Room turnover window
A Socratic walk-through of the room turnover window — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a hotel leave its rooms standing empty for four hours in the middle of every day?
Checkout is at eleven. Check-in is at three. For four hours in the middle of every day, a substantial share of a hotel's rooms are unoccupied, unsold, and generating nothing — in a business whose entire economics rest on the fact that an unsold room-night is gone forever.
The obvious explanation is that the rooms need cleaning. That is true but insufficient, because cleaning a room does not take four hours. It takes something closer to half an hour. So the gap is roughly eight times longer than the task it exists to accommodate, and we should want to know what fills the difference.
Reasoning it through
REASONING #The mistake in the obvious explanation is treating one room as the unit. The hotel is not cleaning a room; it is cleaning all of them, with a crew of fixed size, and the binding quantity is crew-hours rather than minutes-per-room.
Work it out. Take a hotel with 150 rooms, of which say 120 turn over on a given day. At half an hour each, that is 60 hours of cleaning work. A housekeeping team of fifteen, each productive for four hours in the window, supplies 60 hours. The numbers meet — and they meet because the window was sized to make them meet, not the other way round.
Now the four hours is no longer mysterious. It is the total work divided by the crew size. Halve the window and the hotel needs thirty housekeepers instead of fifteen.
So why not employ thirty? Here the second constraint appears, and it is about the shape of the demand rather than its size. Cleaning work does not arrive evenly through the day. It arrives in one spike, bounded on the left by when guests leave and on the right by when guests arrive. Staff hired to compress that spike would be paid for a short shift and idle the rest of the day, or would have to be found other work. The window is long partly because labour is cheaper when it is used in a block.
There is a third element: variance absorption. The gap is not sized for the average day. Guests check out late, some rooms need more than a standard clean, someone calls in sick. If the window were sized exactly to the mean workload, roughly half of all days would end with rooms not ready at three, and a guest told their room is unavailable is an expensive failure — remembered, written about publicly, and possibly costing the booking.
Now notice the asymmetry that fixes the whole design. The cost of a room sitting empty for an extra hour is small and invisible — it is an opportunity cost on a slot that mostly could not have been sold anyway, since almost nobody wants to arrive at noon and leave the same evening. The cost of a room not being ready when a guest arrives is large and highly visible. Facing a lopsided pair of costs like that, the sensible operator buys slack. The four-hour window is that purchase.
And that explains the last odd feature: the times are fixed and public, not adjusted daily to the actual workload. A published time is a promise that lets guests plan and lets staff be rostered. A window that moved with demand would deliver a slightly better average and a much worse worst case.
The analogy
THE ANALOGY #Think of a school hall used for exams in the morning and a concert in the evening, with a single caretaking team to reset it in between.
Resetting one hall takes an hour. But the team also has three other rooms to reset, and the reason the concert starts at seven rather than at half past two is not that the hall takes four and a half hours to clear. It is that the same people must clear four rooms in sequence, that a chair may be broken and need replacing, and that a concert audience arriving to find the hall half-stacked is a far worse outcome than an empty hall standing idle for an hour.
The hall has no alternative revenue in the gap, whereas a hotel room genuinely could be sold to someone — day-use bookings exist — which is why the hotel's version of the gap is a real economic choice rather than a simple scheduling necessity.
Clarifying the model
THE MODEL #The window is a commercial decision wearing operational clothes, and the giveaway is that it can be bought. If four hours were a physical requirement, no amount of money would move it. But early check-in and late checkout are routinely sold, granted free to loyalty members, or offered when occupancy is low. That tells you the constraint is a scheduled average, not a limit — the hotel can produce a ready room earlier when it chooses to prioritise one, and it charges for the privilege because prioritising one room means reordering the queue.
Occupancy changes the answer, and this is testable. At full occupancy, nearly every room turns over and the crew is saturated, so the window binds. At low occupancy, far fewer rooms need cleaning, the same crew finishes early, and early check-in is handed out freely. The same hotel behaves as though it has a hard constraint on some days and no constraint at all on others, which is difficult to square with any account resting on the cleaning itself.
Not all of the gap is cleaning. Rooms are inspected, maintenance is scheduled into the vacancy, linen must be distributed, and the property system updated before the desk will release a key. Several are serial dependencies, so adding housekeepers shortens some parts of the window and not others.
What I am less sure of. How much of the four hours is genuinely crew-limited and how much is convention inherited from an era of different staffing and different labour arrangements is not something I can settle. The strong clustering of checkout and check-in times across properties of very different sizes is a little suspicious: if the window were purely derived from each hotel's own arithmetic, you would expect more variation than you see.
The falsification test. If the window is set by crew-hours against a workload spike plus a variance buffer, then it should shrink when occupancy falls, and hotels with proportionally larger housekeeping teams should offer earlier check-in as standard. If the window were identical at twenty per cent and ninety-five per cent occupancy, and identical across properties with very different staffing ratios, the operational account would be wrong and the answer would be pure convention.
A picture of it
THE PICTURE #How to readRead left to right as one ordinary day. The two guest bars are what the public sees; the gap between them is the question. The work bars show why it is not one task but a queue plus a serial inspection step that cannot start until cleaning does. The highlighted buffer at the end is the part that exists for the bad day rather than the average one — it is the slack that keeps a late departure or a broken shower from producing a guest waiting at the desk. Read the bar lengths as illustrative proportions, not as any hotel's measured schedule.
What became clearer
WHAT CLEARED #The window is not the time it takes to clean a room. It is the time it takes a fixed crew to clean every room, plus a deliberate buffer bought because the two possible failures are wildly unequal in cost: an empty room is a quiet loss, a guest with nowhere to go is a loud one. Once framed that way, the fact that the hotel will sell you an early check-in stops looking like an upsell on a technicality and starts looking like exactly what it is — the price of moving your room to the front of a queue.
Where to go next
ONWARD #- Why day-use bookings have not made the midday gap disappear despite filling exactly that slot.
- How housekeeping is scheduled when a property runs consistently near full occupancy.
- Why overbooking and turnover slack are managed as separate buffers against different failures.