Tell me about a time you escalated a decision you disagreed with to someone more senior.
answer
- name the stake that justified escalating
- show you exhausted the direct route
- you told them you were escalating
- escalate the problem, not the person
- state the ruling and how you carried it
basics
~20 sTests judgment about when going over someone's head is proportionate. Show that you exhausted the direct route, escalated the problem rather than the person, said you were doing it out loud, and committed fully to whatever came back.
how to answer
5 beats- the decision and why the stake was outside your authorityOpen with the consequence, not the disagreement — who or what got hurt, and at what scale. Keep it to a fifth of your airtime; the proportionality is what earns the rest of the story.
- what you tried first with the decision-makerShow at least one genuine attempt at the original level, including an alternative you offered. An escalation with no prior attempt reads as going around someone, and interviewers will hear it that way.
- how you raised it — to whom, in what form, and who knewName the form, ideally written: options, costs, your recommendation, and what would change your mind. Say explicitly whether you told the person you were escalating past, because that single fact decides how the story reads.
- the ruling and how quickly you closed the argumentThis is the bulk of your airtime alongside the next beat. Give the decision that came back — often a middle path — and the moment you stopped arguing, in the words you used.
- how you carried it afterwards, and the relationshipClose with what you owned in the chosen path and how you and the decision-maker worked together next. One number for the outcome, and no trace of a victory lap.
your answer
5 story prompts- Pick the one escalation you would defend again, not the most dramatic one you have.
- Write down exactly what you tried with the decision-maker before you went above them.
- Name the stake — money, safety, people locked out — that made escalating proportionate.
- Be ready to say whether you told that person first, and what you would change.
- Prepare one line on how you and that person worked together six months later.
draft and rehearse your own answer in a learn session
go deeper
Probes judgment, proportionality and respect for how decisions get made. Escalation is a real tool, and refusing to ever use it is as much a signal as reaching for it constantly. A strong answer proves you can distinguish a decision you dislike from one whose consequences you have no standing to accept, and that you commit fully to whatever ruling comes back.
I led mobile platform work on an enterprise app that warehouse and delivery staff use as their shift tool. As part of moving off our old distribution and sign-in stack, the program decided to enforce a hard minimum OS version on the cutover release. Anything older simply would not launch. I pulled the fleet numbers: 9.3 percent of active devices failed that bar, and they were concentrated in two regions whose warehouses had not been through the hardware refresh. That is people arriving for a shift and finding their scanner app dead. I went to the program's decision owner first with a soft-gate alternative — old devices get a read-only degraded mode and a nag screen for one quarter. He had reasons and he held. So I told him, in that conversation, that I was going to take it to our director because the blast radius was bigger than I had any standing to accept on my own, and I sent the director a one-page note: both options, the support cost of each, what reverting would cost after launch, and my recommendation. She picked the middle: soft gate now, hard gate once the refresh completed. I owned building the degraded mode, which I had argued for, and I never mentioned the escalation again. Rating held at 4.39 through the cutover. The comparable hard gate we shipped the year before had taken it to 4.12 and cost us six weeks of support load.
The proportionality test — blast radius beyond what he could accept alone — is what makes this an escalation rather than a complaint, and telling the decision-owner to his face that he was escalating is the detail interviewers listen hardest for. It would downlevel if the fleet percentage were missing and the stake stayed abstract.
Three field-facing mobile apps in my org were being folded onto a single shared app shell owned by a platform group in another org. I did not own that decision and did not try to. My objection was sequencing: the shell had no offline story, and my crew app is offline for most of a working shift. I worked it at the platform level for about six weeks — a spike, a shared doc, a proposed order of adoption — and got nowhere on funding the offline work. So I escalated once, to the executive over both orgs, as a one-page memo. Two options, the cost of each, what unwinding the migration would cost per app after adoption, my recommendation, and the thing I insist on in any memo of mine: the evidence that would prove me wrong. He kept shell adoption and made offline a shell requirement funded by the platform group, with my app migrating last. That is not what I asked for and it was better than what I had. I wrote the close-out myself, including my original objection, so my staff engineers did not have to interpret my silence. Then I named a migration owner in my org and deleted our duplicate storage layer myself. The standing rule I run is one written dissent, one escalation, then it is the org's decision. Two apps are through; the crew app's rating went 4.26 to 4.33, and the second app dipped to 4.09 during its cutover, which we had predicted and staffed for.
The principal signal is the mechanism rather than the outcome: a memo format that names disconfirming evidence, a stated limit of one dissent and one escalation, and a written close-out carrying his own objection so his org does not have to read his mood. Deleting the duplicate layer himself is the commitment made costly and visible.
You are rarely expected to escalate, and doing it often is a warning sign at this stage. A good answer here is usually asking your manager whether something should go further, and being able to say what you had already tried yourself.
Escalation at your scope means bringing a decision to your manager or a tech lead with options rather than a complaint. Show that you arrived with a recommendation and a cost, and that the person you were disagreeing with knew you were raising it.
You are expected to escalate rarely and well. Show the proportionality test you applied — blast radius beyond your authority to accept — the written form you used, and that you closed the argument the moment a ruling came back, including in front of your own team.
Show the norm, not the instinct: how many rounds of dissent your org gets before a decision is final, what an escalation memo of yours must contain, and how you keep multiple teams executing a call you lost without visible hedging.
saying these in an interview costs you the question
- Escalating without ever raising it with the decision-maker first
- Going around a manager and hiding it from them
- A stake too small to justify going over anyone's head
- Escalating as a habit rather than a rare exception
- Winning the escalation and enjoying the win out loud
- No account of what happened after the ruling came back
- How did the person you escalated past react?This is the real test of the story. Say whether you told them beforehand — if you did not, own that as the mistake it usually is. Then describe the working relationship afterwards in concrete terms: what you shipped together next, whether they brought you into the following decision. Interviewers are checking whether escalation costs you your peers.
- How often do you escalate?Give a rate, not a philosophy — once or twice in a multi-year program is a credible answer, and it lands better than any principle you could state. Then name what does not clear the bar for you: preferences, taste, decisions whose cost you can absorb yourself. Frequent escalation reads as an inability to lose.
- What did you do when the ruling still went against you?Show a hard stop. The strongest version names the exact moment you closed the argument and how you told your own team, including that you had objected — so nobody has to guess where you stand. Avoid any hint that you kept the case alive informally, which is the failure mode this question exists to catch.