The booking was the easy part. Eight of you wanted the chef’s counter, the restaurant took a deposit per seat to hold it, and everyone tapped “pay” in the group chat without a second thought. Then, the morning of, the text arrives: “I’m so sorry — something came up, I can’t make it.” Seven people now stare at the same problem. There’s an empty seat at a dinner that’s already been paid for, and nobody knows what that empty seat costs — or who pays it.
Here is the thing almost every group gets wrong in that moment: they argue about getting the money back. They can’t. A deposit is gone the instant it’s charged, whether eight people show or seven. The bill for a pre-paid dinner isn’t decided at the table by what everyone ordered — it was locked in when you booked. So the only number that can still change is how many people split it. Pre-payment quietly moves the whole money decision from the table to the calendar, and the fight over the empty chair is really a fight with a cost that already happened.
This is a guide to that specific situation: a group dinner where money changed hands before anyone sat down, and then someone dropped out. It’s a newer problem than it sounds, because the kind of dinner that asks for a deposit is exactly the kind of dinner more groups are booking now.
Sources: OpenTable, Top Trends Set to Define Dining in 2026; The Philadelphia Inquirer (2026).
Why do so many group dinners now come with a deposit?
Because the dinners worth gathering for have shifted toward the kind a restaurant can’t afford to leave empty. OpenTable’s 2026 data shows experiential dining — tasting menus, chef’s tables, pop-ups, collaborations — up 46% year over year, group dining up 11%, and 61% of Americans saying a meal out now feels like a special occasion rather than an everyday one. Those are the seats with real fixed costs behind them: a set number of covers, ingredients bought to order, staff scheduled for a full room. A no-show doesn’t just lose a tip — it strands a plated cost.
So restaurants started pricing that risk in. Deposits, card holds, and per-person no-show fees are now standard at the top of the market, and the platforms have followed: in early 2026 OpenTable began adding a 2% service fee that rides on top of no-show penalties, deposits, and prepaid experiences. The math is brutal at the high end — the Chicago group behind Alinea once put its annual losses from no-shows and short-seated tables at roughly $260,000 a year, which is what pushed it to sell dinner as a non-refundable ticket in the first place.
Why this lands on you: the deposit exists to protect the restaurant from your table’s no-show. It does nothing to protect the rest of your group from the friend who bails — that gap is yours to sort out, and almost no one sorts it out before it happens.
Sources: OpenTable (2026); The Philadelphia Inquirer (2026); FCSI, “Tock ticketing”.
What actually happens to the money when someone cancels?
Nothing. That’s the whole answer, and it’s the part that’s hard to accept. The deposit or prepayment is already spent — it left your accounts when you booked, and the restaurant keeps it whether the seat is filled or not. A cancellation doesn’t create a new debt anyone can collect. It just leaves the same fixed cost divided among fewer people.
To see why that feels so strange, compare it to a normal dinner. Normally the bill is built backward, at the end, out of what each person actually ordered — your pasta, your two glasses of wine, your share of the tax and tip. A pre-paid dinner is built forward: the total is set the day you book, before a single plate exists, and your share is just that fixed number divided by everyone splitting it. When someone drops out, the price of the dinner doesn’t move. Only the denominator does.
Normal dinner → your share = what you ordered + tax + tip (settled after the meal)
Pre-paid dinner → your share = fixed total ÷ people splitting it (locked before the meal)
A no-show changes exactly one thing: the denominator.
This is the reframe the whole problem turns on. You are not trying to recover a missing payment. You are re-dividing a number that was already paid. Once a group sees it that way, the question stops being “how do we get the money back from the person who flaked” — which has no answer — and becomes “who absorbs the slightly larger share now,” which does.
Why does one empty chair start a three-day argument?
Because of a bias with a name: the sunk cost effect. Hal Arkes and Catherine Blumer defined it in 1985 as “a greater tendency to continue an endeavor once an investment in money, effort, or time has been made,” driven, they argued, by the simple desire not to appear wasteful. Money already spent shouldn’t change what you do next — but it does, and powerfully.
Their cleanest evidence looks a lot like your dinner. In a field study, theater patrons who happened to pay more for a season subscription went on to attend more plays over the next six months than patrons who’d paid less for the identical seats — not because they enjoyed theater more, but because their larger sunk cost pulled them into the seats. Swap “season ticket” for “dinner deposit” and you have the prepaid table exactly: the money already paid is quietly steering the decision.
“Customers who had initially paid more for a season subscription to a theater series attended more plays during the next 6 months, presumably because of their higher sunk cost in the season tickets.”
Arkes & Blumer, The Psychology of Sunk Cost (1985)
That single bias misfires in two directions at a cancelled dinner. The person who wants to bail drags themselves out anyway — “I already paid, I have to go” — even when staying home is the better call, because the deposit is non-refundable either way. And the friends who showed treat the absent person’s share as money owed to them, a debt to claw back, when it’s nothing of the kind. Arkes and a later collaborator, Peter Ayton, went further in 1999: they called the effect maladaptive, showed that young children — and animals — don’t chase sunk costs the way adults do, and traced the adult version to an over-applied “Don’t waste” rule. The urge to not “waste” the seat is a habit, not a calculation. The money is gone whether you eat or argue.
Sources: Hal R. Arkes & Catherine Blumer, “The Psychology of Sunk Cost,” Organizational Behavior and Human Decision Processes (1985); Hal R. Arkes & Peter Ayton, “The Sunk Cost and Concorde Effects,” Psychological Bulletin (1999).
Who should absorb the no-show’s share?
One of three people, and which one is fair depends on a rule you ideally set before booking. There’s no universal right answer — there’s a right answer for your group, your relationship to the person who dropped, and how big the gap is. The three options are: split the empty seat among everyone who came, let the no-show eat their own prepaid share, or transfer the seat to someone else. Here’s how they compare.
Notice that two of the three only feel fair if the group decided in advance. “You eat your own share if you cancel after Thursday” is reasonable when everyone agreed to it at booking, and petty when it’s sprung on a friend the morning of. That’s why the timing of the conversation matters more than the rule you pick. The crude default — just split it evenly among whoever’s left — is fine for small gaps, but if you want a fairer baseline than that, the math of what a fair split actually is is its own question.
What do restaurants know about this that your group doesn’t?
That the fair fix for a paid-for seat is to make it transferable, not refundable. This is the quiet genius of the ticketed-dinner model. When Alinea’s group moved to selling dinner as a ticket — “in the same way that they might book for the theatre or a sports event” — they didn’t promise refunds. They made the ticket yours to hand off. Its sister restaurant Next spells out the policy plainly: after the cutoff, “all sales are final and non-refundable,” but “you may always transfer your reservation to another guest.”
Apply that to your group and the answer to “who covers the empty seat” often disappears. The fairest move for the person who bails isn’t to demand a refund from friends, and it isn’t to be chased for money — it’s to find a replacement. Bring the roommate, the partner, the coworker who’d have loved the seat. The denominator goes back to eight, the cost per person snaps back to what everyone budgeted, and the money that was already spent actually buys a dinner someone eats. The restaurant solved its no-show problem by making seats transferable. Your group can borrow the same move.
The transfer rule: a non-refundable seat is not a dead loss — it’s a gift card with a date on it. The person who can’t come owns the right to give that seat away, and using it is almost always fairer than splitting the gap or eating it.
Sources: FCSI, “Tock ticketing”; Next Restaurant, FAQ.
How do you set the no-show rule before you book?
Say it out loud in the group chat the moment the deposit comes up — not after someone cancels. The reason cancellations turn into arguments is that the rule gets invented under pressure, when one person is already the villain. Decide it while everyone is still excited and no one is the no-show, and the awkward part is over before it starts. Four steps, in order:
Name the all-in number per person
“It’s $75 a head including the deposit” is collectible and clear. A specific number, stated up front, is what everyone is agreeing to — and it’s the figure the no-show rule will reshuffle later.
Borrow the restaurant’s cutoff
Restaurants draw a line — ticketed rooms like Next ask for at least 72 hours’ notice, after which a cancellation is final. Use the same line for the group: “cancel before Thursday and you’re off the hook; after that your seat is yours to fill or cover.”
Pick the rule for a late drop
State which of the three it is — split the seat, you eat your own share, or you transfer it. “If you bail last-minute, first try to send someone in your place; if you can’t, it’s on you” is a complete, fair policy in one sentence.
Collect before you book, not after
Money gathered up front is the price of a seat, not a debt you have to chase — and chasing is the part that wrecks friendships. Settling the share while the dinner is still in the future is also the moment people most willingly pay. splitty can turn the booking into a pre-filled request to each person, so collecting isn’t your job to nag about.
What if someone already bailed and there was no rule?
Stop trying to recover the deposit and recompute the dinner forward. When there was no agreement, relitigating it after the fact only makes the table choose sides over money that’s already gone. Do three quick things instead. First, offer the transfer: text the person who dropped — “totally fine! if you know anyone who wants the seat, send them, it’s paid” — which turns a loss into a save and costs no one anything. Second, if no replacement turns up, name the new number plainly: “with seven of us it’s about $86 each now — works for me, all good?” No blame, just the updated denominator. Third, let the rest go. The difference between splitting the gap and making the absent friend pay is usually a few dollars; the difference in goodwill is not. Decide the principle out loud next time, and the same cancellation never costs you an argument again.
The one line that defuses it: “The deposit’s gone either way — let’s just figure out tonight’s number and enjoy dinner.” It moves the table off the unrecoverable past and onto the only decision that’s still live.
How does splitty handle a dropped headcount?
splitty was built for the variable bill — the one computed from what people actually ordered — and a pre-paid dinner inverts that into a fixed total over a moving headcount. The useful part is the same either way: when the number of people splitting changes, splitty re-divides the real cost instantly and shows each person their exact updated share, so “it’s about $86 each now” is a figure read off the receipt, not a guess made in the parking lot. Each person gets a pre-filled request in their own payment app, which means the recount and the collecting both stop being one organizer’s emotional labor.
The honest limit: splitty doesn’t hold deposits, take reservations, or get anyone’s money back from a restaurant — that’s the restaurant’s system (and Tock’s) to run. And for a balance that runs across a whole multi-day trip, a standing ledger like Splitwise or Splid fits better. splitty’s job is narrower and exactly the one this problem needs: take the cost that’s really owed tonight and turn it into fair, pre-filled shares.
When is a pre-paid no-show no big deal?
When the gap is small, the group is close, or the seat gets filled. If a friend cancels a $30 brunch reservation among four people who never keep score, the extra few dollars each isn’t worth a single sentence of discussion — just split it and move on. If someone sends a replacement, the problem evaporates entirely. The cancellation only stings when it’s a big prepaid number, a vague arrangement, and a group that does quietly keep track — all at once. Take away any one of those and being the person who dropped out goes back to being a minor scheduling hiccup instead of a standoff. The fixed dinners worth booking are getting more common; the argument doesn’t have to come with them.
FAQ
Frequently asked questions
01 Do I get my deposit back if I cancel a group dinner reservation?
Usually not, once the restaurant's cancellation cutoff has passed. Deposits, card holds, and per-person no-show fees exist precisely so the restaurant keeps the money when a seat goes empty — in early 2026 OpenTable even added a 2% service fee that rides on no-show and deposit charges. Some restaurants do refund if you cancel early enough — ticketed rooms like Next, for instance, allow a reschedule with at least 72 hours' notice — so check the specific policy. After the cutoff, the deposit is a sunk cost: it's gone whether you attend or not, which is exactly why dragging yourself to a dinner you no longer want to attend doesn't actually save you anything.
02 Someone bailed on a pre-paid dinner — do the rest of us have to cover their share?
Only if you choose to, and it's one of three fair options. You can split the empty seat among everyone who showed (simplest for close friends and small gaps), let the no-show's prepaid share simply be forfeit to them (fairer for big groups or when a clear cutoff was missed), or — best of all — have the person who dropped transfer their seat to a replacement guest. The right choice depends on your group and how big the per-person gap is. What you can't do is recover the money from the restaurant, so the real decision is only ever about who absorbs the slightly larger share now.
03 Is it fair to make the person who cancelled pay their full share?
It's fair if the group agreed to that rule before booking, and it feels harsh if it's invented after someone cancels. The cleanest approach mirrors how restaurants handle it: set a cutoff in advance — 'cancel before Thursday and you're off the hook; after that your seat is yours to fill or cover.' Agreed up front, that's reasonable and everyone signed on. Sprung on a friend the morning of, the same rule reads as petty. The timing of the conversation matters more than the rule itself, so decide it while everyone is still excited and nobody is yet the no-show.
04 What's the best way to handle a deposit when someone drops out last minute?
Transfer the seat instead of splitting or eating the cost. Ticketed restaurants like Alinea and Next made dinner non-refundable but explicitly transferable — 'you may always transfer your reservation to another guest' — and your group can borrow the same move. Ask the person who bailed to send a replacement: a roommate, partner, or coworker who'd love the seat. The headcount returns to normal, the per-person cost snaps back to what everyone budgeted, and the money already spent actually buys a dinner someone eats. A non-refundable seat isn't a dead loss; it's a gift card with a date on it.
05 Why do I feel like I have to go to a dinner I already paid for, even when I don't want to?
That's the sunk cost effect, documented by Arkes and Blumer in 1985: once you've put money into something, you feel a strong pull to follow through, driven by not wanting to appear wasteful. In their classic study, theater patrons who'd paid more for season tickets attended more plays — not because they enjoyed it more, but because the larger sunk cost pulled them into the seats. A paid dinner deposit works the same way. But the money is gone whether you attend or not, so 'I already paid' isn't a reason to go — it's just a bias. Go if you'll enjoy it; the deposit shouldn't decide for you.
