The queue and the fair share

A clinic replaced its single queue with an appointment system and the waiting room emptied. Complaints rose.

The arithmetic had improved. Average waiting time fell from fifty minutes to nine, and nobody now stood in a corridor. What people lost was something the old queue had quietly supplied: the ability to see that the rule was being followed. In a line, the fairness is visible. You watch the person in front of you go in, you know your position, and you can tell at a glance whether anyone has jumped. With appointments, the order lives inside a computer, and a patient who waits twenty minutes past their slot has no way of knowing whether the delay is bad luck, an emergency, or somebody's cousin.

Research on queueing describes two separate things that people want: to wait less, and to be treated in a knowable order. Systems optimised for the first often damage the second, and the resulting unhappiness is reported as 'the wait', because that is the only word available.

The clinic's eventual fix was cheap. A screen in the waiting area showed the next three appointment times and a line of text when the schedule slipped: 'running 15 minutes late — emergency admitted at 10:20'. Nothing about the waiting changed. Complaints fell by two thirds.

The general principle is that fairness has an evidence requirement. A process can be fair and feel arbitrary, and from inside the waiting room those are the same thing. A process that explains itself is doing a different job from a process that is merely correct, and both jobs have to be done.