What changed since the first round
Two things, and both point the same way.
The build round happened. That was the technical-credibility screen, and the re-scope had already taken most of its teeth out. Whatever you said there is now fixed and quotable. If these two compare notes — and on a panel of two for one seat, they will — the thing they are most likely to compare is how you described the boundary of the role.
Where the role lives is confirmed unsettled. Not your inference this time: the build lead said it directly. That is more useful than either answer would have been. A settled reporting line would have told you which case to argue. An unsettled one tells you the seat itself has not been defined, and that whoever offers a workable definition first is likely to get it.
Why that is leverage rather than a worry
Right now you can shape what the role owns and hear it confirmed out loud by the people who would hold you to it. After you accept, the same conversation costs political capital you have not yet earned. This is the cheapest week of your tenure to ask what the job is.
The one real risk
The role as re-scoped is requirements, strategy, design and some UI/UX. Four of those words already sit on this interviewer's org chart: her team owns strategy, design, process mapping and stakeholder management, upstream of every build.
So her hour is quietly an audition for one question — what does this person do that my team does not already do? Nobody asks it in those words, and a hiring manager will rarely say out loud that the answer was thin. It comes back later as a lukewarm debrief with no specific objection in it, which is the hardest kind to recover from.
The overlap is a property of the work, not of the org chart. It exists whichever team you land in. Only the mechanism for resolving it changes — a shared manager, a peer negotiation, or a formal hand-off.
The move
Defer on the relationship, claim the artefact. Give her the thing her team most needs to keep — the business problem, the process design, and the stakeholder — and claim something she does not currently own end to end: the layer between an agreed design and an engineer's estimate.
That layer is real and it is unglamorous, which is exactly why it is available. A process map says what should happen. It does not say which field is authoritative, what the acceptance criterion is, or what the screen does when the source system is a day behind. Someone closes that gap today by absorbing it on top of their actual job, or nobody closes it and the build team guesses.
Because placement is open, the line should end in an offer rather than a question. The first-round version closed by asking where the boundary sits. This one proposes it — lightly, and framed as taking load off her rather than taking scope from her.
“Since it's still open — the version I'd argue for is that your team keeps the business problem, the process design and the stakeholder relationship, and I take the layer underneath: the spec, the data definitions, the prototype, the interface. Not because that's the better job, but because it's the layer your team has to produce on top of everything else, and it's the one where a developer still has to guess. If that gap already gets closed here, that's genuinely worth knowing today — it would change what I'd be most useful doing.”
- “Since it's still open” — acknowledges the fact she and the build lead both know, so nothing you say next reads as presumption.
- “your team keeps … the stakeholder relationship” — the concession, said first and without being asked. It is the single thing most likely to make her relax.
- “the layer underneath” — a claim on an artefact rather than on territory. Owning a deliverable is what separates a role from a routing function.
- “on top of everything else” — frames your scope as load you remove from her team, not scope you take from it.
- “that's genuinely worth knowing today” — makes it a proposal she can decline rather than a claim she has to defeat. It also happens to be true, which is why it survives follow-up.
Rehearse it until it comes out the same way twice, in under forty-five seconds. The read page has the variants for when she pushes back, and the drill bank has the four position questions it has to survive.
The consistency constraint
Whatever you told the build lead about where your role starts and stops is now the fixed point. If this interviewer hears a materially different boundary, that is the documented failure mode for a two-person panel, and it is the one you can eliminate entirely with an hour of rehearsal.
Practically: before you refine anything on this site, write down what you actually said in the first round, in your own words. If it drifted from the prepared version, keep your version and adapt everything here to match it. A slightly weaker line said twice beats a stronger line that contradicts the first one.
The asymmetry worth knowing
The two of them fear opposite things. The build lead fears you will over-promise upstream and hand down fuzzy work. The strategy lead fears you will bypass her process and own her stakeholders. A line that reassures one can alarm the other — “I'd own the stakeholder relationship” is exactly what a build lead likes to hear and exactly what a strategy lead does not. Check every sentence against both.
Your three things
Three things you want her saying afterwards, each landing at least twice in different answers. Re-cut from the first round for this interviewer specifically.
Makes her team's work survive contact with engineering
Not a second discovery function. The person who takes an agreed design and produces something estimable and testable.
Settles design arguments in a prototype, not in a sprint
Prototypes to de-risk her design rather than to replace it, and knows a prototype is disposable.
Will not surprise her
Stakeholder conflict routes through her. Disagreement gets raised once, privately, then dropped. No relitigating in front of the business.
The four days
You already have more material than you can deliver. The gap is not knowledge — it is having concrete stories with numbers in them, and having said the position answer out loud enough times that it survives an interruption. Reading more is the comfortable option and the least useful one.
Reconstruct round one, then six stories
First, write down what you actually said about the boundary and anything she may hear second-hand. Then fill the six slots, and go and find the real figures for the six numbers. The redesign story is the one that matters most here and it is the one most likely to be thin.
Done when: five of six slots filled, at least four with a specific number in them, and every number has a source you could state out loud.
Position, domain and design
The read page end to end, then the six position drills out loud. Then the learn page — accruals, as-was versus as-is, and AI under controls are the three that make you sound like an insider. Then the design page, because she will weight the UI/UX portion more heavily than the title suggests.
Done when: the position answer comes out the same way twice, under forty-five seconds, without notes — and you can define an accrual and a true-up cold.
Cold mock
All the drills, random order, out loud, timed, answered before revealing. Rate honestly. Then re-drill only the reds. If someone can ask them at you with follow-ups, do that instead — it surfaces the answers that only work when you set them up.
Done when: every card has a rating and you have written down the three that fell apart.
Taper
Thirty minutes on amber and red only, then commit to the five questions you will ask her, then stop. Nothing new the day before — new material displaces what has already settled and adds nothing you can retrieve under pressure.
Done when: you have stopped. Sleep is a study technique.
What not to do
Do not sell strategy. It is her word. Every minute you spend proving you can do strategy is a minute spent proving you duplicate her team. Sell the layer underneath and let her conclude the strategy competence from how you talk, not from you claiming it.
Do not frame the re-scope as a relief. “The role moving away from the technical side suits me” reads, to anyone downstream of you, as someone glad to be further from delivery. Frame it forward: this is the work you are best at and it is the work the seam actually needs.
Do not negotiate the org chart. Offer a view of what the role owns, once, and then let it go. Pressing a second time converts a proposal into a demand, and you are proposing it to the person whose team currently holds the ground.
Do not point out that the two of them describe the boundary differently. If that happens, note it and keep it. It tells you the first three months will be negotiation, and it is a fair thing to raise with the hiring manager at offer stage — in the scope paragraph, in writing, not as a handshake.
Read the interviewer →
What she screens for, her ten questions, and the boundary answer with its pushback variants.
Six stories →
The slots that cover her whole question surface. Start here on day one.
Learn →
The six things worth actually knowing — accruals, headcount definitions, AI under controls, and where eBay is right now.
Drills →
Twenty-four questions with model answers, timed and rated.
Design →
Finance interface principles and how to walk through a redesign.
Ask →
What you ask her, sequenced — and what each answer actually tells you.