Why Good Restaurants Close and Mediocre Ones Do Not

The announcement always lands on a Tuesday. It is always a photograph of a handwritten note taped to the inside of the door, and it always contains the same three sentences about an incredible journey and the best team anyone could have asked for.

Then the room goes dark, and within four months something else is in it selling a burger with truffle oil to people who are perfectly content, and that something else will still be trading in six years. This is not a conspiracy and it is not a tragedy of taste. My position is simpler and less romantic: quality is not the variable that determines survival. The survival function takes different inputs entirely, and food quality is a weak and indirect term in it.

The arithmetic that never makes it into the review

A restaurant is a business with terrible structural characteristics. Food runs somewhere around thirty percent of revenue. Labour takes another thirty or more once you count the people nobody thinks about, the ones who arrive at six to peel things. Rent and rates, if you are lucky, take six to ten. Then energy, insurance, waste collection, card processing, the repair on the walk-in that failed in August, and the accountant who tells you about all of it afterwards.

What is left, in a good year, in a well run independent room, is a net margin in the low single digits. That is the whole cushion. One slow quarter, one landlord with a rent review, one broken piece of refrigeration at the wrong moment, and the cushion is gone. The business was never solvent in the way people assume solvent businesses are. It was solvent the way a cyclist is upright.

Ambition is structurally expensive

Here is the part that makes the pattern feel unjust. Almost everything that makes a restaurant good also makes it more fragile.

• More cooks per cover, because precise food needs hands.
• More waste, because fresh produce in small batches spoils and because a kitchen with standards throws things away.
• Less cross-utilisation, because a short and specific menu cannot hide an ingredient in four other dishes.
• Fewer covers, because a serious room is quiet and the meals are long, and a table that turns once a night has to carry the whole evening.
• More training, which is a cost that walks out of the door whenever someone gets a better offer.

Every item on that list is a virtue. Every item is also a cost, and the costs land immediately while the reputation arrives slowly, if at all. The mediocre place upstairs has none of these problems. It buys frozen, holds a wide menu built from eight interchangeable components, turns the room three times, and needs nobody clever in the kitchen.

Survivorship bias, running in both directions

An objection I should deal with early: perhaps good places do not close more often, and we only think so because we notice them. That is a real effect. Survivorship bias normally describes overrating the survivors, but there is a mirror version here. Thousands of profoundly forgettable places close every year and nobody writes a farewell post about them, because nobody had noticed they were open.

I grant that this inflates the impression. It does not dissolve the argument, because the cost structure above is real and can be worked out on paper. The good room and the mediocre room are not playing the same game with different luck. They face different break-evens.

The strongest objection, which is genuinely strong

The best counterargument runs like this: running a viable business is part of the craft. A restaurant that cannot pay its cooks properly, that burns through staff and folds after eighteen months, was not a good restaurant. It was a good hobby that other people subsidised with their labour. Judging a place purely by what lands on the plate is a critic's luxury.

I concede a lot of that. More than I used to. I have worked for a brilliant operator who could not do numbers and the experience was not noble, it was chaotic and slightly humiliating, and the people who paid for it were the ones on the lowest wages. Costing a menu properly and paying on time are craft. That connects directly to the labour arguments in the tipping piece, which is the same problem seen from the payroll end.

But the objection proves too much. Pushed all the way, it makes survival the definition of quality, and then the sentence good restaurants close becomes incoherent by construction rather than false by evidence. That is a tautology wearing a suit. A place can cook beautifully and be badly run, and both halves of that sentence are true at once.

What the survivors are actually doing right

The part people skip, because it is unflattering to admit: the boring survivor is usually doing something real.

It does the same thing every day. Its menu was costed to work on a dead Tuesday in February rather than on a full Saturday. It tolerates boredom, which is an underrated professional skill and the reason its staff turnover is low. It is not chasing a review, so it never distorts itself to impress someone who was only ever coming once.

Consistency is a form of quality that no award exists for. The place I go to when I am tired and want no surprises is providing something genuine, and my sneering at it says more about my appetite for novelty than about its cooking. It is also, notably, not spending its energy on the story problem I described in the hype gap.

What I want is not for the boring places to fail. It is for the good ones to be a little harder to kill: longer leases, less punitive rent reviews, owners who learn spreadsheets before they learn plating. Beautiful food is not enough on its own. It never was. I just wish that were less true.