I spent a Tuesday night shadowing a supervisor in a twenty-four storey tower off Wormwood Street, EC2M, with a team of nineteen and a single goods lift. By half past ten she had been in that lift eleven times. Two of those trips were productive. The rest were caused by a trolley left on eleven, a consumables shortfall on seventeen that turned out to be a cupboard nobody had unlocked, and a call from a fire warden about a propped door on six. She knew her building, her team liked her, and the operation was losing roughly ninety minutes a night to vertical travel that better planning would have removed. Nobody in the client’s facilities team had ever thought about the building as a queueing problem, which is the only way a tower can sensibly be thought about.

What sets the ceiling on output in a high-rise?

The goods lift. Not headcount, not method, not the machinery.

A single goods lift serving twenty-plus floors gives you a round trip of six to nine minutes once you account for door dwell, the operative loading a trolley, and the near-certainty that a courier or an out-of-hours contractor has called it to the loading bay mid-journey. Multiply that by the number of times your operation moves something vertically and you have the real productivity number for the building. Everything else – technique, machine choice, the price per square metre in the tender – sits downstream of it.

Count the lift movements. It takes one shift with a clipboard and it will tell you more about a tower contract than a year of audit scores.

The design answer is to stop moving things. Satellite stores on alternate floors, pre-stocked weekly during the day, eliminate the trolley-in-lift pattern entirely. Two square metres of cupboard on floors three, seven, eleven, fifteen and nineteen will outperform any amount of exhortation about efficiency, and landlords in the City give up lettable space for it reluctantly and then never notice the loss.

The goods lift arithmetic

Work it out for your own building. Round trip in minutes, multiplied by movements per night, divided by sixty. In that Wormwood Street tower the answer came to just over four hours of collective time per shift spent inside a steel box.

Waste is the largest single contributor. Consolidating each floor into one wheeled bin and descending once, at a scheduled time, beats individual sack runs by a margin that embarrasses the original method.

Should floors be cleaned by zone or by task team?

Zone cleaning gives one operative a defined area and the whole task list within it. Team cleaning splits the work by task – a vacuum specialist, a waste specialist, a washroom specialist, a utility person – who move through the building in sequence.

Team cleaning is oversold in Britain, and I say that having run it. It works beautifully in an American tower with uniform floor plates, generous storage and a lift dedicated to the cleaning operation. It works badly in a converted London building where floors seven and eight have a different plan from the rest, one tenant has restricted access, and the goods lift is shared with a restaurant on the ground floor. The sequencing that makes team cleaning efficient breaks the moment a floor cannot be entered on schedule, and in London a floor cannot be entered on schedule at least twice a week.

The hybrid that survives contact with a real building: zone ownership for general floors, specialist task teams for washrooms and for periodic works.

Washroom specialists earn their place because washroom quality drives every complaint that reaches a building manager, and because a person doing twelve washroom sets in a night gets faster and better at it in a way that a generalist doing two never will. The washroom specialist also carries one stocked trolley to one floor and works down the core using the stairs where the building permits it. Count the lift movements on that route against the generalist alternative and the case makes itself.

Where task teams fall over in a converted tower

Restricted floors. A tenant on eighteen with a fob-controlled lobby and a no-unaccompanied-access rule breaks a task sequence that assumes the vacuum team can follow the waste team by fifteen minutes.

Give restricted floors to a named, permanently assigned operative and take them out of the sequence entirely. It costs efficiency on paper and saves the schedule in practice.

How do you stop the supervisor spending the shift in a lift?

Write the route before the shift rather than improvising it during. A supervisor with a printed route card – floors in a fixed order, with the checks to be done on each and the time she should be there – moves through the building once. A supervisor responding to events moves through it six times.

Radio coverage in a tower is worse than anyone expects. Lift shafts, plant rooms, basement bin stores and the cores of some concrete-framed buildings will swallow a signal, and a team that cannot reach the supervisor will either stop working or invent a solution. Schedules solve what comms cannot. Fixed check-in points at fixed times mean the absence of a person at a place is itself the alert.

Count the lift movements again after you have done this. In a tower near Nine Elms, SW8, the same exercise produced a drop from thirty-one supervisor trips a week to eleven, achieved with a laminated card and nothing else.

The check-in point that does the most work

The bin store, at the end of the run. Everything that went wrong on the floors arrives there eventually, in the form of the wrong sack, the missing bin, or the operative who finished forty minutes early.

Stand there for twenty minutes at the same time each night. You will learn more than any digital audit app will tell you.

Audit scores in towers suffer from a bias worth correcting while you are down there. Supervisors inspect the floors that generate complaints, which are already the floors receiving the most attention, while floor sixteen with its quiet single tenant goes six months unexamined. Rotate the audit sample by lot rather than by instinct.

What does hybrid occupancy do to a five-night specification?

It breaks it, and the industry has been slow to admit as much because the admission costs money.

London office occupancy now concentrates hard into Tuesday, Wednesday and Thursday. A floor that took ninety people on a Wednesday takes twenty-two on a Friday. The waste volumes, the washroom consumption, the desk touch-points and the kitchen mess all follow that curve, and the cleaning specification almost never does. Five identical nights at an identical price is a fiction that both parties have agreed to maintain.

My position, which costs my own business revenue every time I state it in a tender: you should be paying less for Monday and Friday nights in most London towers. A reduced spec on the two light days – waste, washrooms, spot vacuuming, kitchens – with the full spec on the three heavy nights, redeploys hours to where the soil is. The building gets cleaner where it matters and the invoice goes down. I have made this argument nine or ten times and had it accepted four times, usually by facilities managers who had already worked it out and wanted a contractor to say it first.

The counter-argument from clients is presenteeism optics, and it is not a silly one. A tenant who comes in on a Friday and sees a lighter clean concludes the contract has been cut.

Washroom consumables and the Tuesday peak

Stock to the peak, not to the average, and check the towel dispensers on Wednesday afternoon rather than Wednesday night. A dispenser that empties at three on a Wednesday generates a complaint that no night shift can answer.

Occupancy sensors and access control data give you the actual curve for your building, and building managers in E20 and the City have that data sitting unused in a system nobody has connected to the cleaning contract.

Which risks change when the work is thirty floors up at midnight?

Evacuation, primarily. Most London towers operate phased evacuation, and a cleaning team distributed across twenty floors at one in the morning is the population most likely to misunderstand it. Operatives hear an alert sounder on an adjacent floor, decide it does not concern them, and carry on. Or they all descend at once through a single stairwell and meet the fire service coming up.

Fire door propping is the offence I see most in towers and the one with the worst consequences. A mop bucket wedged against a lobby door on floor twelve, for the twenty minutes it takes to run a corridor, defeats the compartmentation the whole building depends on. Every operative knows it is wrong. Every operative does it, because the alternative is a door that cannot be opened with a machine in both hands.

Lone working is the other change. A person cleaning floor twenty-nine alone at midnight, with a lift out of service and a radio dead spot, needs a check-in protocol that someone monitors rather than a policy document.

Fire doors, mop buckets and what to buy instead

Buy the door retainers that release on the alarm signal, the acoustically triggered kind that fit to an existing closer. They cost a fraction of a hard-wired system and they remove the temptation entirely.

Walk the escape routes with the night team, physically, in the dark, at the hour they will be working. A daytime induction in a bright lobby off Great Portland Street, W1W, teaches nobody anything about finding a stair core at one in the morning.

The supervisor at Wormwood Street now finishes her floor walk at eleven-forty and takes the lift twice a night. The cupboards on the odd-numbered floors got fitted out in February.