Scope, and the level you get offered

Here is the thing that makes behavioral rounds feel arbitrary. Two candidates tell the same quality of story, with the same structure and the same specificity, and one gets a hire while the other gets a down-level. Nothing went wrong in the second interview. The stories were smaller.

Read the rubric ladders and it is unmistakable. The question is the same at every level; only the scope of an acceptable answer moves. For proactivity: a junior story is a change affecting their team's focus area, "usually only requiring the candidate themselves to work on"; senior is a change affecting the whole team, "usually requiring three or more people"; staff is a change affecting the org, "usually requiring two or more teams."

WatchHello Interview · 1:06:48

Two chapters from the long conversation cued in the first lesson of this module, picked because both are levelling questions wearing a disguise. Listen for how the managers separate a candidate who owned a change from one who was present for it — that distinction is the whole subject of this lesson.

The common behavioural questions, broken down by ex-Meta and ex-Amazon senior managers Play “The common behavioural questions, broken down by ex-Meta and ex-Amazon senior managers” on YouTube

Scoring your inventory

There are two axes, and they are not symmetric. Headcount ladders all the way up. Team count only separates staff-scale work from everything else — junior and senior stories are both single-team, so teams behave as a gate rather than a ladder.

Now score your stories against a target. Take Staff as the target:

Zero of four clear staff outright, and that is the normal result for a strong senior engineer interviewing one level up. These are not bad stories. The flaky test is a real piece of work. It is just, unambiguously, a junior-scoped anecdote, and telling it beautifully will not change that.

The interesting row is the migration, where the two axes disagree: six people reads senior by headcount, two teams reads staff by breadth. Which axis the interviewer uses is not up to you — but which axis you foreground is:

"It was a small team by headcount — six of us — but the awkward part was that the write path belonged to the platform team and the reconciliation to finance eng, so most of the work was getting three groups to agree on a cutover order none of them individually wanted."

Same facts, and now the breadth is the subject of the sentence rather than a detail the interviewer has to infer.

Raising a story's read, honestly

Three moves, in descending order of how often they apply.

Say the breadth you actually had. Most people describe the work and omit its surface area entirely, because the surface area was obvious to them at the time. Who else had to agree. Whose roadmap moved. Which team's on-call changed. This is the most common free upgrade, and it's free because you're adding facts you left out rather than facts you invented.

Name the durable artefact. A one-off fix reads at the scope of the fix; the thing that outlived it reads wider. "…and the runbook became how the team does cutovers now" or "…the check we added catches this class of bug for everyone." Staff-scope signals are frequently about what persisted.

Say who you convinced, not what you built. At senior and above, the scarce skill isn't implementation, it's influence without authority — getting people who don't report to you to change what they're doing. If a story has a persuasion spine, tell that spine. "The hard part wasn't the schema, it was that two teams had already committed to the old one in their quarterly plans."

Failure, at scope

"Tell me about a time you failed" is the question where scope and honesty pull hardest against each other. Pick something small and you look like you've never carried anything; pick something enormous and you're arguing for your own unreliability.

The resolution is that the failure should be real and the response should be what carries the scope. What the question collects is growth, and growth is evidenced by what changed afterwards.

• It has to be an actual failure. A story where the project succeeded, or the failure was someone else's, or the failure is a disguised strength ("I cared too much about quality"), is a non-answer — and it is the most-seen non-answer in the round, so it is recognised instantly.
• Own the part that was yours. Not everything, but not nothing. "My cutover plan assumed the retries were idempotent. I hadn't verified it, and forty accounts were double-charged."
• Then the response, at the level you want. A junior-scoped response fixes the bug. A senior-scoped response fixes the class of bug for the team. A staff-scoped response changes how decisions get made — the review step that now exists, the assumption that now gets written down.
• Close the loop. What happened after the change. Did it recur? "That check has caught two more since." A learning with no follow-up evidence is a claim.

Conflict, at scope

Same structure, different signals — conflict plus empathy, and the scope ladder runs from implementation detail with a coworker, to direction with team leads, to direction across two or more teams.

The single highest-value habit here is to state the other side's position before your own, in terms they'd accept:

"He'd been on-call through the last two migrations and had watched cutovers go badly when they were rushed, so his position was that we should parallelise the existing job and not touch the write path at all. That's a completely reasonable read of the risk."

Then your reasoning, then how it was resolved — and "resolved" does not have to mean you won. A story where you changed your mind for a stated reason is a strong answer, and it's one many candidates never consider offering.

The questions you ask

The last five minutes are yours, and they're not scored as a formality. Question quality is information about what you care about, and generic questions ("what's the culture like?") return generic answers you could have read on the careers page.

What works is specificity about the work you'd actually be doing: how the team decides what to build next, what the on-call load genuinely looks like, what the last thing to slip was and what happened, what a successful first six months would contain. Asking your interviewer something that draws on their own experience — how their scope changed, what surprised them — tends to produce the most honest answers, because you're asking them about something they know rather than something they're supposed to represent.

And this is not only signalling. You are also deciding whether to work there, and the round is one of very few chances to ask someone who does the job what the job is.

What this module gave you

Four things that decide the round before any particular question: what is being collected and how the recommendation gets assembled; where the four minutes of an answer go and which three sentences carry it; an inventory built by coverage rather than count; and a scope read on every story, so you know which level your material honestly supports.

What comes next in this track is the material itself — the specific answers for each signal area, the company-values variants of these questions, and the harder cases: managing up, being managed badly, work you couldn't ship, and gaps or exits you'd rather not explain.