The “Trip Base Michi-no-Eki Project” that Sekisui House and Marriott International launched in November 2018 was summarized as a network of limited-service hotels at “29 properties in 14 prefectures nationwide” in a notice issued by Sekisui House Hotel Management on March 12, 2026. Within the scope tracked by MetroEngines Research, the Fairfield by Marriott hotels located next to michi-no-eki (government-registered roadside stations), excluding the two urban properties in Sapporo and Osaka Namba, number 29 properties with 2,336 rooms, an average of 80.6 rooms per property, and in September 2026 the estimated ADR was up year on year at 27 of 28 properties. This article lines up the 29 locations by area rate and by room stock within a 5km radius, works backward from construction cost to find the rate at which an 80-room model with no food and beverage operation can stand on its own, and offers a yardstick that municipal attraction officers, regional banks and developers can use to read where the next openings could go.

The conclusion first. Take the published 2018 plan figures (Phase 1: 15 locations and 1,000 rooms, generating about ¥20 billion in construction demand) as a construction cost of ¥20 million per room, and assume 80 rooms with no F&B, estimated year-round occupancy of 75%, an NOI margin of 38% and a target yield of 7%: the required ADR at the same cost level as the existing properties is ¥15,100. Building the same structure at today’s construction cost (steel frame at ¥2.405 million per tsubo in 2025) raises the required ADR to ¥21,300. The 12-month median estimated ADR of the 29 properties is ¥12,800, and it is climbing toward the first line, with a median of +12.0% year on year in July–September 2026 (+18.5% for September alone). Our reading is that room for the next openings narrows to tourist hubs where area rates approach the ¥21,300 band, or to locations where municipal land provision and industrialized construction can bring down the investment side.

The Business Framework — What Public Documents Confirm

First, let us sort out the ownership and operating relationships within the scope of the primary sources. At the time of the November 28, 2018 announcement, it was reported that Sekisui House would handle hotel design and construction, and that the business would take a form in which “a wholly owned subsidiary of the company acts as the operating entity and entrusts actual operations to Marriott” (Travel Voice, November 28, 2018). In Phase 2 (14 locations, 1,184 rooms in 8 prefectures), announced by Sekisui House on October 29, 2021, a special purpose company, “New Tourism Trip Base No. 2 GK”, was formed as the business entity; Sekisui House entered into a tokumei kumiai (silent partnership) agreement with the SPC, the SPC signed a building lease with a Sekisui House subsidiary that runs the hotels, and the land is leased from the municipality that established the michi-no-eki (Kyodo News PR Wire). The March 12, 2026 notice for the “Michi-no-Eki Day” campaign was issued by Sekisui House Hotel Management Co., Ltd., and states that it “operates 29 limited-service hotels nationwide”. The form of the contract with Marriott (franchise or management contract) cannot be determined from these public documents, so this article does not estimate it or go into it.

What matters here is that this is a three-party structure, unusual for a regional limited-service hotel. The SPC owns the building, the municipality owns the land, and an operating company runs the hotel. For the municipality, it converts existing infrastructure, the land next to a michi-no-eki, into a lodging function; for the business, land acquisition cost drops out of the investment. Both feed directly into the required ADR calculation discussed below. For reference, the number of registered michi-no-eki reached 1,234 nationwide with the 65th registration on September 4, 2026 (Ministry of Land, Infrastructure, Transport and Tourism), so the 29 properties sit next to 2.4% of all stations.

The approach to food is also at the core of the business model. Since the 2018 announcement, the design has consistently been that the hotel has no restaurant, and guests eat and shop for souvenirs at the michi-no-eki or nearby shops. An August 2026 feature (@DIME) reported that all properties share the same brand design, locally hired staff guide guests on the area, dog-friendly rooms were introduced at 7 properties from 2024, and occupancy “seems to be rising every year”. Narrowing the function to lodging keeps the building floor area and staffing small, as discussed below, while giving up F&B revenue such as breakfast. With this structure as the premise, the sections below follow the conditions of rate and location in numbers.

The 29 Properties at a Glance — 2,336 Rooms in Four Years, 80.6 Rooms on Average

Within the scope tracked by MetroEngines Research, the 29 michi-no-eki properties began with openings in October 2020 of three properties in Gifu and one each in Mie, Tochigi and Kyoto, and built up at an almost even pace: 8 properties in 2020, 6 in 2021, 8 in 2022 and 7 in 2023. Room counts range from 50 (Kyoto Minamiyamashiro) to 102 (Hokkaido Eniwa), with a median of 85. The facility configuration is also common: all rooms are Western-style at around 25 sqm (21 sqm at the three Hokkaido properties), each property has a lounge, and none has a large public bath, hot spring or restaurant. By prefecture, Hyogo has 4, Gifu 4, Hokkaido 3, Tochigi 3 and Kyoto 3, so the network is concentrated in Kansai and Chubu, while Kyushu has one each in Saga, Fukuoka, Kumamoto and Kagoshima. The urban Sapporo property excluded from this analysis is a case where the brand’s first hotel in Japan was launched through a rebrand of an existing hotel.

Where Rates Stand — 12-Month Median Estimated ADR of ¥12,800, 27 of 28 Properties Up Year on Year

Across the 28 properties for which an estimated settled ADR can be calculated, the 12-month average ranges about 2.5-fold, from ¥8,800 (Wakayama Kushimoto) to ¥21,700 (Kumamoto Aso), with a median of ¥12,800. The direction is more telling than the level: comparing September 2026 estimated ADR with the same month a year earlier, 27 of 28 properties were up, with a median of +18.5%. The largest gains were at Hyogo Awajishima Higashiura (+30.2%), Gifu Mino (+28.0%), Hyogo Awajishima Fukura (+26.8%) and Mie Kumano Kodo Mihama (+26.6%); the two Awaji Island properties outgrew their areas, where municipal ADR moved +2.9% and +16.8% over the same period. This reads as a phase in which properties three to six years after opening are lifting their rates. However, September 2026 includes a five-day holiday run on the calendar (September 19–23), so the single-month gain may be mixed with calendar effects. Comparing the three months of July–September combined, all 28 properties were positive with a median of +12.0% (monthly medians of +8.8% in July and +12.7% in August): the direction is the same, but this range is less affected by the calendar. The background to rates rising on Awaji Island even as lodging supply grew 1.4-fold is analyzed in detail in Awaji Island: Listings Up 38%, Estimated ADR Still Rose 7.7-11.0%.

Overlaying the months also makes seasonality clear. In 2026, the peak is in summer to early autumn, with August at ¥14,000 and September at ¥13,900, and the trough is June at ¥9,900. Compared with the same months of 2025 (August ¥12,600, September ¥11,700), the peak has risen a notch; April 2026 was low at ¥11,000, and January–March ran between ¥11,900 and ¥13,200. October–December 2026, estimated from listed price levels at the time of research, comes in at ¥12,100–¥12,400: October is slightly above the same month a year earlier (¥11,900), while November and December are below their year-earlier levels (¥13,400 and ¥13,200). Because the October–December estimates are on a different basis from closed months, the year-on-year comparison is only a guide to direction. With a guest mix dominated by travelers touring by car, the weekday-weekend gap is not as wide as at urban hotels; instead, demand concentrates in holiday periods and summer.

Three Location Types by Gap to the Area Rate — Leading, Complementing, Parallel

Dividing each property’s estimated ADR (12-month average) by the median of target properties in the same municipality (September 2026, N of 5 or more) splits the locations into three types. We call a ratio of 1.3 or higher “rate-leading”, 0.85 or lower “price-band-complementing”, and anything in between “area-parallel”. There are 6 rate-leading properties: Hyogo Tajima Yabu (1.89x; Yabu City ADR ¥6,300, N=11), Gifu Seiryu Satoyama Park (1.89x), Fukuoka Ukiha (1.89x), Okayama Tsuyama (1.63x), Tochigi Utsunomiya (1.56x) and Kagoshima Tarumizu Sakurajima (1.50x). These are areas where the typical lodging in the municipality is in the ¥6,000–¥8,000 range, and the hotel takes ¥11,000–¥14,000 on the strength of the brand and newly built rooms: locations where the arrival of a foreign brand created a new upper price band in the area.

The price-band-complementing type, on the other hand, has 7 properties: Hyogo Kannabe Highland (0.66x), Hyogo Awajishima Fukura (0.67x; Minamiawaji City ADR ¥20,800, N=18), Saga Ureshino Onsen (0.72x), Kyoto Amanohashidate (0.74x), Tochigi Nikko (0.77x; Nikko City ADR ¥17,600, N=121), Gifu Takayama Shirakawa-go (0.82x) and Wakayama Kumano Kodo Susami (0.47x, N=5). All are tourist destinations where hot-spring ryokan and resorts form the center of the price range; where ryokan with two meals push up the median, these hotels supply room-only Western-style rooms at ¥9,000–¥14,000. For travelers who want to tour by car but cannot reach the ryokan price band, or for independent inbound travelers who want to extend their stay, the reading is that a new price-band option has been added to an existing lodging destination. The area-parallel type covers 6 properties, Kumamoto Aso (1.12x, though at ¥21,700 its estimated ADR is the highest of the 29), Okayama Hiruzen Highland, Gifu Gujo, Wakayama Kushimoto, Nara Tenri Yamanobenomichi and Hyogo Awajishima Higashiura, which sit at roughly the same level as typical lodging in their areas.

The remaining 10 properties have fewer than 5 target properties in their municipality and were excluded from the ratio calculation. Mie Kumano Kodo Mihama, Hokkaido Minamifurano and Kyoto Minamiyamashiro are the only target properties in their municipalities, and no ratio is given for Mie Okuise Odai because the property’s own estimated ADR is outside the estimation scope. This connects to the room stock discussion in the next section. In other words, this business model includes both “locations that slot a new price band into an existing lodging destination” and “locations that put a hotel for the first time at a michi-no-eki with almost no lodging around it”, and the conditions under which rates hold are harder to read for the latter.

Room Stock and Surrounding Demand Within 5km — 12 Sole-Supplier Properties

Looking at each property’s share of operating rooms within a 5km radius (properties with 10 or more rooms and a sales price confirmed in the most recent three months), 12 properties are “sole supplier” locations at 50% or more, 8 are “few-player coexistence” locations at 20–50%, and 9 are “within a cluster” at under 20%. Five properties, Gifu Gujo (87 rooms), Hokkaido Minamifurano (78), Mie Kumano Kodo Mihama (63), Mie Okuise Odai (72) and Kyoto Minamiyamashiro (50), are the only operating property within 5km, for a 100% share. Wakayama Kumano Kodo Susami holds 90 of 107 rooms in 2 properties within 5km (84%) and Gifu Takayama Shirakawa-go 64 of 87 rooms in 2 properties (74%), effectively close to sole supply. At these locations the very concept of an “area rate” is set by the property’s own price, so the yardstick for comparison needs to be found not in the municipality but in lodging destinations on the same touring route (central Takayama or Furano City).

In contrast, the 9 properties within clusters, such as Tochigi Nikko (39 properties and 1,121 rooms within 5km, an 8% share), Saga Ureshino Onsen (29 properties, 1,032 rooms, 8%), Fukuoka Ukiha (17 properties, 621 rooms, 8%) and Kumamoto Aso (14 properties, 682 rooms, 14%), are located in existing lodging destinations. Here the price-band gap to surrounding ryokan and resorts is the product. Supplementing surrounding business demand with the number of establishments and employees within 2km (Economic Census), Nara Tenri Yamanobenomichi has 711 establishments and 9,902 employees, Hokkaido Eniwa 851 and 8,863, and Saga Ureshino Onsen 649 and 6,191: points where business and exchange demand can be expected on top of tourism. Conversely, Hyogo Kannabe Highland (103 and 492) and Gifu Takayama Shirakawa-go (40 and 212) are points specialized in tourism demand. Business demand indicators are not shown for the five points of Gujo, Ukiha, Naganuma, Hiruzen and Minamiyamashiro because they fall outside the coverage of the small-area statistics.

As background on supply and demand, the estimated occupancy for the business hotel segment at the prefecture level (based on inventory listed on major booking sites) in September 2026 ranged from 63.8% in Kagoshima (72 target properties, 5,868 rooms) to 92.8% in Hokkaido (222 target properties, 25,913 rooms). The median for the 14 prefectures is about 86%, showing that limited-service supply broadly absorbed September demand. These estimates are based on how inventory sold on major booking sites is being taken up and differ from the actual occupancy of the whole property. The reason the estimate below assumes 75% occupancy is that, on a year-round basis, the off-season needs to be discounted from this September level.

Estimating the 80-Room Model — Required ADR of ¥15,100 and ¥21,300 at Two Construction Cost Levels

So at what rate does a hotel of this format work? Let us work backward. The premise is 80 rooms with no F&B, close to the 29-property average (80.6 rooms), and total floor area set at 40 sqm per room, 3,200 sqm in total (about 968 tsubo), based on actual examples (Ukiha 41 sqm per room, Awajishima Higashiura 37 sqm per room). Construction cost is shown at two levels. Level 1 is based on public information: the figure from the 2018 announcement, in which Sekisui House expected “about ¥20 billion in construction demand” for Phase 1’s 15 locations and 1,000 rooms, is read as ¥20 million per room, and ¥2.5 million per room is added for FF&E (furniture, fixtures and equipment) and pre-opening costs, for a total investment of ¥1.8 billion (¥22.5 million per room). Level 2 is the case of building the same structure today: applying the 2025 steel-frame cost of ¥2.405 million per tsubo based on the MLIT Building Construction Starts Statistics (the all-structure average was ¥1.95 million in 2024 and ¥1.38 million in 2022, up 41% in two years) gives a construction cost of ¥2.33 billion and a total investment of ¥2.53 billion (about ¥31.6 million per room). Land is assumed to be leased from the municipality, so no acquisition cost is included and ground rent is included in operating costs. How much the cost per room differs between a new build and a brand conversion of an existing hotel is compared with actual Sapporo data in Hotel Brand Conversion Costs 1/7 of New Build: 15 Sapporo Hotels.

On the revenue side, the NOI margin (net operating income after ground rent, fixed asset tax, insurance and FF&E reserve) relative to revenue is set at 38%. A GOP (gross operating profit) margin of around 30–40% is generally regarded as one benchmark across the industry, with business hotels tending to run relatively higher (Relo Hotel Solutions), and the results of 91 operated properties disclosed by Invincible Investment Corporation were 38.9% (fiscal period ended December 2024). Given that this is a limited-service format with no F&B division, we set the GOP margin at 45% (this article’s assumption) and deducted 7 points for the small 80-room scale, the regional location and the ground rent burden on leased land. This 45% is set above the benchmarks above, and if the NOI margin comes in lower, the required ADR rises as shown in the sensitivity table below. Estimated year-round occupancy is 75%, and the target yield (NOI divided by total investment) is 7% as the base for a standalone regional asset, with sensitivity at 6% and 8%.

Applying this sensitivity table to the results of the 29 properties makes their current position clear. Only 4 properties exceed ¥15,100, the Level 1 (same cost level as the existing properties) line at a 7% yield and 75% occupancy, on a 12-month average: Kumamoto Aso (¥21,700), Hokkaido Minamifurano (¥17,000), Hyogo Awajishima Higashiura (¥16,500) and Hokkaido Eniwa (¥16,300). Lowering the bar to ¥13,000, the 6% yield line, brings in 12 properties, and the remaining 16 are also building up their rates toward this line, at a median of +11.0% year on year in July–September 2026. If occupancy can be raised to 80%, the Level 1 7% line falls to ¥14,200 and the 6% line to ¥12,200, and the number of properties exceeding the latter on a 12-month average widens to 16. The biggest structural strength of the existing 29 properties is that they were built before construction costs rose, and they are approaching the viability line on both occupancy and rate.

Another factor that sways the conclusion heavily is the NOI margin. Lowering 38% to 30% raises the Level 1 required ADR to ¥19,200, which only Kumamoto Aso reaches on a 12-month average. Conversely, at 42% it is ¥13,700, which 7 properties reach. Applying the 29-property 12-month median of ¥12,800 at 75% estimated year-round occupancy and a 38% NOI margin, the implied yield is about 5.9% at Level 1 and about 4.2% at Level 2. At the cost level before construction prices rose, the yield is close to the target; at today’s cost, there is a gap.

Level 2’s ¥21,300, meanwhile, is the condition for a new opening that builds the same structure from today. Of the 29 properties, only Kumamoto Aso exceeds it on a 12-month average, and none of the 29 locations has a municipal ADR above ¥21,300 (the highest is Minamiawaji City at ¥20,800), so even building the same format at the same cost, it is hard to make it work on rate alone. That is exactly why room for the next openings shifts from “areas with high rates” to “areas with conditions that can lower the investment”. The next section deals with this point.

Viability Conditions and Room for the Next Openings — Three Levers Municipalities and Regional Banks Can Move

Summed up in one sentence, the viability conditions readable from the results of the 29 properties are: “a michi-no-eki where typical area lodging is ¥17,000 or more, or where municipal land and infrastructure support can bring the investment close to Level 1, and where touring routes or business demand can deliver year-round occupancy of 75% or more”. Looking at the existing distribution, overall review scores are uniformly high at 4.40–4.70 across rate-leading, complementing and sole-supplier types (29-property weighted average of 4.52 for the last 24 months, N=14,288 reviews, HotelBank Editorial Team research), so the acceptance of the product has been confirmed even where the character of the location differs. Put differently, location choice has reached the stage where it is decided not by brand acceptance but by the arithmetic of cost and rate.

The implications for municipal attraction officers are clear. At the time of the 2018 announcement, Sekisui House had entered discussions on location conditions with 10 prefectures, and the 2021 Phase 2 announcement said it was advancing business partnerships with municipalities in 26 prefectures and aiming for about 3,000 rooms in 26 prefectures in 2025, but as of 2026 the rollout remains at 14 prefectures. For the remaining prefectures to become candidates, in addition to the location condition of being next to a michi-no-eki, the municipality needs to be able to present three points in numbers: (1) the level of surrounding lodging rates, (2) the terms of land and infrastructure provision, and (3) the position on touring routes. For regional financial institutions, the structure of the SPC’s building lease and the municipality’s land lease is material for judging how rent stability is secured; for developers, the competitive condition is whether they can control the cost of a compact 80-room, 3,200 sqm structure through industrialized construction.

Caveats — Estimate Assumptions and How to Read Them

The estimated ADR in this article is a settled rate estimated from published prices on major booking sites, not each property’s accounting results. Checked against REIT-disclosed results, the median error is about 7%, so individual property figures should be read with a range of about ±¥1,000. Municipal ADR is the median of target properties in that municipality; locations with N below 5 are treated as reference values and excluded from the location type calculation. Level 1 construction cost is this article’s assumption, converting the total at the 2018 plan announcement into a per-room figure, and differs from actual construction costs and each property’s contract terms. The 38% NOI margin, 75% occupancy and 7% yield are also this article’s assumptions; as this is a simplified estimate, actual investment decisions require a detailed feasibility study.

In addition, room stock within 5km is limited to properties with 10 or more rooms, so small properties such as minshuku and guesthouses are not included. Business demand based on the Economic Census aggregates the 2021 Economic Census for Business Activity at the small-area level; it is a static indicator of residence and employment and does not reflect tourist flows. Please also note that estimated occupancy is based on how inventory sold on major booking sites is being taken up, and differs from the actual occupancy of the whole property.

Frequently Asked Questions

Q. How many Fairfield by Marriott hotels are located next to michi-no-eki?

According to Sekisui House Hotel Management’s notice of March 12, 2026, there are 29 properties in 14 prefectures nationwide. Within the scope tracked by MetroEngines Research as well, excluding the two urban properties in Sapporo and Osaka Namba, there are 29 michi-no-eki properties with 2,336 rooms, an average of 80.6 rooms per property.

Q. At what rate does an 80-room hotel with no F&B work?

In this article’s estimate, at the cost level based on the 2018 plan figures (total investment ¥1.8 billion), with 75% estimated year-round occupancy and a 7% yield, the required ADR is ¥15,100; at the 2025 construction cost level (total investment ¥2.53 billion), it is ¥21,300. Both are simplified estimates and assume the land is leased from the municipality.

Q. What are room rates at the 29 properties?

The 12-month average (October 2025 to September 2026) across the 28 properties for which an estimated settled ADR can be calculated ranges from ¥8,800 to ¥21,700, with a median of ¥12,800. In September 2026, 27 of 28 properties were up year on year, with a median of +18.5%. Please note these are estimates, not each property’s actual results.

Summary

The 29 Fairfield by Marriott properties next to michi-no-eki built up 2,336 rooms in 14 prefectures over the four years of 2020–2023, and they are in a phase of rising rates: in September 2026, estimated ADR was up year on year at 27 of 28 properties, with a median of +18.5% (including a five-day holiday run; a median of +12.0% over the three months of July–September). Compared with area rates, there are 6 “rate-leading” properties taking ¥11,000–¥14,000 in municipalities in the ¥6,000–¥8,000 range, 7 “price-band-complementing” properties filling a room-only Western-style price band in ryokan-led tourist destinations, and 12 “sole-supplier” properties providing the only or majority supply within 5km. A defining feature of this model is that one brand makes the same product work in three kinds of location.

In the estimate for an 80-room, no-F&B hotel, at the same cost level as the existing properties the required ADR is ¥15,100 (7% yield, 75% occupancy); 4 properties exceed it on a 12-month average, and 12 reach the 6% line of ¥13,000. The ¥21,300 required for a new build at today’s construction cost exceeds the municipal ADR of every one of the 29 locations, so rather than searching for areas with high rates, the next openings are likely to head to michi-no-eki where municipal land and infrastructure provision lowers the investment, and where overlapping touring routes and business demand secure occupancy. With hotels next to only 29 of the 1,234 michi-no-eki, the pool of locations still shows ample room for foreign brands to expand into regional Japan, and the stations that actually open will be those where the conditions above come together.