Watch the last five minutes of most conference keynotes and you will see the same small tragedy performed on schedule. The MC checks the clock, announces there is "time for a couple of questions", and the room enters a silence with a very particular texture — the silence of two hundred people waiting for somebody else to go first. Eventually a hand rises, somewhere near the back. The question rambles. The answer, compressed by the clock, restates something the talk already said. Then everyone breaks for coffee, and the agenda records that the session included Q&A.
It did not. It included a formality. Five minutes tacked on the end buys exactly what it pays for: one rambling question and one non-answer — plus, if the day is running true to form, a colleague who stands, thanks the speaker warmly, and delivers a three-minute statement containing no question at all. None of this is the audience's fault. It is a design failure. Q&A is a format, with its own craft, its own timing and its own failure modes, and most events treat it as the packing foam around the real content.
Give it a real share of the session
Our working rule: Q&A deserves roughly a quarter of the session. A 40-minute slot runs about 30 minutes of talk and 10 of properly managed questions; an hour-long session earns 15. Write it into the run sheet as a segment with a start time and an owner, not as whatever remains when the speaker finishes.
The arithmetic changes behaviour on both sides of the lectern. A speaker who knows a quarter of the session belongs to the audience builds a tighter talk — the digressions get cut in rehearsal because the clock demands it, and the ideas most likely to provoke questions get sharpened rather than hedged. The audience listens differently too. People attend to an argument they know they will get a genuine chance to test. And when the inevitable overrun comes, defend the segment. Q&A is the first thing sacrificed when a talk runs long, and every time it happens the room learns precisely where its questions rank.
Moderate it, or gamble it
The open microphone is roulette. Whoever reaches the mic first sets the tone for everything after, and the first question is disproportionately likely to be a niche grievance, an internal in-joke, or that speech in disguise. The fix is moderation — not as gatekeeping, but as advocacy. A good moderator is the audience's voice: they ask the question behind the question, cut the statement-dressed-as-question off with charm, redistribute airtime away from the serial hand-raisers, and push back when the speaker's first answer glances off the point.
Two techniques do most of the work. First, seeded starter questions — two or three prepared in advance, drawn from the briefing and from what the audience is actually facing, deployed so nobody endures the dead air while the first hand deliberates. Once the ice is broken the real questions come; the seeds exist to break it. Second, collection and curation. Gather questions in writing — during the talk, or in advance of the event — and let the moderator sequence them: group the themes, retire the duplicates, put the sharpest one early. Curation is not censorship, and the distinction matters. Keep at least one genuinely hard question in the mix, because a Q&A with no teeth reads as staged, and the room always knows.
Placement — and when to skip it entirely
Position matters nearly as much as duration. Take questions after the body of the talk, not sprinkled through it; interruptions break the momentum a good speaker spends twenty minutes building. Let the talk land, then interrogate it. The instructive exception is the interview-style session, which is really Q&A promoted to the whole hour — every minute is questions, held together by a professional asker. That is why firesides work so well for the accomplished guest who has no forty-minute talk in them, and it is the strongest proof that questions can carry a session rather than merely close one.
Just as important is knowing when to skip it. After-dinner speaking is entertainment: the room is loose, the wine has circulated, and a roving microphone at 9.40pm is a risk with no upside. Awards nights run on momentum and celebration — names read correctly, winners applauded, nobody detained. Bolting Q&A onto either is a category error. Part of treating questions as a format is respecting the formats they do not belong in.
Virtual changed the economics of asking
Here is the quiet lesson of the virtual years: chat questions often beat raised hands. People ask better questions typing than standing. There is no adrenaline spike, no audience for their nerves, and a few seconds to edit — so the questions arrive shorter, sharper and more honest, and they come from the people who would never approach a microphone in a ballroom. The written channel also hands the moderator a full queue to curate in real time, rather than a gamble on whoever is nearest the aisle.
The lesson travels back into the physical room. Collect written questions there too — cards on tables, a screen, a phone — and let the shy majority into a conversation the confident minority has always owned. The hands-up ritual selects for confidence. The written question selects for curiosity, which is the trait you actually want surfaced.
The case for all this care is simple. The talk is what you planned to say. The questions are where the room tells you what it actually heard — which idea landed, which one confused, what your people are quietly worried about and would never volunteer in a survey. That intelligence is sitting in every audience you assemble, free, waiting for ten designed minutes. An event that skips the questions is not saving time. It is choosing not to know.