The nightly rate on our six rooms is reviewed every day, date by date. We keep no seasonal rate card. The decision is ours, the people operating the rooms, and the tool only proposes.
What follows is our own record as of 1 October 2026, covering the 489 reservations that arrived in the twelve months from October 2025 to September 2026, with the method published on our operating data page. No amounts: everything here is a share, a count of nights or a number of days.
So who actually decides the price, and how?
A person does, every day, date by date and room by room. A dynamic pricing tool proposes a rate for each future date from the way bookings have been arriving, and we decide whether to take it or override it.
Dynamic pricing means moving the nightly rate of the same room from date to date with demand. It is a different thing from switching a fixed rate card at the start of a season: inside one week the price moves with the day and with how many days are left.
Four things get looked at daily, a few minutes per building.
- How much of each of the next ninety dates is sold, judged against the same room at a comparable number of days out rather than against a calendar date
- Where a gap of one or two nights has opened with reservations on both sides. Those are a different problem from ordinary empty nights
- The asking rates of comparable rooms, similar capacity and size, within the same walk of the same station
- Whether a local event, a public holiday or a school break falls on that date
Why no seasonal rate card?
Because the card fixes a year of prices to the assumptions held on the day it was written, and when demand moves away from them the rooms keep selling at the old number.
The swing shows as soon as our twelve months are laid out by month: 18.6 points across the six rooms, 35.2 points for the three Tokyo rooms alone. The two cities also move in opposite directions in the same month. October 2025 was Tokyo 95.2% against Sapporo 80.6%; September 2026 was Tokyo 60.0% against Sapporo 83.3%.
So the 83.7% average fits neither city's calendar, and an owner with one building has no such cancelling out (seasonality). What we keep fixed is not the price list, it is the order the decisions are made in.
| Scope | Best month | Worst month | Spread |
|---|---|---|---|
| Our six rooms | 90.3% (October 2025) | 71.7% (September 2026) | 18.6 points |
| The three Tokyo rooms | 95.2% (October 2025) | 60.0% (September 2026) | 35.2 points |
| The three Sapporo rooms | 94.0% (February 2026) | 68.3% (November 2025) | 25.7 points |
What does a 42-day median do to the order of the work?
It means the price you change today mostly sells the month after next, not next month.
Across those 489 reservations the median gap between booking and arrival was 42 days, exactly six weeks. Half arrive earlier than that, so next month's occupancy is already half decided before next month starts (booking lead time).
The weight of the daily review therefore sits 30 to 90 days out, not on the coming week. Occupancy of 83.7% is 16.3% empty, about 4.9 nights per room in a 30-day month, and those nights are not one block: they are ones and twos sitting between stays.
Which levers move, and in what order?
The first lever is not the price. It is the minimum stay and how gap nights are handled.
They work differently. A price set on a date lands on every night of any booking that crosses it, while minimum stay and long-stay discounts change the shape of the booking itself. For one empty night with reservations on both sides, dropping a two-night minimum is usually cheaper than dropping the rate.
The long-stay discount has a boundary. Merging two 3.3-night bookings into one of 6.6 nights saves one changeover, and at 83.7% occupancy with an average of 3.3 nights a room turns over 7.6 times a month, so with cleaning at 17.0% of revenue one changeover is worth about 0.56 nights of rate. The discount applies to all 6.6 nights, so 0.56 / 0.066 puts the boundary near 8.5%. The previous twelve months gave 8.8% (an average stay of 3.3 nights). Redraw it from your own cleaning cost and your own average stay.
| Lever | What it changes | When it moves |
|---|---|---|
| Minimum stay | The shape of the booking that can land on that date | When one and two-night gaps are left behind |
| Long-stay discount | How easily longer stays arrive | Quiet season, when changeovers need reducing |
| How far ahead the calendar is open | How much demand you can even see | When long lead-time bookings are being missed |
| Photographs and listing copy | Being chosen once you are seen | Any time, but it lands beyond 42 days out |
| The nightly rate | How fast a date sells | After the four above |
Why does discounting come last?
Because a discount is set on a date, but it lands on every night of the booking that crosses it.
Cut the rate on one empty night and you cut it on the 3.3 nights an average booking runs. A d% discount gives away 0.033d nights of rate, so against selling one empty night it breaks even at 1 / 3.3 = 30.3%. If the booking would have arrived at full rate anyway, the discount is pure loss.
Cutting the whole rate card is starker. Higher occupancy also means more changeovers, which makes the largest cost line heavier, so occupancy only means something next to the average rate (occupancy).
| Rate cut | Occupancy needed to hold revenue | Points to climb from 83.7% |
|---|---|---|
| 10% lower | 93.0% | +9.3 points |
| 15% lower | 98.5% | +14.8 points |
| 20% lower | 104.6% | Unreachable |
When does a person override the tool?
Only when there is something the tool cannot see. We never override to make an occupancy figure look better.
- A fault in the room that is not fixed yet, on dates we would rather not sell
- Building work or a neighbourhood event, where the dates need handling quietly
- A long-stay enquiry in progress, where the dates around it are worth holding
- An event announced too recently to have reached the surrounding prices
- A week when the cleaning team is short-handed and cannot absorb more changeovers
What does the owner get to see?
Occupancy and average daily rate, side by side, in the monthly report, always.
Either one alone can be manufactured: occupancy can be bought with a discount, and the average rate rises nicely if you hang up a price nobody pays. Only the pair shows what was done that month (reading a monthly P/L).
With them go the next 90 days of pace, what we changed on price that month and why, and the floor rate. The floor is agreed when the contract is signed and we do not go below it on our own judgement: a discount moves more than an owner's income, it moves which guests the room attracts (guest ratings).
What should I ask a management company about pricing?
Ask these five before the fee percentage. Every one of them is answerable before a contract is signed.
- How often the rate is reviewed: daily, weekly, or a seasonal rate card
- What proposes and who decides. Is the tool's output simply passed through
- Who sets the floor rate, and whether it can be lowered without the owner's agreement
- Whether the monthly report carries both occupancy and average daily rate, and the next 90 days of pace
- What moves before the price does when a date is not selling. If the only answer is a discount, there is one lever in the building
Can you just do the pricing?
No. The rate sits on the same circuit as the number of changeovers, the minimum stay, how guests are answered and what they then write in a review, and holding one end of it alone does not move the result. What moves is the operation, and price is one lever inside it.
We work in Tokyo's 23 wards and the city of Sapporo, with properties that hold a hotel-business permit (旅館業許可, the framework a hotel runs on, with no 180-night cap) or are going to get one. We cannot run a property that stays under the 180-night notification framework (permit or notification). We hold at most five properties at a time and we do decline after a viewing (the properties we turn down).
The fee is performance-based from 20% of revenue, with nothing fixed to pay in an empty month. The figures here are our own record, not a promise of occupancy or rate for another property. How to compare fee models is in what management costs, and how to read the contract is in reading a management contract.
Send the address and the number of rooms for a read on pricing