We tracked the booking progress of ryokan in Kusatsu, Gunma, across the three nights of the Labor Thanksgiving long weekend, Saturday Nov 21 to Monday Nov 23 (national holiday), from 45 days before arrival up to the latest observation (Nov 6). The first night, Saturday Nov 21, had already reached an estimated occupancy (Est. OCC, based on OTA-listed inventory) of 96.4% at 45 days out, and by 15 days out it had climbed to 99.6% with just 5 rooms left (N=52 properties, 1,391 rooms). The second night, Sunday Nov 22, moved from 90.2% at 45 days out to 98.8% at 16 days out, with 17 rooms left. The final night, Monday Nov 23 (holiday), however, stood at 67.5% at 45 days out, 76.4% at 30 days out and 80.8% at 17 days out: about 19 points below the other nights of the same long weekend. 269 rooms remain unsold (N=53 properties, 1,401 rooms), more than ten times the 22 rooms left across the first two nights combined. This article examines where this final night stands against previous long-weekend final nights (Oct 12, Nov 3 and Sep 23), checks whether the weakness is specific to Kusatsu ryokan by comparing it with Gunma Prefecture as a whole and with all property types in Kusatsu, and sets out the options for the remaining two weeks.

Scope: ryokan in Kusatsu, Gunma Prefecture, N=45-54 properties, 1,319-1,409 rooms (varies by stay date; based on listings on major booking sites). This article does not cover price levels; occupancy is shown only as an estimate based on OTA-listed inventory (Est. OCC). Definitions are at the end of the article. Data as of Nov 7, 2026 (observations up to Nov 6).

Booking Curves for the Three Nights — The First Two Are Effectively Full, Only the Final Night Paces Like a Weekday

First, we overlay the curves for all three nights on the same axes. Saturday Nov 21 moved from 96.4% at 45 days out to 98.6% at 30 days out and 99.6% at 15 days out, a “pinned to the ceiling” pattern that steadily absorbs the little inventory that remains. Sunday Nov 22 started slightly lower at 90.2% at 45 days out, then accelerated to 95.0% at 35 days out and 96.5% at 30 days out, drawing level with the first night at 98.8% at 16 days out. Monday Nov 23 (holiday), by contrast, added 8.9 points from 67.5% at 45 days out to 76.4% at 30 days out, but picked up only 4.4 points in the not-quite-two weeks from 30 to 17 days out, stalling at 80.8%.

What stands out is that the final night’s curve has almost the same shape as the weekdays around the long weekend. Friday Nov 20, the night before, moved 72.4% at 45 days out → 80.0% at 30 days out → 85.2% at 14 days out, and Tuesday Nov 24, the night after, moved 64.2% → 73.0% → 79.5% at 18 days out; Monday Nov 23 at 67.5% → 76.4% → 80.8% falls between those two. In other words, judging by booking pace, the night of Nov 23 looks less like “part of a long weekend” and more like “the eve of a post-holiday weekday.” The gap to the first night was 28.9 points at 45 days out, and comparing both at 17 days out it is still 18.8 points (99.6% vs 80.8%), with no sign of narrowing over time. The same structure, in which the last day of a long weekend stays in the weekday band despite being a holiday, was observed as a 16.3pt gap to the first day in Yufuin Ryokan Oct Long Weekend: Final Day Trails Day 1 by 16.3pt.

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan; Est. OCC based on OTA-listed inventory; observations from 45 days before arrival to the latest date)

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan. Latest = observed Nov 6, 2026. Sold-out property share = estimated share of properties with no listed inventory found on OTAs and other channels)

The table makes the difference in sold-out property share even clearer. For Saturday Nov 21, 94.2% of the 52 properties already have no listed inventory, and for Sunday Nov 22 the figure is 84.6%, whereas for Monday Nov 23 (holiday) only 26.4% of properties show no inventory. Put the other way, about three quarters of properties still have rooms to sell for the final night, and the market as a whole is entering a phase of “competing for the same 269 rooms’ worth of demand.” For reference, the following Saturday, Nov 28, moved 85.9% at 45 days out → 92.9% at 30 days out → 96.7% at 22 days out (N=52 properties, 1,388 rooms), and Saturday Dec 5 moved 81.4% at 45 days out → 90.1% at 30 days out (N=54 properties, 1,409 rooms): for Kusatsu ryokan, the standard pattern is that even an ordinary Saturday passes 90% by 30 days out.

How the Last Three Long-Weekend Final Nights Landed — From Around 80% at 17 Days Out, Adding 9-14 Points

So is 80.8% at 17 days out late for a final night, or is it the usual pace? Overlaying the long-weekend final nights (the night of the last day of a long weekend) that Kusatsu ryokan went through this autumn on the same axes gives a fairly clear answer. Monday Oct 12 (Sports Day) moved 66.5% at 45 days out, 73.7% at 30 days out and 82.7% at 17 days out before landing at 92.0% just before arrival; Tuesday Nov 3 (Culture Day) went from 64.8% → 72.6% → 79.2% to 93.2% just before arrival; and Wednesday Sep 23 (Autumnal Equinox Day), the last day of Silver Week, went from 60.9% → 70.2% → 83.2% to 92.9%. All three nights were in the 79-83% band at 17 days out, then added 9.3-14.0 points by the day before arrival and landed together at 92-93%.

Against this benchmark, this year’s Monday Nov 23 (holiday) at 80.8% at 17 days out sits right between Oct 12 (82.7%) and Nov 3 (79.2%), so its pace is not particularly slow. The structure in which only the final night ends low within a long weekend, however, does not change. Against Saturday Oct 10 at 99.6%, Sunday Oct 11 at 99.4% and Tuesday Sep 22 (holiday) at 99.7%, the final nights landed at 92.0% and 92.9%, about 7 points lower. Note that Nov 3 was preceded by Monday Nov 2, a regular weekday in between, so strictly speaking it was not a long weekend; we included it because it shares the same conditions of a holiday night followed by a weekday. Its 93.2% was 6.1 points below Sunday Nov 1 (99.3%), and 2.9 points below the weekday Monday Nov 2 (96.1%). The sold-out property share just before arrival was also 49.0% for Oct 12, 59.6% for Nov 3 and 53.3% for Sep 23, far below the 87.8-93.8% of the 99%-level nights above (73.1% for Monday Nov 2). Given that the October monthly average for Kusatsu ryokan was 92.8% (daily N=43-50 properties), a long-weekend final night can be described as a night that settles at the level of “a perfectly ordinary day.” In Toyooka Ryokan Nov 21-23 Long Weekend: Day 1 95.7%, Final Night 85.5%, which tracked the same Labor Thanksgiving long weekend for ryokan in Toyooka, home to Kinosaki Onsen, the same picture emerged: the first night is essentially settled, and the nights that can still be moved are the night before and the final night.

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan; Est. OCC based on OTA-listed inventory; observations from 45 days before arrival to the latest date. The past three nights are final through the day before arrival)

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan. The past three nights are final as observed 1 day before arrival; Nov 23 is the in-progress value observed Nov 6, 2026)

Applying the past three pickups directly to this year’s 80.8% puts the landing in a range of 90.1-94.8%. Of the 269 remaining rooms, that works out to 130-196 rooms sold by arrival. Looking only at the 92-93% landing level, this year’s pace looks on track, but if pickup is only on par with Oct 12 or Sep 23, the night will stop just above 90%. These are reference values that apply past pickup, not a forecast.

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan N=53 properties, 1,401 rooms; calculated from 80.8% and 269 rooms left as observed Nov 6, 2026. Not a forecast)

Compared With All of Gunma and All Kusatsu Properties — The Final-Night Weakness Is Region-Wide, and Kusatsu Ryokan Are Actually Ahead

To check whether the slow final-night pace is specific to Kusatsu ryokan, we looked at the same three nights for all properties in Kusatsu (including hotels and other non-ryokan types) and for ryokan across Gunma Prefecture. As of Nov 6, Saturday Nov 21 stood at 99.6% for Kusatsu ryokan versus 98.1% for all Kusatsu properties (N=85 properties, 2,245 rooms) and 96.1% for Gunma ryokan (N=180 properties, 5,630 rooms); Sunday Nov 22 at 98.8%, 96.3% and 93.1%; and Monday Nov 23 (holiday) at 80.8%, 75.1% (560 rooms left) and 76.2% (N=180 properties, 5,624 rooms). Kusatsu ryokan were the highest on all three nights, and on the final night they were 5.7 points above all Kusatsu properties and 4.6 points above Gunma ryokan.

At the same time, the structure in which the final night sinks well below the first two is common to all three aggregates: the gap between Saturday Nov 21 and Monday Nov 23 (holiday) is 19.9 points for Gunma ryokan and 23.0 points for all Kusatsu properties, wider than the 18.8 points for Kusatsu ryokan. The starting point at 45 days out was also higher for Kusatsu ryokan at 67.5%, versus 64.0% for Gunma ryokan and 60.2% for all Kusatsu properties, so in terms of final-night sales, Kusatsu ryokan can be read as outperforming both comparison aggregates. The 269 remaining rooms therefore seem to stem not from “Kusatsu not selling” but from “thin demand for the long-weekend final night as a date in itself,” and the logical first step is to focus on how to create demand (this article does not cover price levels, so price-related factors have not been examined). For the category-specific pattern in which Gunma ryokan reach high levels earlier than other property types and Est. OCC and sold-out property share move in the same direction, see Gunma Booking Curves by Type, which used the October long weekend as its case.

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Est. OCC based on OTA-listed inventory, observed Nov 6, 2026. Kusatsu ryokan N=52-53 properties, all Kusatsu properties N=85-86, Gunma ryokan N=180)

Finally, let us look at how the final night’s inventory is distributed across properties. For Monday Nov 23 (holiday), we tallied the remaining allocation share (listed inventory ÷ total rooms) for the 61 Kusatsu ryokan (1,197 rooms) whose listed allocation has at some point reached 30% or more of total rooms. 12 properties (19.7%) already have zero remaining allocation, 16 (26.2%) have 20% or less, 18 (29.5%) have more than 20% up to 50%, and 15 (24.6%) have more than 50%, with a median of 28.6%. In other words, even among Kusatsu ryokan, a group that has already sold out the final night (about one fifth) coexists with a group that still holds more than half (about one quarter), and the right moves for the remaining two weeks depend on which group your property is in.

Source: Compiled by HotelBank Editorial Team from MetroEngines Research (Kusatsu ryokan N=61 properties, 1,197 rooms, as of Nov 7, 2026. The property selection differs from the booking-curve aggregate)

For Revenue Managers at Kusatsu Ryokan — Implications and Action Plan

Source: Prepared by HotelBank Editorial Team (based on figures in this article; results are not guaranteed)

Summary — Three Yardsticks for Measuring a Long-Weekend Final Night

First, measure the final night with a “weekday eve” yardstick, not a “long weekend” one. Kusatsu ryokan’s Monday Nov 23 (holiday) stands at 80.8% at 17 days out, between Friday Nov 20 at 85.2% (14 days out) and Tuesday Nov 24 at 79.5% (18 days out) from the same Nov 6 observation, and it moves at a different level from the 98-99% of the first two nights. Second, the landing benchmark is 92-93%, with 9-14 points of pickup from 17 days out. The fact that all three long-weekend final nights this autumn fell within this range makes it usable as a baseline for evaluating your own property’s pace this time. Third, the final-night weakness is region-wide, and Kusatsu ryokan are ahead. Kusatsu ryokan at 80.8% are above Gunma ryokan (76.2%) and all Kusatsu properties (75.1%), so it is natural to read the 269 remaining rooms as stemming from the date’s demand structure rather than a weakness specific to Kusatsu (price-related verification is outside the scope of this article). We plan to revisit where the night of Nov 23 actually landed once the stay date has passed.

About the Data

Source: Compiled by HotelBank Editorial Team from MetroEngines Research