Supply volume as background — 1,739 openings averaging 24 rooms in 2025, of which 374 with 50+ rooms are the practical universe

Supply volume has been covered in earlier articles, so it is addressed here only insofar as it defines the sample. Within the scope tracked by MetroEngines Research & Consulting, 1,739 properties were observed opening in 2025 (based on confirmed OTA listings), averaging 24 rooms. The breakdown is 745 vacation rentals, 265 business hotels, 122 guesthouses, 113 cottages, 99 resort hotels, 98 hostels and 59 ryokan, among others — on a count basis, small-scale, dispersed inventory makes up the large majority. For 2026, 1,284 properties averaging 27 rooms have been observed so far.

What actually enters a business plan or an investment decision, however, is the subset above a certain scale. Restricting to 50 rooms or more leaves 166 hotels in 2024 and 208 in 2025, for a combined 374 — barely more than 20% of the count-based universe. This article takes those 374 as its starting point. Note that opening counts are based on confirmed OTA listings; because listings appear several months before opening, recent months and years will continue to rise as further listings appear. The 1,284 figure for 2026 is likewise not final.

Study design — normalized to the market level, with the opening month excluded

Simply asking whether ADR is low immediately after opening mixes in seasonality and market-wide rate inflation. A July opening and a February opening have entirely different absolute levels in their first month, and the 2024 cohort simply rides the market’s rise through 2025. We therefore converted each property’s monthly estimated transacted ADR into a market-relative index by dividing it by the median estimated transacted ADR for the same prefecture and same property type. An index of 1.20 means that in that month the property sat 20% above a typical property of the same type in that market.

Next, we defined each property’s median index over months 12-17 after opening as its steady-state level, and normalized the entire series to it. The comparison is therefore neither the market nor the competitive set but the property’s own profile one year later — a design intended to isolate the opening effect alone.

One methodological point deserves emphasis. The month of opening itself was excluded from the analysis. In the opening month, the median number of days on which a price was observed is 18, well short of the 30-31 days of a normal month. Including that partial month produces an opening-month index of 0.928 (-7.2%), but restricting to properties with 25 or more observation days brings it to 0.998 — the gap essentially disappears. In other words, the “steep opening-month discount” is an artifact of the observation window, not a real phenomenon. From here on, the month after opening is treated as month 1, and only months with 25 or more observation days are used.

182 hotels satisfy these conditions and have a determined steady-state level (138 business hotels, 27 city hotels, 16 resort hotels and 1 ryokan; 126 opened in 2024 and 56 in 2025), giving 4,138 observed hotel-months.

At the median, no “first-year discount” is observed

To lead with the conclusion: across 182 hotels, the median market-relative index for months 1-3 after opening is 1.006 (+0.6%) versus the steady-state level — there is no meaningful discount. The first-year average (months 1-11) is 1.011, again effectively identical to the steady-state level. Median time to recovery is 2 months, and 48% of properties are already at their steady-state level in the first month after opening.

Broken out by month, the index is 0.964 in month 1, 0.997 in month 2 and 1.020 in month 3: there is a slight dip in the month after opening, but it resolves by month 2. (Taking each property’s own median across months 1-3 gives the 1.006 cited above.)

In yen terms, the median estimated transacted ADR for months 1-3 across the 138 business hotels is approximately ¥8,100, against roughly ¥8,300 that the same property would have commanded at its steady-state level in the same month and same market — a gap of about ¥200, or -2.5%. That is an order of magnitude away from the conventional “-15% in year one” assumption.

The interquartile range shows that behind this flat median, individual properties vary widely. For months 1-3 after opening, the first quartile is 0.885 and the third quartile 1.096; the bottom decile is 0.768 and the top decile 1.188. This is a market in which the average property does not discount, but any individual property can move +/-20%. What matters for a business plan is not the median but which side of that distribution the property lands on.

The distribution is the real story — 39% open at -16.5%, 33% at +12%

Classifying properties whose months 1-3 index is below 0.95 of steady state as “discount openings,” those above 1.05 as “premium openings,” and the rest as “at-par openings,” the 182 hotels split as follows.

The discount-opening group (71 hotels) starts month 1 at 0.835 of steady state — that is, -16.5% — and recovers gradually over the following ten months. Months to recovery run 5 at the first quartile, 8 at the median and 12 at the third quartile; for this group alone, the conventional assumption of discounting in year one is broadly reasonable. In yen terms, the median estimated transacted ADR for months 1-3 is about ¥8,700 against roughly ¥10,600 at the steady-state equivalent — a gap of about ¥1,900 per room.

The premium-opening group (61 hotels) moves in the opposite direction. It launches at +12.4% in month 1, holds around +10% through month 7, and then declines toward steady state over months 12-13. The price premium from novelty and buzz immediately after opening is real, but it is the kind of premium that disappears within a year. In a business plan, the elevated ADR seen just after opening must not be extrapolated into year two and beyond.

That these three groups are not merely an artifact of measurement noise can be confirmed from the shape of their convergence. During the stabilized period (months 12-24), month-to-month variation in the market-relative index has a median of 7.3%, so a single month’s observation is far from quiet. But if the split were pure noise, the groups would revert to the mean the following month and the difference would vanish. In practice the discount group recovers monotonically over ten months, and the premium group holds +10% for seven months before decaying. That persistence cannot be explained by monthly dispersion. The group assignment itself, moreover, is based on the median of months 1-3 rather than any single month.

Distribution of months to recovery — half within 2 months, 9% still short after a year

The distribution of months to recovery (the first month in which the three-month moving average reaches 98% of steady state) is strongly right-skewed: 1 month at the first quartile, 2 at the median, 7 at the third quartile, 12 at the ninth decile, with a mean of 4.2 months. Cumulatively, 48% reach steady state by month 1, 63% by month 3, 74% by month 6 and 91% by month 12.

Translated into a business plan, using the single figure of 4.2 months is not appropriate. It is closer to reality to carry the bimodal shape straight into the scenarios: half the properties sell at the steady-state rate immediately, one in four takes seven months or more, and one in ten takes a year.

What determines the discount side — not local supply volume, but the price position being targeted

Can we predict which properties fall on the discount side? Conventional cuts such as property type, room count and metro versus regional location produced almost no difference.

Only one variable clearly mattered: where in the market’s price band the property ultimately lands. Splitting the 182 hotels into terciles by their steady-state market-relative index (the multiple of the market median at which they sell during the stabilized period) gives the following.

Properties aiming to sell at twice the market median post a month-1 index of 0.910, roughly 9% below steady state; nearly half are classified as discount openings, and they take a median of three months to reach steady state. Conversely, properties targeting the more accessible 0.84x position are already selling at close to their steady-state rate from month 1 (0.982). The higher the position a property reaches for, the longer the run-up needed before the reputation and track record that justify that position accumulate — a result consistent with practitioner intuition. Setting the ADR ramp assumption in a business plan as a function of how far above the market the property is aiming, rather than of property type or room count, improves accuracy.

One result ran counter to intuition. The volume of new supply nearby explains almost nothing about a new property’s own opening price. Correlating supply intensity by prefecture — new rooms added in 2024-2025 divided by existing rooms — against that prefecture’s median months 1-3 index gives r=0.148 (n=18 prefectures), essentially no relationship. The distribution of that same “share of existing stock” measure taken down to the municipal level is set out in Hotel Supply Pressure Ranking 2026-2027.

Osaka, with the highest supply intensity at 13.7%, shows the strongest launch of all at 1.097 for months 1-3. Hokkaido follows with a supply intensity of 8.0% and an index of 1.139. Conversely Aichi, restrained at 4.3% supply intensity, is the weakest at 0.879. The reading is that pricing latitude at opening is set not by the absolute volume of new supply but by the growth in demand absorbing that supply. In development decisions, the premise that “this area is in an opening boom, so year one has to be sold cheap” is not supported by the last two years of observed data.

What about the demand side, then? We re-measured the “growth in demand absorbing supply” referenced above using the same series as this analysis. For each of the 18 prefectures, we compared the 12-month average of the prefecture’s median estimated transacted ADR for 2024 against the same figure for 2025, and used that growth rate as a proxy for how far the market was able to push rates through (the demand side’s absorption capacity), testing it against the months 1-3 index.

The direction matches the hypothesis: prefectures where market prices grew show higher opening positions. But the effect is weak. The correlation between growth and the opening index is r=0.251 (n=18 prefectures) — higher than the r=0.148 for supply intensity, yet still close to no relationship. Across terciles the medians run 1.014 at the top versus 0.990 at the bottom, a difference of only 2.4 points. The correlation between supply intensity and market price growth is itself r=0.320, meaning the obvious fact that stronger-demand prefectures also receive more supply is all that thinly links the two. Read faithfully to the data, prefecture-level macro supply and demand — on both the supply and demand sides — explains almost nothing about an individual property’s opening price. There is little room to deduce a business plan’s ADR ramp assumption from area-level supply and demand; the property’s own target price position, discussed in the preceding section, is decisively more powerful. Note also that each prefecture covers only 4 to 26 properties, so prefecture-level values are volatile.

Building it into the business plan — the ADR assumption alone moves first-year yield by 0.65pt

What happens when the observed values are reflected in a business plan? Assume a 150-room business hotel with a stabilized estimated transacted ADR of ¥8,300 (the steady-state equivalent level for the 138 business hotels in this analysis), occupancy of 60% in the first year of operation and 80% at stabilization — the conventional opening ramp-up assumptions in DCF-based real estate appraisal practice — and a GOP margin of 45% (the lower end of the 40-60% guideline range for limited-service hotels). Construction cost is taken as ¥2.405 million per tsubo for steel-frame construction in 2025, based on MLIT’s Statistical Survey on Building Construction Starts, and at 28 sqm of gross floor area per room (1,271 tsubo) gives approximately ¥3.06 billion (excluding land). How the rise in that per-tsubo cost itself distorts future supply plans is covered in detail in Construction +41% & Labor Crunch Erase Japan’s 2027-2029 Hotel Supply.

Case B’s +1.1% is the median of first-year (months 1-11) averages across the 182 hotels; Case C’s -7.3% is the first-year average for the 71 discount-opening hotels. Between conventional Case A and median Case B, first-year room revenue differs by ¥43.9m and GOP by ¥19.8m, translating into a 0.65-point gap in first-year yield on construction cost. For the same property and the same occupancy assumption, the ADR ramp assumption alone moves the result that far.

More important as a sensitivity is the spread between the discount-opening and premium-opening groups. Their first-year averages are 0.928 versus 1.090; that 17.5% difference equates to ¥44.2m in first-year room revenue and 0.65 points of first-year yield. This spread is the single largest risk factor in the opening year, and a single discount rate as in Case A cannot express it. In practice, the closest fit to what this analysis observes is to apply Case B or Case C according to the target position relative to the market median, and to treat the upside (premium opening) as temporary — assuming a return to steady state within 12 months.

Two-axis sensitivity: ADR assumption x occupancy — the wider swing is on occupancy

All three cases above fix first-year occupancy at 60%. In practice that 60% is itself only an assumption, so the ADR assumption and occupancy need to be flexed together on two axes. The other conditions (150 rooms, GOP margin 45%, construction cost approx. ¥3.06bn) are the same as in the table above.

There are two things to read here. First, the range of first-year yield driven by the ADR assumption (¥7,100 to ¥9,000) is 0.92 points at 60% occupancy. Second, holding ADR constant, moving occupancy from 50% to 70% shifts the yield by 1.35 points (3.38% to 4.73% in Case B at ¥8,400). In other words, however precisely the ADR ramp assumption is set, opening-year earnings swing more with how occupancy ramps than with the difference between ADR assumptions. The finding this article presents on the ADR side — that the discount is smaller and shorter than expected — does not displace the uncertainty on the occupancy side.

It should also be kept in mind that this analysis measures the price term only. The ramp-up of occupancy after opening cannot be tracked for the 2024-2025 opening cohort, because our estimated occupancy data only accumulates from 2026 onward. The DCF convention of 60% in year one rising to 80% by year three is left in place, and it remains likely that most of the earnings risk in the opening year sits on the occupancy side. The implication of this article is that the discount on the ADR side is smaller and shorter than expected — not that the opening year can be viewed optimistically.

Summary

Four points emerge from the observed data on 182 hotels with 50 or more rooms opened in 2024-2025, covering 4,138 hotel-months.

Writing “-15% in year one, stabilizing within three years” into the ADR ramp line of a business plan is, against the observed data, either too conservative or simply beside the point. It is excessively harsh for the median property, and it misjudges the months required for recovery at properties aiming well above the market. What belongs there is not a single discount rate but a monthly ramp curve keyed to the target position relative to the market median. The tercile ramps in this analysis (bottom third 0.982 in month 1, middle third 0.971, top third 0.910 rising to steady state over six months) can serve as a starting point.