Cheat Sheet: Leadership Round

Read the transcript

1. Prepare the night before, in ADR shape

Host: So you know the leadership round is coming, and the temptation is to think you can wing it — you lived these decisions, right? But this round isn’t scored on what you did, it’s scored on whether your account of it holds together when someone keeps pushing on it. That’s a different skill, and it needs real prep, not just a quick recall in the lobby before you walk in.

Guest: Right, and the prep that actually works is doing it the night before, in something like the ADR shape — Context, Problem, Options, Decision, Consequences. Pick two or three decisions that were genuinely yours and write them out that way. Options is where it gets uncomfortable, and it’s the one that does the work.

Host: Because that’s where you have to remember what you rejected and why, not just what you chose.

Guest: Exactly — if you can’t reconstruct the alternatives you turned down, that’s a huge signal, and you want to discover it alone at your desk, not in the room. Better to find the gap in private than have it found for you in the room.

2. Build the answer to survive follow-up: constraint, decision, rejected options, cost, outcome, mind-change

Host: So say you’ve done that private audit and you’ve got your two or three decisions in ADR shape. How do you actually structure the spoken answer so it doesn’t fall apart when they start poking at it?

Guest: There’s a sequence, and each piece exists to survive a specific probe, not to sound polished. Start with the constraint — a number or a named party, not an adjective, because ‘tight timeline’ invites a follow-up you can’t answer. Then the decision, and you say it was yours, not the team’s, because an undivided ‘we’ just gets probed and that probing eats your time. Then the options you rejected and why — that’s the piece that turns it from a default into an actual decision — followed by the cost, which is the cheapest credibility in the room and almost nobody names it. Close with the outcome as a number that isn’t a duration, because six months of migrating is effort, not impact, and finally what would’ve changed your mind, and if it did change, how long the gap was between the evidence showing up and you acting on it. Then you stop talking and let them dig.

Host: And that last one about the gap — that’s clearly not a one-off question, that’s a whole category, right? Reversed decisions, disagreements you lost, things you standardized.

Guest: Right, and that’s why you want a handful of situations loaded four questions deep, not eight shallow ones. A disagreement you lost is really testing whether you can separate being overruled from being wrong. A decision you reversed is testing that evidence-to-action gap again. Something you standardized is testing whether you did it where it compounds — identity, observability, deployment — or just on personal taste. Three decisions you actually owned beats eight you can merely describe, and if you catch yourself reaching for a project where you were just nearby, that reach is the finding.

3. The tells that give you away

Host: So let’s make it concrete — what are the actual tells that give an interviewer the sense they’re listening to a sanded-smooth answer? What should someone listen for in their own practice runs?

Guest: Start with no stated cost — if the decision cost nothing, it was trivial, and if it clearly wasn’t trivial but you can’t name what it cost, you’ve smoothed it over. Next is the undivided ‘we,’ which reads as either modesty or concealment and the interviewer genuinely can’t tell which, so you have to say what you specifically did. Watch for depth running out one question early — the fix there is fewer examples prepared deeper, not a more polished summary. ‘We discussed it and aligned on the best approach’ is the single most common non-answer, so ban it from your own mouth. And impact measured in effort — duration, headcount, lines changed — is a tell you’re describing motion, not judgment; same with claiming you’ve never been wrong, or not being able to say how long it took you to notice you were. Really, recall under follow-up is a different task from recall itself, and the difference shows within four minutes — so don’t treat this as the round that needs no preparation.

Generated from this page by Claude Sonnet 5 on 2026-08-23, spoken by Kokoro-82M running locally. Two synthetic voices, not a recorded conversation. Every claim is drawn from this page — where it differs from the text above, the text is correct.

Use This When

You are about to be asked about decisions you have made. Read it the evening before, not in the lobby — the work is checking which decisions you can actually go deep on, and that check takes longer than the reading.

Depth lives in Track: Technical Leadership and Module 14.

The Sequence

Preparing — write two or three of your own past decisions in the ADR shape: Context, Problem, Options, Decision, Consequences. Options is the one that does the work. If you cannot reconstruct what you rejected and why, you have found that out in private rather than in the room.

Answering — the interviewer is checking internal consistency, so build the answer so it holds:

Constraint. What had to be true — with a number or a named party, not an adjective.
The decision, and that it was yours. Say what the team did, then what you decided. Undivided “we” gets probed, and the probing eats your time.
The options you rejected, and why. This is what separates a decision from a default.
The cost. Every real decision traded something. Naming it is the cheapest credibility available and most candidates skip it.
The outcome, in a number that is not a duration. “Six months migrating” is effort. What changed is impact.
What would have changed your mind — and, if it later did, how long the gap was between the evidence arriving and you acting.

Then stop and let them dig. Depth on their question beats breadth on yours.

Quick Reference

Have these loaded, four questions deep each

Situation	What it samples
A decision hard to reverse	Whether you recognized it as such at the time
A disagreement you lost	Whether you can separate being overruled from being wrong
A decision you reversed	The gap between evidence arriving and you acting
Something you standardized	Whether you standardized where it compounds, or on taste
An incident you led	Whether the outcome was a mechanism or a culprit
Quality degrading quietly	Whether anyone owned the quality signal
Work you handed off	Whether you delegated decisions or only tasks

Three decisions you owned beats eight you can describe. If a row makes you reach for a project where you were merely nearby, that is the finding.

Disagreement, decomposed — constraint (resolve it first; most arguments end here) · conclusion (name the evidence that would settle it and what it costs) · preference (two-way door — concede and spend the credibility elsewhere).

Standardize where it compounds — identity, observability, deployment. Say what you deliberately left alone.

Red Flags

An account with no stated cost. Either the decision was trivial or it has been sanded smooth.
Undivided “we” — reads as modesty or concealment, and the interviewer cannot tell which.
Depth that runs out one question early. The fix is fewer examples, not more polish.
“We discussed it and aligned on the best approach.” A non-answer, and the most common one.
Impact measured in effort: duration, headcount, lines changed.
Treating this as the round that needs no preparation. Recall under follow-up is a different task from recall, and the difference shows within four minutes.
Claiming you have never been wrong, or being unable to say how long it took you to notice.