The bus stop that moved fifty metres

Moving a bus stop fifty metres down the road cut use of that stop by a fifth. The buses were the same, the fare was the same, and the new shelter was better.

The fifty metres crossed a junction. Passengers coming from the estate now had to wait at a crossing with a long cycle, and on average the change added about ninety seconds. Ninety seconds is not a hardship. It is, however, enough to tip a decision for people whose alternative was walking the whole way, and the people with that alternative were the ones living closest.

Transport planners use the phrase 'access cost' for everything that happens before the vehicle arrives: the walk, the crossings, the kerbs, the sense of whether it is sensible to stand there after dark. Access cost is where most of the variation in whether a service gets used actually lives, and it is systematically underestimated, because it is invisible from inside the vehicle and does not appear in a timetable.

This explains results that look strange otherwise. A faster bus with a worse stop can lose passengers. Two routes with identical journey times can differ by a factor of two in use. A service can be improved and decline at once.

The cheap diagnostic is to walk the last two hundred metres yourself, at the time of day the service is meant to be used, and count the decisions you have to make. Every crossing, every unlit stretch and every place where it is unclear which way to go is a reason somebody with another option takes it.