview signal is timely, clear, and fair management accountability, not pride in a termination decision.
Step 8: Account for results, contribution, and reflection
Report at least three layers: the person's performance on comparable work, impact on team or customer, and removal or continuation of temporary guardrails. State the observation window and limits. If the evidence shows short-term improvement only, say they met the standard in the defined scope; do not claim a permanent transformation.
Attribute precisely: “They chose to practice and delivered the improvement. A peer provided two technical reviews. I clarified the standard, removed conflicting work, arranged support, protected the critical path, and made the decision from shared evidence.” Then name your error—feedback two weeks late, reliance on raw output, or checkpoints that were too intrusive—and the earlier condition that now triggers action.
High-Quality Sample Answer
The following is fictional practice material and must not be presented as personal experience. Six weeks, eight milestones, five missed milestones, four weeks, two parallel projects, two paired reviews, seven met milestones, four review rounds, two review rounds, and zero customer incidents are all placeholder data that must be replaced.
“I managed a backend team responsible for a payments platform. After a senior engineer took ownership of the settlement service, eight committed design and incident-review milestones fell within six weeks, and five missed the agreed date. Three designs required other engineers to add critical risk analysis before release, and review averaged four rounds. Six weeks, eight, five, three, and four rounds are placeholder data that must be replaced.
I was responsible for making the role expectation clear and giving the engineer a fair opportunity to improve, while preventing the team from absorbing continuing rework and customer risk. I checked milestone records, task scope, and review comments without using colleagues' personality labels. I found that I had contributed: I assigned two migration projects without making the priority explicit, and our definition of a reviewable design did not require risk and rollback sections. I stopped treating the entire gap as an individual problem.
In a private one-on-one, I shared the observations and impact and asked how the engineer understood the standard and what blocked delivery. They said the new settlement domain lacked context and urgent reviews repeatedly interrupted planned work. They also acknowledged that risk analysis began only after comments arrived. The interrupt record supported a capacity factor, but the role still required surfacing risk early and escalating a changed forecast.
We agreed on a four-week plan. I would reduce two migration projects to one, arrange two paired reviews on settlement boundaries, and have the product owner make priority explicit. The engineer would submit a weekly design by Wednesday covering alternatives, major risks, rollback, and open decisions. If the date became risky, they would escalate within one working day with scope, resource, or date options. We reviewed the same record weekly and stated in advance that, after four weeks, comparable milestones would support restoring full ownership, adjusting the plan, or moving to the next company-policy step. Four weeks, two projects, two reviews, Wednesday, and one working day are placeholder data that must be replaced.
To protect customers, changes affecting money consistency temporarily gained a second approver. The engineer still led other reversible designs. The second approval lasted only until two consecutive designs met the team standard, and I coordinated the extra review time rather than transferring it indefinitely. I told the team that critical settlement changes temporarily required two reviewers; I did not disclose personal performance information.
At the first check-in, I noticed that I had rewritten the document line by line. That made the artifact look better but could not prove independent performance. I changed my approach: the engineer first explained the recommendation, evidence, and risk, and I intervened only at a safety boundary or critical omission. At the four-week review, seven of eight comparable milestones met the defined standard on time. Average review fell from four rounds to two, no customer incident occurred, and the final delay had been escalated that day with an approved scope change. Four weeks, eight, seven, four rounds, two rounds, that day, and zero incidents are placeholder data that must be replaced.
The engineer performed the improvement, and peers contributed specialist feedback. My contribution was correcting conflicting priorities, clarifying the role standard, arranging targeted support, bounding customer risk, and keeping decisions tied to evidence visible to both of us. We removed the second approval gradually and restored full responsibility. In retrospect, I should have addressed the pattern at the third missed milestone instead of waiting for the fifth. I now check standards, capacity, and support and hold a direct conversation when two comparable commitments fail consecutively. Third, fifth, and two consecutive events are also placeholder data that must be replaced.”
When adapting this story, remove the payments, settlement, migration, and dual-review setting. Substitute your real role, standard, evidence, support, policy, and result. Preserve the causal structure: sustained gap, management contribution, direct conversation, assessable plan, time-bounded guardrail, independent evidence, next decision, and specific reflection. A non-improvement ending can still be strong if the process was timely and fair and the decision followed the applicable process.
Common Mistakes
- Calling the person an underperformer → The label hides standard and evidence → name role expectations, comparable tasks, observation window, and impact.
- Concluding from one incident → There is no sustained pattern and task novelty may be ignored → use comparable work across a reasonable period.
- Reducing the cause to attitude → Motive is guessed and support cannot be chosen → ask for context and test knowledge, judgment, capacity, process, and conduct factors.
- Ignoring management responsibility → Ambiguity and priority conflict may keep producing the gap → correct conditions within your control first.
- Showing empathy without stating the standard → Nobody knows what must change → make support, required outcome, and consequence clear together.
- Using a PIP as the first sentence or as punishment → Facts, dialogue, and applicable policy disappear → begin with timely feedback and support, then follow the process for formal escalation.
- Designing an impossible plan → The outcome is predetermined and unfair → use relevant, assessable role outcomes and reasonable time.
- Making high performers rescue indefinitely → The team pays the management cost → bound extra work, set an exit condition, and absorb coordination as the manager.
- Exposing customers to create an opportunity → Customers become the performance experiment → set temporary review by impact and reversibility.
- Doing the work yourself and declaring improvement → The result cannot be attributed to the person → track independent judgment, artifacts, and intervention.
- Explaining private performance to the team → Privacy and trust are damaged → communicate only necessary work allocation.
- Promising that everyone can be saved → Role mismatch and non-improvement are avoided → define a fair branch for each outcome.
- Using termination to prove toughness → Development responsibility and process are skipped → show evidence, support, protection, and a timely decision.
- Skipping verification on later work → Short-term movement may not be stable → use comparable tasks and removal of guardrails.
- Reflecting only that you should communicate more → No behavior can change → name an earlier trigger and action.
Follow-Up Questions and Responses
Follow-up 1: How did you know the standard was fair rather than your personal work style?
Return to role and level expectations, communicated outcomes, task scope, and multiple comparable cases. Separate outcome standards from implementation style. If another method meets quality, timing, collaboration, and risk boundaries, do not demand your habits. Have your manager or HR check consistency when appropriate, and let the person challenge both facts and feasibility.
Follow-up 2: What if the person says a personal or health issue affected work?
Express concern and ask what workplace support and formal channel they need without probing for a diagnosis or irrelevant private details. Address role outcomes and immediate risk while involving the responsible HR, employee-relations, or accommodation professional under local policy. Do not reveal real health information in an interview or describe a legally required adjustment as special treatment.
Follow-up 3: Other team members are already frustrated. What do you do?
Acknowledge the workload issue and adjust critical assignments, rotations, or reviews immediately without discussing the private performance process. Give anyone carrying extra work a defined scope, end date, and priority offset. Continue watching morale and quality. The manager owns coordination and decisions; the team should not vote anonymously on whether somebody stays.
Follow-up 4: What if the first support method does not work?
Recheck the barrier hypothesis. Training may fail because the actual issue is capacity, judgment, or role fit. Frequent check-ins may create dependence. Change one evidence-backed factor while preserving the role standard and final review date. If reasonable support still produces no improvement, follow the stated process instead of adding coaching forever to delay a decision.
Follow-up 5: Is this still a good interview story if performance did not improve?
Yes. Explain when you recognized the gap, how you corrected management factors, what realistic support you supplied, how you protected the team and customer, which evidence still missed the standard, and who made the next decision under which policy. Add what you could have done earlier. A fair and timely non-improvement outcome is more credible than an invented turnaround.
Follow-up 6: What did you contribute, and what belongs to the person and others?
Your contribution may be establishing the evidence baseline, clarifying the standard, removing system barriers, arranging support, bounding risk, maintaining review discipline, and making decisions within your remit. The person owns practice, choices, and delivery. Peers own their specialist help. HR or senior leadership owns its formal decisions. This attribution prevents another person's growth or departure from becoming a manager-hero story.
Follow-up 7: What would you do earlier now?
Give a concrete trigger: two comparable commitments missed consecutively, rework moving beyond the team baseline, or the team beginning to rescue the same responsibility routinely. Before the next one-on-one, check the standard, task scope, and record, then discuss the pattern directly. The trigger starts inquiry and communication; it is not an automatic finding of unacceptable performance.