How many rooms a management company can run without quality slipping is decided by how many of them move inside the same five hours on the same day, not by the monthly total. Our own ceiling is five properties, and it sits there because of whether a person can physically reach the door, not because of what the software can process.
Here is the evidence from our own portfolio. Over 365 days across three buildings and six rooms there were 486 checkouts, an average of 1.33 rooms a day. That average is not what happens on site. The day-by-day figures below are our own reservation data for October 2025 to September 2026 recounted one day at a time; the monthly aggregates, occupancy and cost lines included, are published at operating data (measured 1 October 2026).
How many rooms can an operator run without quality slipping?
As many as one team can get inside on the same changeover afternoon. The monthly count is the wrong unit. Counting our six rooms day by day across 365 days, the checkouts fell like this.
The mean is 1.33 rooms a day, but the days that need staffing are the 51 with three or more rooms (14.0%) and the 15 with four or more (4.1%). The peak was five rooms, which happened twice in the year. So the peak ran about 3.8 times the mean, and a roster is built for the peak. In the other direction, 88 days (24.1%) had no checkout at all. That is why a quiet week is no evidence that a company has capacity for one more property.
The method, so it can be repeated. Rooms per day counts reservations whose departure date falls inside the 365 days, counted per room, so a day when two rooms were let as one booking counts as two. The 489 bookings we publish are the same reservations cut by arrival month, which is why that figure and these 486 do not match. Any operator can produce this table from their own PMS with the same two lines of counting.
| Rooms checking out in a day | Days | Share of 365 days |
|---|---|---|
| None | 88 | 24.1% |
| One | 132 | 36.2% |
| Two | 94 | 25.8% |
| Three | 36 | 9.9% |
| Four | 13 | 3.6% |
| Five | 2 | 0.5% |
Why same-day overlap and not the monthly total?
Because a changeover has a deadline that day, and the work cannot be spread across the month. Our checkout is 11:00 (10:00 at one room) and check-in is 16:00, so the usable window is five hours.
And 322 of those 486 checkouts (66.3%) were followed by an arrival in the same room on the same day. Two changeovers in three have no option of tomorrow. An average stay of 3.3 nights and 83.7% occupancy, the monthly figures, show none of that deadline. What one changeover actually contains is broken out in cleaning, on site.
| Time | What has to fit in the five hours | Where the overflow shows up |
|---|---|---|
| 10:00, 11:00 | Checkout, check for belongings left behind | A call to the previous guest, not the next one |
| 11:00 to 16:00 | Clean, linen, restock, inspection and photographs | The last room of the day gets the rushed version |
| 16:00 | The next guest arrives | Pre-arrival messages, then a guest who cannot get in |
| That evening | Keys, equipment, a neighbour calling | Nobody rested for tomorrow's changeovers |
Is the bottleneck software or people?
People moving. Messaging, daily pricing and monthly accounting run on the same machinery whatever the property count; a changeover clean and a trip to the building at 23:00 need one more person for each one added.
Travel also refuses to cross cities. In our two cities, checkouts landed on the same day in both on 104 days (28.5%). Nobody stands in Tokyo and Sapporo on the same afternoon, so each city has its own team, and the same fact is why we take properties only in the 23 wards of Tokyo and in Sapporo. Seen from the cost side, cleaning is 17.0% of revenue, our single largest line, inside a published operating total of 27.8%. A conversation about an operator's staffing is a conversation about the biggest cost line there is.
How is a ceiling of five properties counted?
Not in properties, but in rooms in the same city and in the days where changeovers collide. Five properties can be five rooms or fifteen, so the property count on its own carries no information.
In our own portfolio, split three rooms and three rooms, Tokyo had two or more rooms checking out on the same day on 46 days and Sapporo on 52. With three rooms in one city, a team handles two rooms inside the same five hours roughly four times a month. At a viewing we look at four things, and the honest consequence is that the answer is sometimes fewer than five properties. We say so at the viewing rather than afterwards.
- City and travel time: can the existing team reach it on an afternoon that already has a changeover?
- Rooms: one room, several, or a whole floor let as one (our Katsushika building is let as a single booking across two rooms on some days)
- Checkout and check-in times: is the window set shorter than five hours?
- Collision: how many additional three-room days does this property create?
If the rooms double, do twice the people cover it?
That is decided by the peak, not the mean. Going from six rooms to twelve takes the daily average from 1.33 rooms to about 2.7, but what the peak becomes is not knowable in advance. Across our own year the peak ran 3.8 times the mean, and the roster has to survive that day, not the average one.
Which is why we stop intake at a stated ceiling rather than at a forecast. When a forecast is wrong, the cost is not paid by us: it is paid by whichever owner's room moved down the queue that afternoon. The problem is not a company without a ceiling, it is a company that has never counted one. Seven questions get at it.
What should you ask an operator before signing?
Seven questions. Ask every one of them about your city and this month, never about the company as a whole. A head-office property count will not tell you where your room sits in the queue.
- How many rooms does one cleaning team cover in my city right now?
- On how many days a month do three of your rooms check out together? Do you count it?
- On that day, who cleans the third room? A named person, or whoever is available?
- Who goes to the building at 23:00, and from where? How many minutes away?
- How many hours sit between checkout and check-in, and what share of changeovers are same-day?
- When a cleaner falls ill on a busy day, how is the order decided?
- What is your intake ceiling, and do you keep taking properties once you reach it?
Where does a stretched operator show first?
In cleaning remarks clustering on particular days of the week. A roster past its ceiling fails on the overlapping days first, not evenly. Our six rooms hold 398 reviews at 4.9, and what we defend is not the star rating but the causes that pull it down.
Four symptoms are visible from the owner's side. Which clauses tell you what leaving will cost is set out in reading a management contract.
- Cleaning remarks bunch on the same weekday or after a long weekend
- Replies are slow between 11:00 and 16:00 and quick again in the evening
- The monthly report arrives late, or shrinks to a single total line
- The answer to "who went?" stops being a name and becomes "our staff"
When not to hire us
If you want twenty properties handed over at once, or your property sits outside the 23 wards of Tokyo and Sapporo, we are the wrong company. The ceiling of five is published for a plain reason: the day we take a sixth, an existing owner's room moves down the queue. The framework comes first too. We hold no registration as a residential accommodation management operator (住宅宿泊管理業者), so we take on properties running under the hotel-business permit (旅館業許可) or working toward it.
Every figure here is the record of our own three buildings and six rooms, not a level offered for a property we have not seen. What we promise is not an occupancy number but that a day with three checkouts is run the same way we run it in our own six rooms, and that the staffing behind it is counted and shown.
Ask about your property