Running It: Airtime, Silence, and the Difficult Person

Group discussion has a structural property that undermines its main purpose, and understanding it explains most facilitation technique.

When a question is opened to a room, the first substantive contribution does disproportionate work. It establishes what kind of answer is expected, which considerations are relevant, and what the range of acceptable positions looks like. Everyone who speaks afterwards is responding to it, whether they agree or not.

That is anchoring, and it is one of the most robust findings in judgement research, going back to Tversky and Kahneman in 1974. Numerical estimates shift toward whatever number was mentioned first, even when the number is obviously arbitrary, and knowing about the effect does not remove it.

In a meeting the anchor is not usually a number. It is a framing, and it is set by whoever speaks first, which is systematically the most senior, most confident, or most extroverted person present rather than the most informed.

The compounding problem. Once two or three people have expressed a view, the cost of disagreeing rises. Solomon Asch's conformity experiments in the 1950s showed people denying the evidence of their own eyes on a trivial perceptual task when a unanimous group said otherwise. Real meetings involve ambiguous questions and colleagues you will see tomorrow, where the pressure is considerably stronger.

So a discussion does not aggregate independent views. It aggregates one view and a series of responses to it, which is a much weaker instrument.

Which yields the single most useful principle in facilitation. If you want a group's independent judgement, you must collect it before people hear each other. Everything in the next step follows from that.

The family of techniques that follows from anchoring, all of which do the same thing: separate the generation of views from their exchange.

Silent writing first. Pose the question, then everyone writes their answer alone for three minutes before anyone speaks. This is the highest-value two minutes available in any meeting. You get independent positions rather than variations on the first one, quiet people arrive with something written, and the range of genuine disagreement becomes visible immediately.

Round-robin. Go around and hear from everyone once before open discussion. Slower and it guarantees that the eighth person's view enters the room, which in open discussion it would not.

Reverse the order of seniority. If the senior person speaks last, everyone else has already committed to a position. If they speak first, the meeting is about their view. This is a small change with a large effect, and senior people who understand it usually adopt it willingly.

Anonymous collection for anything sensitive. Where honest input carries a cost, the only way to get it is to remove attribution.

Written pre-reads with comments, which the previous lesson covered, and which are silent-writing-first done in advance.

All of these trade a few minutes of efficiency for a substantial improvement in what the group actually knows. And the crucial point is that the cost is visible while the benefit is not: the meeting feels slower, and the alternative was a fast meeting that surfaced one person's view.

One further reason to use them. They change who contributes. People who think before speaking, who are less senior, or who are working in a second language contribute far more in writing than in open discussion, and their input is not systematically worse.

Airtime in an unmanaged meeting distributes badly and predictably, and a facilitator's main job is correcting it.

The default state. A few people speak most of the time. This correlates with confidence, seniority, extroversion and comfort in the dominant language, and only loosely with having something useful to say. Meanwhile several people say nothing at all, and their silence is read as agreement when it frequently is not.

The interventions, from lightest to heaviest, because using the heaviest first damages the room.

Redirect by invitation. Ask a specific quiet person a specific question. Not any thoughts, which invites a pass, but you worked on this last year, what did you see. Naming the reason gives them standing to speak.

Bound the talkers by structure rather than by correction. Let us hear from people who have not spoken yet. This constrains without singling anyone out.

Use the written round from the previous step, which sidesteps airtime competition entirely.

Go direct and private if it persists. A quiet word afterwards works and a public correction humiliates, and the second costs you the room.

And notice the silence rather than accepting it. Someone who has not spoken for forty minutes is either disengaged or disagreeing quietly, and both are worth knowing about before the decision is made rather than afterwards when it fails to be implemented.

The general point. Airtime is the resource being allocated, and if the facilitator does not allocate it, the room allocates it by status.

Meetings contain people who derail them, and the useful move is diagnosis rather than management, because the behaviours have different causes and different remedies.

The repeater, who makes the same point several times. Almost always someone who does not believe they were heard. The remedy is to demonstrate hearing: write their point down visibly, say it back, and confirm you have captured it. Repetition usually stops immediately, because it was serving a function.

The re-opener, who returns to a settled question. Sometimes they were absent when it was decided, sometimes they never agreed and it was never surfaced. Ask which: is this new information, or is this a disagreement we did not resolve? Both are legitimate and they need different responses.

The tangent-taker, following interesting but irrelevant threads. Capture it visibly in a parked list and return to the agenda. The visible capture is what makes it work; a promise to come back to it does not.

The blocker, who objects to everything without offering an alternative. Ask for the alternative directly and specifically. If none comes repeatedly, that is worth a private conversation, because it is usually about something not on the agenda.

The dominator, covered in the previous step.

And the quietly-disengaged, who is the most consequential and gets the least attention. They have stopped contributing, will not implement what is decided, and nobody will find out until later.

The pattern across all of these. Disruptive behaviour in meetings is usually a person with an unmet need: to be heard, to have an objection addressed, or to understand why a decision went against them. Managing the behaviour without addressing the need moves it elsewhere, usually to after the meeting where you cannot see it.

A room that agrees quickly feels like a well-run meeting and is frequently the opposite, and this failure mode is worth recognising because it is invisible from inside.

Irving Janis described groupthink in 1972, studying policy disasters including the Bay of Pigs. His account identified symptoms including an illusion of unanimity, self-censorship of doubts, pressure on dissenters, and an illusion of invulnerability. The specific historical analysis has been debated and refined considerably since, and the core observation has held up: cohesive groups under pressure converge prematurely and suppress the doubts that would have corrected them.

The practical signature in an ordinary meeting. A proposal is made, two people support it, nobody objects, and it is agreed in four minutes. That can mean it was obviously right. It can also mean nobody wanted to be the one to object, which produces exactly the same observable behaviour.

Distinguishing them is the facilitator's job, and there are cheap ways.

Ask for the case against, explicitly and as an assigned task rather than as an invitation. Someone is asked to argue the opposite, which removes the social cost because they are performing a role.

Ask what would have to be true for this to be a mistake. Easier to answer than do you disagree, because it does not require anyone to position themselves.

Run a pre-mortem, which is Gary Klein's technique: imagine it is a year from now and this failed badly, and write down why. This is markedly more effective than asking for risks, because it presupposes the failure and removes the need to predict it.

And watch for the silent. Unanimity in a room where three people said nothing is not unanimity; it is three unrecorded positions.

The underlying point. A decision nobody objected to and a decision everyone agreed with look identical in the moment and behave very differently during implementation.

Working through genuine disagreement is the thing meetings are best at and the thing facilitators most often mishandle, usually by trying to resolve it too quickly.

The first move is diagnosis, because disagreements have types and mixing them produces circular argument.

Disagreement about facts. We believe different things about the world. This is the easiest, because it is resolvable in principle: identify what would settle it and go and find out, rather than continuing to argue.

Disagreement about predictions. We agree on the facts and expect different futures. Not resolvable now, so the useful move is to make it concrete: what would we expect to see in three months under each view, which converts an argument into a test.

Disagreement about values or priorities. We want different things. This is not resolvable by evidence at all, and treating it as factual is why some arguments recur for years. Naming it as a priority difference moves it to whoever decides priorities, which is usually a legitimate and quick resolution.

And disagreement about something unsaid. Someone's real objection is about resourcing, credit, or a previous decision, and it is being expressed as a technical concern. This is the most common and the least addressed.

The facilitation moves that help.

Ask each side to state the other's position to that side's satisfaction, before arguing. This is slow and it repeatedly reveals that people were disagreeing about different things.

Write the two positions where everyone can see them, which converts a contest between people into an examination of options.

And know when to stop. Some disagreements should be escalated, some should be tested rather than argued, and some should be noted and set aside. A facilitator who forces resolution on an unresolvable disagreement produces false agreement, which is worse than the disagreement.

Remote meetings have different failure modes, and hybrid ones are considerably worse than either pure format, which is worth stating because organisations keep defaulting to hybrid.

What is harder remotely. Reading the room, since you cannot see hesitation or disagreement in body language. Interrupting, because the turn-taking cues people use are absent or delayed, so people either talk over each other or wait too long. Informal contribution, since the side remark that adds something is not available. And sustained attention, because the meeting is in the same place as everything else on the screen.

What is easier remotely, and it is underused. Simultaneous written contribution, which is genuinely better than in person: everyone can type at once, which removes the airtime problem entirely. Chat as a parallel channel, where quieter people contribute more. And a natural fit with the written-first techniques from earlier in this lesson.

Why hybrid is worse than either. The people in the room have a different meeting from the people at home. They can see each other, react to each other, and continue talking before and after. Remote participants get a partial feed, are slower to enter a conversation, and are systematically forgotten.

The interventions that actually help hybrid, in order of effectiveness.

If anyone is remote, everyone joins individually from their own device, including the people sitting together. This equalises the experience and is the single most effective fix.

If that is impossible, assign someone in the room to watch the remote channel and interrupt on their behalf, as an explicit role rather than a hope.

Go to remote participants first in any round, since they are the ones who will otherwise be missed.

And put everything in writing, because the shared artefact is the only thing both groups genuinely have equal access to.

The end of a meeting determines whether it produced anything, and it is where meetings are most consistently mismanaged, because the time is spent on whatever the last agenda item was.

What must happen before people leave.

State the decisions, out loud, in the words that will be written down. Not we agreed to move forward, but we are proceeding with option B, and we are not doing C this quarter. Saying it out loud repeatedly surfaces that two people understood the discussion differently, which is far cheaper to discover now than in three weeks.

State the actions with an owner and a date. An action without a named person is not an action, and we should all is nobody. This is the single most common failure in otherwise good meetings.

Name what was not decided, and what happens next with it. Unresolved items disappear silently otherwise, and reappear as surprises.

Check for the unspoken objection, once. Is anyone leaving this room with a concern they have not said? Asked directly, at the end, it occasionally produces the most important contribution of the meeting.

And send the notes quickly, the same day where possible. Notes sent a week later are archaeology; notes sent within hours are the record everyone works from, and whoever writes them has substantial influence over what the meeting is understood to have decided.

The practical way to protect this. Reserve the last five minutes and defend them, which means stopping the discussion while it is still going. That feels wrong and it is correct: a discussion that ran five minutes longer and produced no recorded decisions has produced nothing, and a decision nobody wrote down did not happen.