The queue that was a signal

A restaurant with a nightly queue reduced its seating by a quarter. Revenue rose.

The decision looks irrational and rests on something real. A visible queue is information, and in a market where quality is hard to judge in advance, information of that kind is valuable. A passer-by cannot taste the food, cannot evaluate the kitchen, and is choosing between a dozen places on one street. A line of people who have already chosen is the cheapest available evidence, and it works because joining a queue is costly: nobody waits forty minutes as a favour.

Economists call this a costly signal, and the cost is what makes it credible. Claims are cheap, so a sign saying 'best noodles in the city' carries almost no information. Waiting is expensive, so a queue carries a lot. The restaurant that cut its seating was converting capacity into credibility, and the trade was favourable because the street was crowded with alternatives and nothing else distinguished them.

The strategy has limits that are easy to state. It fails where customers have other information, since a place with reliable reviews gains nothing from a line. It fails where waiting is impossible, as at a station. And it is fragile: a queue that forms because of a queue can disperse the same way, and a single slow evening removes the signal entirely.

For a customer the useful reading is narrow. A queue is evidence that other people, with roughly your information, chose this place. It is not evidence that the food is good, and the two are different claims. Where you can get better evidence cheaply, the queue should be ignored.