Most Meetings Fail Before Anyone Speaks

The arithmetic nobody does, and it changes how you treat the invitation list.

An hour with eight people is eight hours. At any professional salary that is a substantial expense, and it is incurred whether or not anything is decided. Organisations that would require three approvals to spend the equivalent in cash spend it on a recurring meeting nobody has questioned in two years.

The cost is also larger than the hour. Work adjacent to a meeting is fragmented: the twenty minutes before it are not usable for anything demanding, and returning to concentrated work afterwards has its own cost. A meeting in the middle of the morning does more damage than its duration suggests, which is why a calendar with four scattered meetings can consume a day.

None of that argues for fewer meetings as a general policy. Some of the most valuable work in any organisation happens when people are in a room together, and the alternative to a meeting is frequently a long email thread that costs more and decides less.

What it argues for is that the meeting should be a deliberate choice with a stated purpose, rather than the default response to any unresolved question.

The practical test, which is unpopular and effective. For any recurring meeting, ask what would happen if it were cancelled for a month. If the answer is that people would find another way to get what they need, the meeting is a habit. If the answer is that something specific would break, name that thing, because it is the meeting's actual purpose and it is probably not what the agenda says.

Meetings fail most often because they are trying to do several incompatible things at once. Separating them is the single largest improvement available.

Deciding. A choice must be made, and the people who can make it are present. This needs options, information, and someone with authority to conclude.

Generating. Producing ideas, options or possibilities. This needs divergence, psychological room to say something half-formed, and explicitly no evaluation.

Aligning. Making sure people share an understanding of something: a plan, a change, a situation. This needs explanation and questions, and it is the one most often achievable without a meeting.

And connecting. Building the relationships that make everything else work. Genuinely valuable, frequently disguised as one of the other three, and better when it is honest about itself.

Why mixing them fails, and this is mechanical rather than a matter of discipline.

Generating and deciding are directly opposed. Generating needs people to offer bad ideas safely; deciding needs them to be evaluated critically. Do both in one session and the evaluation kills the generation, usually within the first three contributions.

Aligning inside a decision meeting means the people who needed the context spend the meeting acquiring it while the people who had it wait, and the decision gets made by whoever arrived informed.

The practical rule. Name the type in the invitation, and if a session genuinely needs two, separate them in time with an explicit transition: we are generating for twenty minutes, we will not evaluate anything, then we will break and come back to choose.

Saying that out loud does more work than any facilitation technique.

A decision path that removes a substantial share of meetings before they are scheduled.

Start with what you actually need. If it is to transmit information, write it. A document is faster to consume than a presentation, can be read at the reader's pace, is searchable, and reaches the people who could not attend. Status meetings are the largest single category of meeting that should be a document.

If you need input rather than transmission, ask whether it needs to be simultaneous. Comments on a document gather more considered responses from more people, including the ones who think slowly and the ones who will not speak up in a room.

If you need a decision, ask who actually makes it. If one person decides, they need input rather than a meeting, unless the discussion itself will change their view.

Meetings are genuinely right for four situations. Where disagreement needs to be worked through, because that requires people responding to each other in real time. Where the topic is sensitive and tone matters. Where you need genuine generation, which is better with the energy of a room. And where relationships are the point.

The pattern worth noticing. Most meetings exist to transmit information or to collect input, and both are better in writing. The meetings that survive this filter are the ones where the interaction itself is doing the work, and those are worth protecting.

Attendance is decided by politics and habit far more than by need, and correcting it is both the cheapest improvement and the most socially awkward.

Why lists grow. Nobody wants to exclude someone who might be offended. People are invited so they are informed. Someone's manager is added because their report is going. And once on a recurring invitation, nobody is ever removed.

Why it matters beyond cost. Group size changes behaviour, not just expense. In a group of four everyone speaks. In a group of twelve, a handful speak and the rest attend. Adding people reduces the participation of those already there, so a large meeting is not a small meeting with more input; it is a presentation with an audience who could have read it.

The useful discipline is to give every attendee a reason.

They decide, or their agreement is needed.

They hold information the discussion requires.

They will do the work that results.

Or they are directly affected in a way that means they should be in the conversation rather than told afterwards.

Anyone who is there to be informed should receive the notes instead, and that is a better outcome for them.

How to shrink a list without causing offence, which is the real obstacle. Make it about their time rather than their value: I did not want to take an hour of your week, the notes will be with you by three, and tell me if you want to be in the next one. Almost nobody objects to that, and many are relieved.

And for a recurring meeting, review the list every few months. Attendance accumulates by default and is only ever removed deliberately.

Most agendas are lists of subjects, which is why most meetings drift. An agenda item should state what has to happen, not what will be discussed.

Compare. Budget is a topic and it could absorb four hours. Decide whether to approve the additional headcount for the support team is an item, and you know when it is finished.

What a usable item contains.

The outcome required. Decide, generate, agree an approach, or resolve a disagreement.

Who owns it, so that someone is responsible for reaching the outcome rather than for describing the topic.

A time allocation, which is a statement about relative importance and forces the conversation about priority to happen while writing the agenda rather than in the room.

And what people must read beforehand, with the material actually attached.

Why that last one matters more than it seems. A decision meeting where the information arrives during the meeting is a reading session with an audience, and the decision will be made by whoever had prior context. Sending material in advance is not a courtesy; it is what makes the hour a decision rather than a briefing.

One practice worth knowing, associated with Amazon and worth considering even outside it. A meeting begins with everyone reading the document in silence for the first portion, because sending it in advance does not reliably mean it was read. This converts a room where three people are informed into one where everyone is, and it takes less time than the explaining it replaces.

And the discipline that makes any agenda real. If an item has no owner, no outcome and no time, it will consume whatever is left and produce nothing, which is what happens to every agenda's item called AOB.

A single sentence at the start of any decision item prevents most of the frustration meetings generate, and it is almost never said.

The frustration comes from a mismatch. People participate believing they are deciding. Someone else believes they are consulting. At the end a decision is announced, and those who thought they were deciding experience it as their input having been ignored.

Nothing about the discussion causes that. It is caused by nobody having said which mode they were in.

The modes worth distinguishing, stated plainly.

I am deciding, and I want your input first. Entirely legitimate and the most common in practice. People are usually content with it when it is stated, and resentful when it is discovered.

We are deciding together, and I will not overrule the group.

We are deciding together, and if we cannot agree I will decide.

We need consent rather than agreement: I am looking for whether anyone has a serious objection, not whether everyone prefers it.

And this is already decided; I am here to explain it and hear concerns about implementation. Saying this honestly is far better than staging a discussion whose outcome is fixed, which people detect and which costs real trust.

The discipline is to say it before the discussion, not after. Stating the mode at the end reads as a justification for the outcome.

And where you are consulting rather than deciding jointly, say what you will do with the input. I will weigh this and decide by Thursday, and I will tell you what I concluded and why. That last clause is what makes consultation feel real rather than performative, and it costs one message.

Recurring meetings are where most organisational meeting time is spent, and they decay in a predictable way that is worth recognising.

The pattern. A meeting is created for a real reason, at a sensible cadence. The reason resolves or changes. The meeting continues, because cancelling requires someone to decide it is no longer needed and nobody owns that decision. The agenda softens into status updates. Attendance grows. And it becomes something people complain about and nobody stops.

The underlying cause is asymmetry. Creating a meeting is a small act by one person. Ending one is a visible decision that implies the meeting was waste, which nobody wants to say about a thing their colleagues have been attending.

What prevents it.

Set an end date when you create it. Weekly until the launch, then we review. This makes continuation the deliberate act rather than cancellation.

Review cadence and attendance periodically, as a normal item rather than as a criticism.

Cancel individual instances when there is nothing to decide. A meeting that is never cancelled teaches everyone that attendance matters more than purpose. And a leader who cancels their own meeting when it is not needed does more for meeting culture than any policy.

And separate the status content out. Most recurring meetings are carrying a status update that should be written, plus a small amount of genuine discussion. Moving the status to a document usually shrinks the meeting by half and improves both parts.

The test from the first step, applied regularly. If this were cancelled for a month, what would break? Answering honestly, out loud, with the attendees, is a legitimate agenda item and it is one of the more valuable half-hours an organisation can spend.

Everything in this lesson, as something to run before sending an invitation.

Could this be a document? If the purpose is transmitting information or collecting input, it probably could, and the document reaches more people and is faster to consume.

What type is it? Deciding, generating, aligning or connecting. Name it in the invitation. If it needs two types, separate them in time with an explicit transition, because generating and deciding actively destroy each other.

What has to be true at the end? Write the outcome, not the topic. If you cannot state what finished looks like, the meeting will not finish, it will run out of time.

Who is here and why? Every attendee decides, holds needed information, will do the resulting work, or is directly affected. Anyone attending to be informed gets the notes instead.

How will the decision be made? Say it at the start. Whether you are deciding with input, deciding together, or explaining something already decided. And if you are consulting, say what you will do with the input and when.

What must people read first? Attach it. And consider reading it together at the start, since sending it does not mean it was read.

How long, per item? An allocation forces the priority conversation into the preparation rather than into the room.

And for a recurring meeting: when does it end, and what would break if it were cancelled for a month?

None of this is facilitation technique, which the next lesson covers. It is all decided before anyone arrives, and it determines more about whether the meeting works than anything that happens inside it.