Tell me about a time you disagreed with a decision your team made and had to support it anyway.
answer
- name the decision and your position
- argued in writing, before the call
- state the moment debate closed
- commit visibly — what you built
- result plus one honest reflection
basics
~20 sTests whether you can argue hard and then execute without hedging. Give one decision you lost, show where you voiced the objection while it was still open, then what committing actually looked like in your work.
how to answer
6 beats- the decision on the table and what you wanted insteadTwo or three sentences, no more — this beat plus the stakes should be about a fifth of your airtime. Name the actual choice and your alternative in one line each, so the interviewer can hold both positions in their head.
- why it mattered enough to argue aboutGive the stake in concrete terms: users affected, risk carried, time lost. This is what separates a real disagreement from a preference, and it earns you the right to have pushed back.
- how you made the case while the call was still openName the venue and the form — a comment on the proposal, a question in the review, a message before the meeting. Written dissent with specific failure modes is the most convincing detail you have.
- the moment the decision landed and how you took itGive the hinge sentence: what argument you had no answer to, and the words you used to close it. This one line is what makes everything after it read as commitment rather than resignation.
- what committing actually looked like in your workThis is the bulk of the answer, roughly sixty percent of your airtime. Show something you did for the chosen plan that you would not have done for your own — built its safety net, took the dull piece, defended it to someone still grumbling.
- the outcome and what you would argue differentlyClose with a number and one honest concession, in about a fifth of your airtime. If you turned out to be wrong, say what you had missed; if you were right, say it once and move to what the team changed.
your answer
5 story prompts- Pick a decision that went against you where you can still state your original objection in one sentence.
- Choose one where you argued in a visible place — a doc, a review — not only in your head.
- Write down what you did afterwards that actively helped the plan you had argued against.
- Have one number for the outcome ready, even if it shows you were wrong.
- This can be the same episode as your conflict story, re-angled onto the committing half.
draft and rehearse your own answer in a learn session
go deeper
Probes disagreement hygiene and ownership. Interviewers are screening for two failure modes at once: silent compliance while the decision is open, and passive resistance after it is made. A strong answer proves you will raise the hard objection in a place people can see it, and then put full effort behind a call that went against you.
I was the newest of four engineers on an enterprise field-service app — technicians use it offline all shift in warehouses — and we were moving the local data layer off a hand-rolled cache onto a proper persistence framework. I wanted to migrate one screen at a time behind a flag. Our tech lead wanted the whole store swapped in a single release, because otherwise two caches would disagree with each other for months. I wrote my case in the design thread: three specific failure modes, and my estimate for the staged path. I asked about it again in the review. He answered every point, and I had no counter to the coexistence argument, so I said I still preferred staged but I was in. Then I tried to be the person who made his plan safe. I built a seed-and-verify harness that replayed a technician's full day against both stores and flagged mismatches, and I owned the rollback switch for the release. It caught two write-ordering bugs that would otherwise have shipped. Our store rating slid from 4.42 to 4.28 for about three weeks after the cutover, then came back to 4.46 once the sync fix landed. I still think staged would have been calmer for us, but the cost of two stores disagreeing was real and I would not have seen it if I had kept arguing instead of listening.
The signal is carried by three details: the objection was written down with named failure modes, there is a clear hinge sentence closing the debate, and the commitment shows up as work that helped the plan he argued against. Downlevels instantly if the harness disappears and he only says he supported the decision.
I owned the scheduling and dispatch flows on an enterprise mobile app used by delivery crews. The platform lead decided to move two of those flows onto a shared cross-platform UI layer so the phone and the ruggedised handheld builds could share code. I argued against it for my flows specifically — the handhelds the crews carry are three generations behind, and I had measured our first shared-layer experiment there: noticeably worse scroll behaviour and two accessibility regressions. I lost that argument, and I think fairly — the duplication cost across builds was larger than I had priced. Before committing I asked for one thing in writing: a revisit trigger. If crash-free sessions on the handheld fleet dropped below 99.5 percent within six weeks of a flow going live, we would revert that flow, no fresh debate needed. After that I stopped reopening it. I moved my two contractors onto the port, and I was the one who explained the reasoning to them rather than saying it was not my call — which mattered, because one of them had been in the thread agreeing with me. Dispatch held at 99.64 percent and shipped. Scheduling hit the trigger in week four, we reverted it that day, and nobody had to re-argue anything. Our rating moved 4.31 to 4.44 across the quarter. The written trigger is the part I would repeat.
The asked-for condition is the whole answer: committing with a written revisit trigger rather than blind compliance, then honouring it without drama when it fired. The line about explaining the rationale to a contractor who had agreed with her is the peer-scope move. It would downlevel if the trigger were only a private intention.
Your scope is your own task. Show that you raised the objection once, clearly, to the person making the call, and then did the work as well as if it had been your idea — not sulking, not doing a deliberately literal job.
Work at feature and peer scope. Show the mechanics: where you argued (a design thread, a review), what you asked for before committing, and how you stopped the team reopening it — including explaining the rationale to people who still disagreed.
Team and production scope. Give the decision a real chance and instrument it: a written revisit trigger, a metric that would prove you right or wrong, and what you changed in how the team makes that class of call afterwards.
Org and strategy scope. Show the norm you carry: written dissent, one escalation at most, then the decision belongs to the org. Name how you kept multiple teams executing without hedging, and what evidence you said in advance would change your mind.
saying these in an interview costs you the question
- Still re-litigating the decision in the interview room
- Quiet compliance — the objection was never voiced before the call
- Committing in words while slow-walking the work
- Blaming the decision-maker for an outcome you predicted
- No result at all, just the argument
- A disagreement so trivial nothing was at stake
- Did the decision turn out to be the right one?Answer honestly either way, and keep your tone identical in both directions. If it went well, say what you had missed — that is the strongest sentence in the whole story. If it went badly, describe the impact plainly and what the team did next, without a told-you-so. Interviewers read a gloating or grudging tone here as a preview of how you will behave on their team.
- How did the decision-maker know you disagreed?Point to a specific, visible channel: a comment on the design doc, a question in the review, a one-to-one before the meeting. Vague answers — 'I made my concerns known' — suggest you never really said it. Also say when: dissent that arrives after the call is not dissent, it is second-guessing, and interviewers are listening for the ordering.
- What would you have done if the outcome had been much worse?Describe a threshold you would have named up front rather than a heroic intervention afterwards. Good answers sound like: I would have asked for the condition under which we revert, and raised it the moment we hit it. That shows you can commit without going blind, which is the distinction the question is really probing.
## Two halves, one episode This prompt has two halves and most candidates answer only one. - The **first half** is whether you can disagree at all — whether you will say the uncomfortable thing while the decision is still open. - The **second half** is whether you can then execute a call that went against you with full effort. Answering only the first half gives you a story about being right that never resolves. Answering only the second gives you a story about compliance, which reads as someone with no technical opinions. The interviewer wants both, in that order, from one episode. ## Wording variants you will hear - 'Tell me about a time you disagreed with your manager.' - 'Describe a decision you did not agree with but had to implement.' - 'When have you had to get behind something you argued against?' - 'Tell me about a time you were overruled.' They all take the same story. The only variant that changes your material is one that names a person — a manager, a peer, a partner team — because the power dynamic changes what counts as a good escalation path. Have one story and be ready to swap the counterpart. ## Where the dissent happened matters more than how strongly you felt A strong answer names a place: - a comment thread on the proposal, - a question asked in the design review, - a message sent before the meeting so the decision-maker was not surprised in public. **Written dissent** is the single most convincing detail available to you, because it is checkable in spirit — a person who writes their objection down is a person who has thought it through and is willing to be wrong on the record. Vague phrasing — 'I pushed back', 'I made sure my concerns were heard' — is the commonest weak signal in this whole question family. ## Name the moment debate ended Good answers have a **hinge sentence**: the lead answered the coexistence argument and I had no counter, so I said I disagreed and I was in. That sentence does three things at once: 1. it shows you were listening, 2. it dates the end of the argument, 3. and it makes everything after it read as commitment rather than resignation. Without the hinge, the story sounds like you were simply outvoted and complied. ## Commitment has to be visible in the work Do not assert it; show it. The strongest evidence is that you did something for the chosen plan that you would not have done for your own — built the safety net it needed, took the least glamorous piece, or explained the rationale to a teammate who was still grumbling. Contrast that with the weak version, where the candidate says 'I fully supported it' and then describes doing exactly the minimum assigned to them. ## Evidence types that land, in rough order of strength 1. a written objection with specific failure modes; 2. a condition you asked to be written down before committing (a revisit trigger, a metric threshold); 3. an action that materially helped the plan you argued against; 4. a number for the outcome; 5. and a reflection that concedes something real. ## How the bar moves with level - **Early on**, the whole signal is that you spoke once and then worked well — nobody expects you to have shaped the process. - **In the middle**, the interviewer starts listening for how you handled other people's continued disagreement, because a feature owner who lets a decision be reopened weekly is expensive. - **At senior level** they want the decision instrumented: what would have told us we were wrong, and when. - **At principal level** they want a repeatable norm rather than a good instinct — a standing rule about how dissent is recorded and how many times something can be escalated — applied consistently enough that people below you can predict it. ## The failure to avoid above all others It is arguing the case again in the interview. If four of your six sentences are about why your approach was better, you have shown the interviewer exactly what living with the decision looked like: it never ended.