You said it half-aloud, a little dry aside while the screen-share loaded for the fourth time. Nobody laughed. Someone's eyes flicked to someone else. Then the meeting moved on, as meetings do, and you spent the next forty minutes rehearsing exactly how many people heard it and what they now think of you.

That sting is out of proportion to the joke. A failed joke is a small, public prediction about other people's minds, and the result comes back in under a second. No wonder many people simply stop trying and settle into a careful, humorless version of themselves that nobody objects to and nobody remembers.

That's a real loss, because humor at work isn't decoration. And the good news is that why jokes land or die is not a mystery of personality. It follows a pattern.

What humor is actually made of

In 2010, psychologists A. Peter McGraw and Caleb Warren proposed the benign violation theory. Its claim is simple: something is funny when it is, at the same time, a violation and benign.

A violation is anything that threatens how things are supposed to be: a social norm, a rule, your sense of dignity, the way the world usually works. A benign appraisal is the simultaneous sense that it's okay, safe, harmless.

Both conditions have to hold at once. A pure violation (a genuinely cruel remark, a real disaster) is just upsetting. A pure benign (a perfectly normal, correct statement) is just boring. Humor lives in the narrow overlap, which is why the same sentence can be hilarious in one room and appalling in another. Nothing about the words changed. The room's judgment of "is this safe?" did.

That reframes your failed joke. It probably wasn't "not funny." It was likely missing one of the two ingredients: either there was no violation (too bland to register) or the benign half wasn't established (the room couldn't tell it was safe).

Three ways a violation becomes benign

McGraw and Warren describe several routes by which a threatening thing gets read as harmless. Three of them translate neatly to a conference table.

Psychological distance. Things are easier to laugh at when they're far away in time, space, or stakes. This is why "we'll laugh about this someday" is true: the same mishap is funnier a year later. In a meeting, a project that has already shipped is safer to joke about than one that's currently on fire.

Weak commitment to the norm being violated. If nobody in the room is deeply invested in the rule that's being bent, bending it is playful. Teasing the absurd formatting of the quarterly template works. Teasing something a person is genuinely proud or anxious about doesn't, because their commitment to it is strong.

An alternative interpretation. The listener needs a second reading in which nothing bad is happening. Tone, a grin, and an established relationship all supply this. It's also why the same joke from a trusted colleague and from a stranger gets opposite reactions: the stranger hasn't given the room a reason to pick the harmless reading.

Why the stakes are higher than they look

Research by Alison Wood Brooks, Hengchen Dai, and Maurice Schweitzer (and related work by Bitterly, Brooks, and Schweitzer on humor and status) found something both encouraging and sobering: a joke that lands tends to raise how much status and confidence others attribute to the teller, while a joke that bombs or reads as inappropriate tends to lower it. Humor is a signal, and signals can be read both ways.

That's the honest case for caution. It's also the honest case against abandoning humor entirely. The same mechanism that punishes a misjudged joke rewards a well-judged one, because a well-judged joke proves you can read a room, and reading the room is much of what people mean by "good judgment."

The safest funny: aim at the situation, not the person

If you want a default, here it is. The most reliable target in a meeting is the shared, low-stakes absurdity everyone is already experiencing together: the call that won't connect, the agenda item that has appeared on seven consecutive agendas, the calendar invite titled "Quick Sync" that is ninety minutes long.

This works because it is nearly pure benign. Nobody is the victim. The violation is mild (the meeting isn't working the way meetings are supposed to) and the observation bonds the room, because it says "I see what you see." Researcher Robert Provine, who studied laughter by listening to people in ordinary conversation, found that most laughter follows unremarkable remarks rather than punchlines. A great deal of laughing is social glue, not comedy, and an observation about the shared moment is glue in its purest form.

Self-directed humor about a trivial flaw ("I'm going to mute myself now, as I've been doing for the last ten minutes, in my head") also tends to read as benign, as long as it's about something that doesn't threaten your credibility on the actual topic. Mocking your own competence at the thing you're presenting turns the joke into evidence.

Avoid targets where the benign half is fragile: a specific colleague's work, anyone's appearance or background, a decision someone just made, anything involving a person who isn't present, or any topic where someone in the room has real stakes right now.

Your next moves

  • Run a two-second check before any joke. Ask silently: "Is it a violation of something, and is it clearly safe?" If you can't answer "safe" about every person present, hold it.
  • Pick one low-stakes, shared target this week. The tech glitch, the recurring agenda item, the length of the meeting. Make one light observation about it, once. You're practicing, not performing.
  • Establish the benign reading with your delivery. Smile slightly, slow down a touch, and keep your voice warm. Tone is the "alternative interpretation" that tells the room it's safe.
  • Let a laugh be small. Don't wait for a reaction or repeat the line. A quiet smile from two people counts as a hit, and chasing more turns the moment into a performance.
  • Write down one thing that landed and one that didn't, with the room's mood. A two-line note after the meeting builds the one thing no theory can give you: your read of this particular team.

When it falls flat anyway

It will, sometimes. The recovery is dull and effective: don't explain it, don't apologize at length, and don't retreat into silence for the rest of the hour. A light "tough crowd" or simply moving on keeps the moment at its true size. Remember that others are far more absorbed in their own contribution than in yours; the spotlight you feel on a failed joke is mostly your own.

If a joke genuinely landed wrong, because it touched something sensitive, a short, private "that came out sharper than I meant" afterward does more than any public damage control.

Where a record helps

The hard part of all this is calibration. You learn what's safe on your team by noticing patterns over weeks, and memory is a poor witness: it replays the one joke that died and forgets the three small ones that quietly worked. Meetingmortem is a place to jot a short reflection after a meeting, what you said, how the room moved, what you'd do again, so your read of the people around you gets sharper instead of just more anxious. If that sounds useful, it's at meetingmortem.lumenlabs.works, and the five steps above work just as well with a scrap of paper.