How do you prioritize which two of your prepared questions to ask in a round's last five minutes?

## The window is a triage problem, not a memory test At the end of an interview round the interviewer typically hands the floor over with something like "what questions do you have for me?" What follows is short — often about five minutes, sometimes less if the round ran long. The candidate has a prepared card of roughly four questions. The craft is deciding, live, which two get asked. ## Step one: size the window out loud Before the first question, ask how much time there is: "I have a couple of questions — do we have about five minutes?" This costs eight seconds and buys three things. It tells you whether to ask two questions or one. It signals that you know the answers take time and you are not going to run over. And it occasionally buys you more time than you would have taken, because the interviewer says they have longer. If the round has run long and there are two minutes, say so and adapt: "we're short on time, so I'll ask the one that matters most." Cutting deliberately reads far better than starting a four-part question with ninety seconds left. ## Step two: rank on two axes The card is already ordered by decision value from the night before. Live, re-rank it against two tests: - **Exclusivity — can only this person answer it?** A question the recruiter can answer, and so can three other people later in the loop, is a bad use of the recruiter's five minutes. Spend the slot on what dies with this round. - **Decision weight — would the answer change what you do?** Not "is it a good question" but "if the answer is the bad version, do I behave differently?" A question whose every possible answer leaves you in the same place is entertainment, not triage. Cross out anything already answered during the round. What remains, in order, is what you ask. ## Step three: spend the second slot last Ask one question. Listen to the whole answer. Then decide the second slot in the moment: the next-ranked prepared question, or a follow-up on what you just heard. The follow-up often wins, because a specific probe into a real answer gets you further than a fresh general question — and it demonstrates you were listening rather than reading a list. This is why two, not four. A worked shape for a five-minute window: about forty seconds for the question and its framing, ninety seconds to two minutes for the answer, then the same again for the second slot, leaving a little room to close. Four questions in the same window means either the interviewer is rushing every answer or you are cutting them off. ## Worked example: the recruiter round on a mobile role Your card for the recruiter has four lines, ranked: (1) how the loop is sequenced and who you would meet, (2) how the role was levelled and what the mobile team wants that it lacks today, (3) whether the team is building new surfaces or maintaining what ships, (4) how long the role has been open. During the call the recruiter volunteers the sequencing while explaining scheduling — strike line one. That promotes line two, which is also exclusive: the recruiter knows how the role was written and levelled, and the engineers later in the loop mostly will not. You ask it. The answer mentions that the team has been shipping fixes rather than features for a while. Your second slot is now obvious: rather than reading line three off the card, you follow the thread you were just handed and ask what changed. That is one prepared question plus one earned follow-up — two asks, five minutes, and you leave knowing something you could not have read anywhere. ## Timing inside the round, not just at the end The window is not always at the end. If an interviewer opens the floor mid-conversation — "any questions so far?" — take it, and ask the question that is most relevant to what you were just discussing; it costs you nothing, because the end-of-round window usually still exists. Conversely, if you are asked at the very start whether you have questions, ask one short logistical one and hold the rest; the good questions depend on what the round reveals. ## The two ways this goes wrong The first is racing: asking all four because they were prepared, which converts a conversation into a quiz and produces four one-line answers. The second is the opposite — freezing when the top line is already answered, and filling the silence with something improvised and vague. A ranked card with strike-throughs is the defence against both: there is always a live top line, and you never need more than the top two.