Waiting in Line

In the 1950s a New York office building received persistent complaints about slow lifts. Engineers concluded that little could be done without replacing the entire system at great expense. A consultant then suggested installing full-length mirrors beside the lift doors. Complaints fell sharply. The lifts were no faster; the wait simply felt shorter because people had something to occupy them.

The episode is often retold, and although details vary between accounts, it illustrates a principle now well supported by research: the perception of waiting time is only loosely related to its actual length. David Maister set out a series of propositions about this in 1985, several of which have been repeatedly confirmed. Unoccupied time feels longer than occupied time. Anxious waits feel longer than calm ones. Uncertain waits feel longer than known, finite waits. And unfair waits feel longest of all.

The last of these has practical consequences for how queues are designed. In a multiple-queue system, customers choose a line and frequently watch a later arrival in another line reach the counter first. Each such event registers as an injustice. A single serpentine queue feeding several servers removes this possibility: whoever arrived first is served first. Curiously, the serpentine arrangement usually produces a longer average wait for any individual, because a single slow transaction holds up everyone behind it rather than one line only. Customers nonetheless report greater satisfaction with it.

Information also alters perception. Airports and railway stations that display expected waiting times generally receive fewer complaints, even when the displayed figure is no shorter than the actual wait. Some operators deliberately overstate the estimate by a small margin, so that most customers are served sooner than promised. The practice has been criticised as manipulative, though its defenders point out that an accurate estimate which is occasionally exceeded produces more irritation than a cautious one that is usually beaten.

Physical design matters too. Houston airport once faced complaints about the wait for baggage. Rather than employ more handlers, the airport moved arrival gates further from the baggage hall, so that passengers walked for several minutes instead of standing still for the same period. Complaints dropped almost to zero. The total time from landing to luggage was unchanged.

Not every intervention works. Attempts to fill waiting time with advertising have produced mixed results, and there is evidence that entertainment which cannot be paused becomes irritating if the queue moves during it. What appears to matter most is whether the person waiting feels that the delay is explained, bounded and fairly distributed.