Stand on one side, walk on the other, and the argument against it

A rule that nobody wrote and everybody follows

On busy escalators in a great many cities, people stand on one side and leave the other clear for walking. Nobody is taught this. There is rarely a sign, and where there is one it usually confirms a practice that already existed rather than establishing it.

It is a good example of a convention that behaves like a rule. Breaking it produces sighs, a tapped shoulder, occasionally a word — a whole enforcement apparatus operating without any authority behind it. And the convention is genuinely local: the walking side differs between countries, and in some cities the practice barely exists, which is a reliable way for a visitor to look briefly foreign.

Why it emerges without being organised

The mechanism is unremarkable. Two populations use the same device at different speeds, and mixing them is worse for both. Separating them requires only that each person can see which side the people ahead have chosen, which on an escalator is trivially easy.

That visibility is the whole reason the convention is stable. A newcomer with no knowledge of the rule sees a column of people on one side and a gap on the other, and copies it within a step or two. The convention therefore reproduces itself from its own appearance, which is why it survives complete turnover of the population using it and why it does not need enforcement to persist.

The throughput argument against it

Here is where it gets contested. On a long and heavily used escalator, the standing side is packed and the walking side is not, because comfortable walking needs a gap between people that standing does not. The result is that a substantial share of the surface carries a small share of the passengers.

Where demand is high enough, having everybody stand can move more people per minute, and trials of standing-only escalators have been run in busy transit systems for exactly this reason. The trade is real and it is not free: those who would have walked arrive later, and they are the ones who notice. Total throughput improves while the experience of a subset gets worse, which is a difficult thing to explain on a platform and a large part of why the idea has never spread far.

Height and length change the answer

The argument is not uniform across escalators, and this is often lost when it is repeated. On a short rise, most people will walk anyway, so the two sides fill similarly and the convention costs little. On a very long rise, few people walk the whole way, the walking side empties, and the imbalance becomes severe.

So the honest position is that the best arrangement depends on length, crowding and the mix of people using it, and that a rule which is right in one station can be wrong in the next. That is an unsatisfying conclusion and it is the correct one. It also explains why transit operators are cautious: a convention people follow reliably has real value, and replacing it with one that applies only sometimes risks getting neither.

Why the convention resists change

Conventions of this kind are held in place by expectation rather than by belief. Nobody stands on one side because they have concluded it is optimal; they stand there because they expect everybody else to, and being the one person who does otherwise is uncomfortable regardless of the arithmetic.

That makes unilateral change nearly impossible and coordinated change quite hard. It also explains why signs asking people to stand on both sides are frequently ignored while staff standing at the bottom asking directly are much more effective. The instruction is not what changes behaviour. Visible evidence that others are complying is what changes behaviour, and only a person present can supply that in real time.

The smaller courtesies attached to it

Around the main convention sits a set of finer ones. Luggage goes on the step in front rather than beside you. Groups compress into single file as they approach. A person who intends to stand does not step on immediately behind somebody walking, because the gap they are filling is that person’s stopping distance.

None of these has ever been written down either, and all of them are noticed when broken. That is roughly the definition of a working convention: invisible when followed, obvious when not, and enforced by nothing more than the mild collective discomfort of several strangers who will never say anything about it.