What Airlines Already Know About Telling People No

Somebody is going to be told there is no room. That part is fixed the moment demand exceeds capacity. What is not fixed - and what turns out to matter far more than the refusal itself - is when they find out, whether they had any choice in it, and whether anyone gave them somewhere else to go.

Quick answer

Research published in the Journal of Service Research found that compensating a bumped airline passenger costs roughly $75–$150 if they are told at home, about $1,350 if told at the gate, and around $7,200 if told after boarding.

That is a fifty-fold difference in the price of identical bad news, driven entirely by timing. The mechanism runs through perceived justice: the later and more public the news, the more it reads as betrayal rather than bad luck.

Voluntary offloading - asking for volunteers before removing anyone - required roughly half the compensation of an involuntary bump, even when the outcome for the person was the same.

An unusually well-priced piece of bad news

Most of what organisers know about disappointing people is folklore. Airlines are the exception, because overbooking forces them to do it at industrial scale - denied boarding across major US carriers ranged from roughly 45,000 to 150,000 passengers a year between 2008 and 2019 - and that scale produced actual research.

A 2021 study in the Journal of Service Research by Nazifi, Gelbrich, Grégoire, Koch, El-Manstrly and Wirtz asked a precise question: how much compensation does it take to keep a bumped passenger from feeling unfairly treated? The answer depends almost entirely on when they were told.

When the passenger learns - Compensation required - Relative to legal minimum

Proactively, before leaving home - $75–$150 - Well below

Reactively, at the gate - ~$1,350 - 100%

Very reactively, after boarding - ~$7,200 - 500%

Nothing about the outcome differs across those three rows. The same passenger does not get on the same flight. The only variable is how far into their journey they had travelled - physically and psychologically - before being told, and that variable moves the price by a factor of roughly fifty.

The same refusal costs fifty times more at the gate than at home. Nobody's plans changed. Only the timing did.

Why timing does this

The researchers describe a chain rather than a single effect, and the chain is what makes the finding portable to contexts that have nothing to do with aviation.

Timing shapes perceived justice. Perceived justice shapes whether the experience registers as ordinary misfortune or as betrayal. And betrayal - not disappointment - is what drives the angry public response, the negative social media post, the decision never to return.

The reason lateness feels like betrayal is that every step a person takes toward an event is an investment they made on the strength of your confirmation. Someone told at home has invested a decision. Someone told at the door has invested an afternoon, a journey, possibly a babysitter, and the social cost of having told people where they were going.

By the time they arrive, they are not being denied a seat. They are being told that the confirmation they organised their day around was never reliable, which is a claim about you rather than about capacity.

The volunteer effect

The second finding is the more actionable one, and it is close to free to implement.

Voluntary offloading - asking who is willing to give up their seat before removing anyone involuntarily - required roughly 50 percent less compensation than an involuntary bump. The person still does not travel. What differs is that they chose it.

That gap is worth sitting with, because it is entirely psychological and entirely real. Agency converts something done to a person into something done by them, and the difference in how it lands is worth about half the compensation budget.

For an event organiser the translation is direct: before you cut anyone from a full list, ask. "We're oversubscribed for Thursday - is anyone able to move to the March session?" will produce volunteers, and every volunteer is a person you did not have to disappoint.

There is a revenue argument too

It is easy to read all this as being about goodwill, which organisers often treat as a soft consideration that loses to practical constraints. The same research connects it to money.

Airlines managing proactive notification well could see net revenue rise by as much as 1.3 percent, while aggressive late-stage compensation strategies were associated with revenue losses of up to 1.0 percent. In an industry with thin margins, that spread is substantial.

For a community event the equivalent is not revenue but return rate, and it behaves the same way. A person told early with a real alternative comes back. A person turned away at the door has learned something about how much your confirmations are worth, and that lesson persists long after the specific evening is forgotten.

Turning this into a waitlist policy

The research produces three design principles, and they are unusually clear about what to do.

• Know your real capacity early, and communicate against it. The single largest lever is how far in advance you can tell someone. Every day of delay moves the cost up a curve that gets steep near the end.
• Ask before you cut. A request for volunteers costs one message and halves the goodwill cost of the people it converts. Nothing else in this article is that cheap.
• Never let anyone find out at the door. This is the $7,200 row. Whatever it takes - a message the night before, a check-in list sent in advance, an explicit standing-room offer - is cheaper than the alternative.
• Offer a next step, not just an apology. Priority for the next session, a recording, a guaranteed seat. An apology closes a door; an alternative moves the person to a different one.
• Tell people they are on a waitlist, and where. Ambiguous status is its own harm. "You are fourth" is manageable; "we'll let you know" is a slow, silent version of being told at the gate.

The last one is underrated. A person who knows they are fourth on a list can make their own plans, which is precisely the agency the volunteer finding says is worth half the compensation. A person told only that they might hear something has been given the disappointment without the choice.

"You are fourth on the list" is a plan. "We'll let you know" is the same bad news with the agency removed.

How overbooking applies to events, carefully

If your measured attendance rate is 70 percent, confirming exactly your room capacity guarantees a room that is 30 percent empty every single time. Overbooking is the obvious response and it comes with the obvious risk.

The research does not say do not overbook. It says the cost of getting it wrong is concentrated almost entirely at the last moment, which suggests a specific approach rather than a blanket prohibition.

Approach - Upside - Cost when wrong

Confirm exactly to capacity - Never turn anyone away - A visibly empty room at every event

Overbook against a measured rate - Full room most events - Occasional overflow - manageable if planned

Overbook against an optimistic rate - Full room, briefly - The door scenario, which is the expensive one

The middle row works only if two things are true: the rate is measured from your own history rather than assumed, and you have decided in advance what happens on the day everyone arrives. Standing room, an overflow space, or an honest early message are all acceptable. Improvising at the door is not.

The lottery question

For heavily oversubscribed events, organisers often reach for a random draw on fairness grounds, and it is worth being precise about what a lottery does and does not fix.

A lottery makes the selection procedurally fair, which genuinely matters - perceived justice is the mechanism the whole research chain runs through, and "chosen at random" is much easier to accept than "chosen by someone, somehow."

What a lottery does not fix is timing. A random draw held two days before the event still delivers late bad news to everyone who loses it. The fairness of the method and the cost of the moment are independent variables, and a well-designed process gets both right rather than trading one for the other.

If you run a draw, run it early, announce the date of the draw in advance so nobody is waiting in ambiguity, and tell the people who did not get in what happens next rather than simply that they did not get in.

Three questions before your next event fills

• How early can you actually know your final capacity? That date is your single biggest lever, and it is usually earlier than you have been treating it.
• Have you asked for volunteers before cutting anyone? It costs one message and halves the cost of every person it converts.
• What is the alternative you are offering, not just the apology? A next step turns a refusal into a deferral, which is a different message entirely.

What to take from this

Running out of room is not the failure. Telling someone at the wrong moment, in the wrong way, with nowhere to go afterwards is the failure - and the research prices that difference at roughly fifty to one.

So tell people early, even when it is uncomfortable and even when you are still hoping something opens up. Ask for volunteers before you remove anyone. Tell people exactly where they stand rather than leaving them in ambiguity. And always attach a next step, because the difference between a rejection and a deferral is the difference between somebody who is finished with your event and somebody who is coming to the next one.

Frequently asked questions

Is it better to ask for volunteers or just cut people from a waitlist?

Ask. Research on airline bumping found voluntary offloading required roughly 50 percent less compensation than an involuntary bump, even when the outcome for the person was identical, because agency changes how it lands.

How early should I tell someone they didn't get a place?

As early as your capacity picture allows. The same research prices proactive notification at home at $75 to $150, against roughly $1,350 at the gate and $7,200 after boarding - the cost curve gets very steep near the end.

Does offering an alternative actually help?

Yes. A concrete next step - priority for the next session, a recording, a guaranteed place - converts a rejection into a deferral, which is a materially different message even though the capacity constraint is unchanged.

Should I run a lottery for oversubscribed events?

A lottery makes selection procedurally fair, which matters because perceived justice drives the whole effect. But it does not fix timing: a draw held two days out still delivers late bad news, so run it early and announce the draw date in advance.

Is deliberate overbooking a good idea for events?

It can be, if you overbook against a measured attendance rate from your own history rather than an optimistic one, and if you decide in advance what happens on the day everyone arrives.

What is the worst way to handle an oversubscribed event?

Letting someone find out at the door. That is the scenario the research prices at roughly five times the legal compensation requirement, because by then the person has invested a journey and told others where they were going.

Should I tell people their position on a waitlist?

Yes. A specific position lets someone make their own plans, which restores the agency that the volunteer finding shows is worth about half the goodwill cost. Ambiguous status delivers the disappointment without the choice.

Does this research really transfer from airlines to events?

The exact dollar figures are aviation-specific, but the mechanism - timing driving perceived justice, which drives whether the experience registers as bad luck or betrayal - is a general finding about being denied something you were promised.

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.

Keep your guest list current so nobody finds out at the door.