How to Get Out of a Meeting

Most advice about escaping meetings is about excuses. Excuses are the least useful part, because the meetings worth escaping are usually ones you could simply have declined — and the reason people invent conflicts instead of declining is that nobody is taught how to decline without it reading as a slight.

Deciding whether to go

Before the excuse, the judgment. Four questions settle most invitations.

Is a decision being made, and do I own part of it? If yes, go. Decision meetings are the ones where absence is genuinely expensive — decisions made without you get made anyway, and re-opening one later costs far more than the hour would have.

Am I here to inform, or to be informed? If you are there purely to receive information, that is a document. If you are there purely to deliver information, that is also a document. Meetings earn their cost when information moves in more than one direction and the directions interact.

Would anyone notice if I were not here? Uncomfortable and clarifying. If the honest answer is no, you are attending for visibility, and there are cheaper ways to be visible than an hour of silence with your camera on.

Is this the meeting where the real conversation happens? Sometimes the agenda is trivial but the room is where decisions actually get made. Recognizing that is not cynicism, it is literacy, and it is a legitimate reason to attend a meeting that looks pointless on paper.

Declining well

A good decline does three things: it is prompt, it is specific about why, and it offers a path to the outcome the organizer wanted. Miss the third and you have created work for someone else.

"I'll skip this one — I don't think I'd add much to the pricing discussion. Happy to review the notes and comment on anything that touches onboarding." That works because it is early, it names the reason without inventing a conflict, and it keeps the organizer's objective intact.

What fails: declining silently, declining an hour before, or declining with a vague conflict that everyone can see is a scheduling choice. The vagueness is what causes the damage — an invented conflict is a small lie that a colleague can usually detect, and detecting it costs more goodwill than an honest "I don't think I'm needed" ever would.

Recurring invitations deserve a different move. Declining a single occurrence of a weekly is noise. Asking whether you need to be on the series is a real conversation and usually a welcome one, because the organizer has probably wondered the same thing and has no more idea how to raise it than you do.

Leaving early

The phrase "hard stop" exists because it works. Announced at the start — "I've got a hard stop at half past" — it is a neutral fact everyone can plan around. Announced at the moment you leave, it reads as an escape, because it is one.

That is the entire technique: front-load the constraint. A departure the room expected is invisible; the same departure unannounced is the thing people remember about the meeting.

If you must leave unexpectedly, say less rather than more. "Sorry — something's come up, I'll catch up on notes" is complete. Detailed excuses invite follow-up questions and are harder to keep consistent, which is the practical argument against elaborate ones, quite apart from the honesty argument.

On the excuse generator

The tool on this page produces plausible reasons to leave, scored for believability. It is a joke, and it is worth being clear-eyed about what using it actually involves: lying to a colleague to avoid a conversation you could have had directly.

Sometimes that trade is worth it. A meeting you were added to by mistake, a call that has run forty minutes past its purpose, a recurring invite from three reorganisations ago that nobody has the authority to cancel — these are situations where a small social fiction costs less than the alternative, and pretending otherwise is precious.

But the excuse is a workaround for a calendar problem, not a fix. If you need one weekly, the problem is the invitation, and the durable solution is the conversation in the section above.

Could it have been an email?

The complaint is so common it has stopped meaning anything, which is a shame, because the underlying question is a good one. Three tests separate the meetings that genuinely should have been documents.

Was anything contested? Disagreement resolves faster in conversation than in writing — tone, hesitation and the ability to ask a follow-up immediately all do real work. A meeting where nobody disagreed with anything was an announcement.

Did the conversation change direction? If the outcome matched the agenda exactly, the agenda was the content and could have been circulated. Meetings earn their cost when something unplanned surfaces.

Did more than two people speak substantively? Two people talking while nine listen is a conversation with an audience. It could have been those two people, plus notes.

And the honest counter-case, because this cuts both ways: some meetings that fail all three tests are still worth holding. A team that never assembles loses the incidental awareness that makes collaboration cheap later. A new joiner learns more from twenty minutes of watching how decisions get argued than from any document. Ritual has value that an efficiency audit will always score as zero.

"Could have been an email" is a useful question and a poor verdict. Ask it about a recurring series rather than a single occurrence — that is where the answer actually changes something.

Looking engaged when you are not

Included because pretending otherwise would be dishonest, with one caveat worth stating: the techniques below are for meetings you could not decline and do not need to follow. Using them in a meeting you should be following is a way of wasting the hour twice.

Position matters more than expression. On video, sit slightly closer than feels natural and keep your eyes near the camera rather than the faces — the gaze reads as attention, and it is the single largest factor in how present you appear.

Nodding is over-rated and easy to overdo. Occasional, slightly delayed nodding reads as processing; continuous nodding reads as a metronome and is more conspicuous than stillness.

The one genuinely effective technique is cheap listening: keep a fraction of attention on the last sentence spoken, so you can repeat it if asked. This costs far less than following the argument and covers the only failure that actually matters — being addressed and having nothing.

Asking a question near the end is the classic move and it carries real risk. A question the presenter already answered is worse than silence, because it proves inattention rather than merely suggesting it. If you are going to ask something, ask about implications rather than content: "what does this mean for the timeline?" is nearly always unanswered and nearly always reasonable.