The Psychology of Why Group Plans Fall Apart

A message goes out, everyone replies with enthusiasm, three people drop a calendar emoji, and then eleven days pass with nothing at all. That silence is not flakiness. It is six specific forces acting on a message that was addressed to everybody and therefore to nobody.

Quick answer

Group plans stall because enthusiasm and commitment look identical in a chat but behave completely differently. A message addressed to the whole group creates an obligation in no one, and saying maybe costs nothing while saying no is socially expensive.

Add optimism about a future self with a clear diary, and the fact that agreeing is cheapest on day one while attending is most expensive on day twenty, and the silence is close to inevitable.

The fixes are structural rather than motivational. Propose a real date, give one task to one named person, set a decision deadline, and keep the answers somewhere that is not the conversation.

Eleven days of nothing

Someone posts the message on a Tuesday. "We should do something for Priya's birthday, she's back in town the weekend of the 14th."

The response is immediate and warm. Seven replies inside four minutes, two of them in capitals, three people drop a calendar emoji like a small ceremonial pledge.

Then nothing happens for eleven days.

Nobody booked anything. Nobody picked a place. On day twelve the original poster says "so are we doing this or not?" and the group performs its enthusiasm all over again, slightly quieter this time.

The tempting explanation is that your friends are flaky. It is almost never true, and it is the least useful thing you could conclude.

What actually happened is that a specific set of forces acted on a specific kind of message, and they produce the same silence in every group, every time. Most of them have nothing to do with how much anyone wants to see Priya.

A message to everyone is a message to no one

Start with the addressing problem, because it comes first and causes most of the rest.

"We should do something" has no recipient. It is grammatically addressed to the group, which means it lands in twelve inboxes and creates an obligation in zero of them.

This is the most reliable finding in the social psychology of groups. In Darley and Latane's 1968 bystander study, people who believed they were the only witness to an emergency reported it 85% of the time, while those who believed five others were present reported it 31% of the time.

That is a laboratory study about a staged emergency, and a birthday dinner is not an emergency. The mechanism travels anyway, because the mechanism is not about danger.

It is about the arithmetic of felt responsibility. When a task is visible to everyone and assigned to no one, each person's share of it divides by the size of the group, and nobody experiences the remainder as their problem.

Twelve people did not each decide to do nothing. Each of them assumed the doing was happening somewhere slightly to their left.

The group chat sharpens this to a point. Everyone can see everyone else reading, replying and being enthusiastic, which is exactly the evidence a brain needs to conclude that the situation is handled.

Read the thread back and you will find a striking absence. Twelve messages of agreement, and not one sentence containing a person's name and a verb.

Why "someone should book it" never books it

Notice how planning language is almost entirely passive by default. "We should", "someone could", "it'd be good to", "maybe we do".

That grammar is not accidental, and it is not laziness. It is politeness doing its job.

Naming a person and a task inside a group of equals feels like giving an order, and friends do not give each other orders. So the group speaks in a voice that assigns nothing, and then quietly wonders why nothing was assigned.

There is a related effect on effort itself. The APA's definition of social loafing describes the tendency to put in less individual effort on a task when working in a group than when working alone, which is precisely the shape of a plan where six people are notionally organising and none of them are.

You did not promise. You expressed a feeling.

Now the second force, and it is the one people find hardest to accept about themselves.

When you replied "yes! love this!" you were not lying, hedging or stalling. You were reporting your genuine emotional state on a Tuesday afternoon.

An intention and a commitment feel almost identical from the inside. They behave completely differently in the world.

An intention - A commitment

What it describes - How you feel about the idea now - What you will do on a named date

What it costs to make - Nothing, and it feels lovely - A held slot and a closed option

Who else can see it - Everyone, as enthusiasm - Everyone, as a name on a list

What happens if you drop it - Nothing, nobody counted it - Somebody has to be told

Where it lives - In a message that scrolls away - In a record that persists

What it survives - Only the mood that made it - A changed mood, mostly

The distance between the two has a name in behavioural research, and it is well measured. Gollwitzer and Sheeran's review of goal pursuit, hosted by the US National Cancer Institute, reports that across a large body of studies people act on their stated good intentions roughly half the time.

Their fix is worth knowing, because it is small and it works. An if-then plan, naming the situation and the action together, closes a good part of that gap compared with intention alone.

"Yes, sounds fun" is an intention with no situation attached to it. "I'll book the table on Thursday morning" is an if-then plan wearing casual clothes.

Most group chats are built to collect the first kind and have no way to hold the second. This is the same confusion that makes a poll answer feel like an attendance when it is really a statement about your calendar.

The cheapest word in the language

Third force, and it explains the shape of the replies rather than their number.

Ask twelve people if they are coming and count the answers. You will get a pile of yes, a scattering of maybe, and almost no no.

That distribution is not what the group actually believes. It is what the group can afford to say.

Saying no in a group chat is socially expensive. It is a public statement, permanently visible to eleven people, and it invites the question of why.

Saying maybe costs nothing at all. It keeps you warm, keeps the option open, and requires no explanation from anyone.

The answer - What it costs to say - What the organiser can do with it

"Yes, I'm in" - Low now, higher later if you drop out - Almost everything, if it is recorded

Thumbs-up reaction - Nothing, and it is deniable - Very little, and it decays fast

"Maybe, depends on work" - Nothing at all - Nothing, but it feels like something

Silence - Nothing, and it is invisible - Nothing, and it cannot be chased politely

"Can't make it, sorry" - High, and it is public - Genuinely useful, immediately

"Yes" then vanishing - Low, if nobody wrote it down - Worse than nothing, it inflates the count

Look at the last column and the problem becomes obvious. The two answers that help the organiser most are the two that cost the responder most.

Every group therefore optimises itself into a headcount of nine that is really a headcount of five. Nobody deceived anyone, and the number is still wrong.

This is also why emoji reactions make such poor attendance records. A thumbs-up is a social gesture that costs nothing, and the organiser is quietly treating it as a promise.

The asymmetry gets worse as the group grows

In a group of four, declining is nearly free. Everyone knows your circumstances, the reason is usually obvious, and the plan visibly bends around you.

In a group of thirty, declining is a small public announcement to twenty-nine people you have varying relationships with. So the maybes multiply, the yeses get softer, and the true headcount hides.

Group size does not change how much people want to come. It changes the price of honesty, which changes the data, which changes everything downstream.

The person who says yes lives in a different week from the person who attends

Fourth force. When you accept an invitation for the 14th, you are making a prediction about a person you have never met.

That person is you, three weeks from now, and you are extraordinarily generous about their availability. Their diary is clear, their energy is high, and they have no laundry.

The tendency has been measured directly. In Buehler, Griffin and Ross's 1994 work on the planning fallacy, students predicted they would finish their thesis in about 34 days and actually took about 55, and fewer than a third finished inside their own estimate.

Thesis deadlines are not Saturday nights, so do not over-read that. What the work establishes is the direction of the bias, and the direction is what matters here.

We are systematically optimistic about our own future capacity, and notably less optimistic about other people's. Which means every yes in the thread is slightly inflated, and the organiser is adding up inflated numbers.

None of this is dishonesty. It is a forecast made in good faith by someone with a structural blind spot about their own next fortnight.

Enthusiasm is cheap early and attendance is expensive late

Put the last three forces together and a shape appears. The cost of participating in a plan is not constant, and neither is the reward.

On day one, agreeing costs one tap and buys you a genuine hit of anticipation. You get to be someone who is going to a thing.

On day twenty, attending costs a train fare, a cancelled other plan, four hours and the energy to be charming. The anticipation was spent weeks ago.

So the group's collective enthusiasm peaks in the first hour and declines steadily, while the actual demands of the plan sit patiently at the far end, unchanged. The two curves cross somewhere around the second week, and that crossing is where plans go to die.

This is why the eleven days of silence are not neutral. They are not a pause in the plan, they are the plan's fuel burning off.

It is also why fast, small, boring decisions beat slow, careful, comprehensive ones. A mediocre venue confirmed on Wednesday outperforms a perfect venue agreed in principle a month later, because the confirmed one converts enthusiasm into a commitment while the enthusiasm still exists.

Why the silence is self-reinforcing

There is one more thing happening in those eleven days, and it is the reason the gap gets longer rather than shorter.

Silence is ambiguous, and people read ambiguity in whichever direction costs them least. Nobody in the thread concludes that the plan is dying, because that conclusion would obligate them to do something about it.

Instead they conclude, quite reasonably, that it is early days. Everyone independently arrives at the same comfortable interpretation, and each person's calm is evidence for everyone else's.

Then the silence starts to protect itself. Speaking up on day nine means being the person who noticed, and the person who noticed is next in line to be the person who does it.

So the longer nothing happens, the more expensive it becomes to say anything. That is a genuinely nasty feedback loop, and it is why the plan almost never restarts on its own.

The organiser tax

Which brings us to the fifth force, and the least fair one.

In every group there is a person who ends up doing it. Not because they volunteered, and usually not because they enjoy it.

They do it because they have the lowest tolerance for the plan collapsing. That is the whole selection mechanism, and it is brutally efficient.

The group is running an auction where the prize is unpaid work, and the winner is whoever cracks first. Everyone else is not free-riding on purpose, they are simply more comfortable with uncertainty than the organiser is.

The three jobs nobody counted

What that person actually does is worth breaking apart, because groups consistently underestimate it.

The first job is deciding, and it is the only one anyone notices. Picking the place, picking the time, absorbing the blame if either is wrong.

The second job is being the database. Holding the headcount in their head, remembering that Sam is a maybe and Tom said yes but has a wedding, tracking that the time moved from eight to half seven. This is the invisible one, and it is the reason attendance tracking in a chat is so exhausting.

The third job is chasing. Asking the six silent people, individually, in a way that does not sound like nagging, then doing it again a week later.

Chasing is the job that burns people out. It is repetitive, it feels socially costly, and it produces a version of you that you did not want to be in this friendship.

The organiser tax is not measured in hours. It is measured in the number of times you have to be the slightly annoying one.

What happens when the tax gets too high

Nobody announces that they are retiring from organising. They simply stop starting things.

The group notices nothing, because the absence of a plan is not an event. It just gradually becomes a group that used to do things.

That is the real cost of all of this, and it never appears in any thread. A wrong meeting time is visible and annoying, and the dinner that was never proposed is invisible and much worse.

Six forces, one silence

It helps to see them listed, because the instinct is to blame one big thing when it is really several small ones stacking.

Diffusion of responsibility A request addressed to everyone creates an obligation in no one, and the bigger the group, the smaller each person's felt share. The intention-commitment gap Enthusiasm about an idea is not a decision about a date, but the two are indistinguishable when both arrive as the same kind of message. Social cost asymmetry Maybe is free and no is expensive, so groups produce far more maybes than they actually contain. Optimism about the future self The person accepting the invitation has an empty diary and infinite energy, and they are fictional. Cost inversion over time Agreeing is cheapest when enthusiasm is highest, and attending is most expensive when enthusiasm is lowest. The organiser tax One person absorbs the deciding, the record-keeping and the chasing, and eventually stops absorbing it.

Notice how few of these are about character. Five of the six are properties of the situation, and the sixth is a property of being human.

That is genuinely good news. Situations can be changed, and there is no version of this where you fix it by having better friends.

What actually breaks the silence

Everything above suggests the same handful of moves. None of them require you to be more assertive than you are comfortable being.

• Propose a date, not a plan. "Dinner on Saturday the 14th, 8pm, somewhere central" gives people something to answer. "We should do something soon" gives them something to agree with, which is not the same and never converts.
• Name one person for one task. Diffusion of responsibility dies the moment a task has an owner. "Marcus, can you pick the place by Thursday?" is not bossy, and Marcus will be quietly relieved that the ambiguity is gone.
• Make no cheaper. Say "just tell me if you can't, no explanation needed" and mean it. You are deliberately lowering the price of the answer you most need and least often get.
• Set a decision date, not just an event date. "I'm booking on Thursday for however many we have" converts a vague future into a deadline, and deadlines are what turn intentions into behaviour.
• Write the answers down somewhere that is not the chat. The moment the headcount lives in one person's head, that person has become infrastructure, and infrastructure gets tired.
• Ask again once, then stop. One clear follow-up is fair and effective. Three is chasing, and chasing is what makes people stop organising things.

The second and fourth do the heaviest lifting. An owner and a deadline between them dismantle most of the silence, which is a suspiciously small amount of work for the result.

Why the usual fixes disappoint

Groups do try to fix this. The attempts fail in predictable ways, and the failures are instructive.

Another poll. Polls are excellent at one job, which is closing a single question at a single moment. They are terrible at attendance, because voting that Saturday works is a claim about your diary rather than a promise about your evening, a distinction chat polls are structurally unable to make.

A pinned message. Pinning is the best move available inside a thread and it still decays. It is a photograph of the plan at one moment, so when the time moves it is either stale or manually swapped, and the people who read the old one hold wrong facts with total confidence.

More reminders. Reminders address forgetting, and almost nobody in this story forgot. They knew, and the knowing was never the bottleneck.

A stricter organiser. This works exactly once, and costs the organiser something they do not get back. Turning yourself into the group's enforcement mechanism is a solution that consumes the person applying it.

A new group chat for the event. Now there are two places to look and the same problems in both. This is how event information ends up scattered across several apps with no version of it that is definitively correct.

Waiting for consensus. Groups often hold off booking until everyone has weighed in, which sounds fair and is quietly fatal. The people you are waiting for are the ones least invested, so the plan is paced by its least committed members while its most committed ones lose interest, which is the same trap that makes chatting about an event feel like planning it.

Every one of these treats the symptom. The forces underneath are untouched, which is why the same group has the same eleven days of silence next time.

What a plan that holds its own state removes

Here is the constructive part, and it is more mechanical than motivational.

Most of the six forces are not psychological facts you have to overcome. They are consequences of the plan living in a medium that cannot hold a state, and several of them disappear when that changes.

A record with a date and a place on it converts a vague intention into a specific thing you either accepted or did not. That handles the intention-commitment gap, because there is now a difference between agreeing and answering.

A guest list with a default of "no answer yet" gives silence a name. Nobody has to be chased on a hunch, because the twelve people are visibly not one undifferentiated blob of enthusiasm.

That idea is old and well specified. The IETF's calendar data standard, RFC 5545, defines an attendee's participation status with values including needs-action, accepted, declined and tentative, and needs-action is the default until somebody actually answers.

A private answer makes no cheaper. Declining to a record rather than to an audience removes most of the social cost that produced all those maybes in the first place.

And a shared record removes the organiser's second job entirely. They are no longer the database, which is the job that made them the organiser and the job that eventually made them stop.

None of this fixes optimism about your future self, and nothing will. It just means the group finds out about it early enough to matter.

Where Ontaym fits

This is the honest bit, and it is one section rather than a running theme.

Ontaym exists because of the paragraph above. An event is a record with one address rather than a conversation you have to reconstruct, so the current time and place and guest list are things you can look at rather than things you have to remember.

People answer without joining a group, installing anything or handing their phone number to eleven strangers. Your chat carries on exactly where it is, doing the part it is genuinely good at, which is the conversation rather than the coordination.

It does not make anyone want to come. It removes the structural reasons that people who did want to come never got round to saying so.

The thing worth remembering

Next time a plan evaporates, resist the flakiness explanation. It is emotionally satisfying and it predicts nothing.

Ask instead which of the six forces did the work. Usually it was the first one, and usually the message had no name in it.

Then do the smallest possible thing. Propose a real date, hand one person one task, and tell everyone that no is a perfectly good answer.

You will be surprised how much of what looked like indifference was just twelve people politely waiting for somebody to go first. Someone has to, and there is no version of group planning where nobody does.

The trick is making sure it does not have to be the same person every time. That is a structural problem, and structural problems have structural fixes, which is the most cheerful sentence in this entire article.

Frequently asked questions

Why does everyone say yes and then nobody actually commits?

Because enthusiasm about an idea and a decision about a date feel the same from the inside but arrive as the same kind of message. Saying yes on day one costs nothing and buys real anticipation, while attending on day twenty costs money, time and energy. The gap between the two is where most plans quietly disappear.

What is diffusion of responsibility in group planning?

It is the tendency for felt responsibility to divide by the number of people who can see a request. A message saying we should organise something is addressed to the whole group, so nobody experiences it as their task. Naming one person for one task removes the effect almost entirely.

Why do so many people answer maybe?

Because maybe is free and no is expensive. Declining in a group chat is a public statement to everyone in it and tends to invite a follow-up question, while maybe keeps you warm and requires no explanation. Groups therefore produce far more maybes than they actually contain.

Is it true that people overestimate their own future free time?

Research on the planning fallacy has repeatedly found people underestimate how long their own tasks will take, and are less optimistic when predicting for others. In the 1994 Buehler, Griffin and Ross studies, students predicted finishing a thesis in about 34 days and took about 55. The direction of that bias is what applies to a Saturday three weeks out.

Why does the same person always end up organising?

Because the group is effectively running an auction where the prize is unpaid work, and the winner is whoever has the lowest tolerance for the plan collapsing. That person then takes on three jobs: deciding, holding the headcount in their head, and chasing the people who never answered. The third one is what eventually makes them stop.

Will a poll fix this?

A poll closes one question at one moment, which it does well. It cannot hold attendance, because voting that a date works is a statement about your calendar rather than a promise about your evening. Preferences also keep moving long after the votes stop.

How do I ask for a firm answer without sounding pushy?

Lower the price of the answer you want. Saying just tell me if you cannot make it, no explanation needed, removes most of the social cost that produces vague maybes. Pair it with a decision date, such as booking on Thursday for however many people have answered.

Does moving the plan out of the chat really help?

It removes several of the forces at once rather than treating symptoms. A record with a date, a place and a guest list makes the difference between agreeing and answering visible, gives silence a name instead of hiding it, and takes the headcount out of one person's head. The conversation stays exactly where it already is.

Ontaym Editorial Team

Ontaym builds tools for organising real-world gatherings, so the team spends its days on the coordination problems this article describes. Articles are researched against primary sources, reviewed before publication, and revised when the underlying facts change rather than on a schedule.

Give your next plan one address instead of one more thread.