Perceived lift waiting time

Why did mirrors beside the lifts silence complaints that faster motors never could?

A building's tenants complain that the lifts are too slow. The engineers do the obvious thing: faster motors, smarter dispatching, perhaps another car. The complaints continue. Then someone hangs large mirrors in the lift lobbies, and the complaints stop.

Before deciding what that story means, it is worth being careful about it. The mirror anecdote is retold constantly in service-design and operations circles, usually attached to an unnamed mid-century office tower, and I have never seen a primary source for it. Treat it as a parable rather than a case study. But the effect it dramatises is well documented, and the question it raises is a real engineering question: if the measured wait did not change, what did?

Start by noticing that the complaint was never about seconds. Nobody in the lobby is holding a stopwatch. What they report is a feeling of having waited too long, and that feeling is the actual output of the system as far as the tenant is concerned. So we have two different quantities that we habitually treat as one: the wait the building produces, and the wait the person experiences.

If those are two quantities, then in principle each has its own inputs. What might feed the second one that does not feed the first?

Try a thought experiment. You wait ninety seconds for a lift in a bare corridor with nothing to look at. Then you wait ninety seconds while reading a notice board that happens to interest you. Same ninety seconds. Which felt longer? Almost everyone answers the same way, and David Maister set this out in 1985 as the first of a set of propositions about waiting: unoccupied time feels longer than occupied time. Several of his other propositions matter here too — uncertain waits feel longer than known finite ones, unexplained waits feel longer than explained ones, and anxiety makes any wait feel longer.

So what does a mirror actually do? It gives the waiting person something to attend to, and specifically something people will attend to almost involuntarily — their own appearance, and, more subtly, the other people in the lobby, whom a mirror lets you look at without being caught looking. The interval is now occupied. Nothing about the lifts has changed at all.

Now push on why the engineering fix disappointed. Faster motors reduce the measured wait, and reducing the measured wait does reduce the felt wait somewhat — the two are not unrelated. But look at the shape of the two costs. Cutting the average wait meaningfully in an existing building means new machinery, possibly a new shaft, and it buys you a handful of seconds. Occupying the wait costs a mirror. And here is the sharper point: perceived duration is not proportional to real duration. A short but empty, uncertain interval can feel worse than a longer, occupied, predictable one, so the cheap intervention can outperform the expensive one on the measure that generates the complaints.

Is this a trick played on the tenants? That is the honest objection, and it deserves a straight answer. It would be a trick if the felt wait were an error to be corrected and the real wait were the truth. But the tenants' complaint was about their experience, and the mirror genuinely improves their experience. What would be a trick is using perception management to conceal a service that is actually failing — a lift that regularly strands people is not fixed by a mirror, and the same mirror that soothes a two-minute wait will not soothe a ten-minute one.

Notice, finally, that the same reasoning explains the modern versions of the mirror. A display showing which car is coming and where it is converts an uncertain wait into a known finite one. Destination-dispatch panels in the lobby, which assign you a car when you enter your floor, do something more interesting: they genuinely reduce travel time by grouping passengers, and they also give the waiting person information and a sense of having already started. Two mechanisms at once, which is why it is hard to attribute the improvement to either.

The misconception worth correcting is that "perceived" means "not real". The perception is the outcome the building's occupants actually live in, and it is measurable — you can ask people to estimate how long they waited, and their estimates respond systematically to occupancy, certainty and explanation while the clock does not.

The second refinement is that this is not a licence to stop measuring. Perceived and actual wait are coupled: past some point no amount of occupation rescues a genuinely inadequate lift bank, and a building that lets its real service degrade while decorating the lobby will find the mirror stops working. The right reading is that you now have two levers instead of one, with very different price tags, and that ignoring the cheaper lever is the mistake the original engineers made.

Third, the general principle transfers. Any system where people wait — call queues, baggage claim, a loading spinner, a hospital waiting room — has both quantities, and the design question is always which lever is cheaper for the improvement you need. Airports that moved the arrivals gate farther from baggage claim did not speed up the baggage; they converted standing time into walking time.

• Destination-dispatch control, which shortens the real wait and the felt one by different mechanisms.
• Whether occupying a wait works when the occupation is obviously artificial and recognised as such.
• The ethics of perception management when the underlying service really is inadequate.

Every term the collection defines is gathered in the glossary.