According to the Monthly Survey on Service Industries (preliminary figures for August 2026) released on October 23 by the Statistics Bureau of Japan (Ministry of Internal Affairs and Communications), monthly revenue in the accommodation industry was ¥718.85 billion, up 2.3% year on year. For the same month of August, the Japan Tourism Agency’s Overnight Travel Statistics Survey (first preliminary figures) shows total guest-nights of 62.68 million, down 6.7% year on year. In other words, revenue rose even as the number of guest-nights fell by nearly 7%: revenue divided by guest-nights — revenue per guest-night — came to ¥11,469, up 9.7% from a year earlier. This article overlays these two government statistics across the 20 months from January 2025 to August 2026, explains why they point in a different direction from our estimated settled ADR as a matter of differing populations, and sets out how to read the autumn figures.

What This Statistic Is — What “Accommodation Industry Revenue” Counts

First, let us check the background of the revenue-side statistic. The Monthly Survey on Service Industries is a monthly fundamental statistical survey conducted by the Statistics Bureau of Japan since the January 2025 reference month. It was created by merging the former Survey of Service Industries (Ministry of Internal Affairs and Communications, a general statistical survey) and the Current Survey of Selected Service Industries (Ministry of Economy, Trade and Industry), and it serves as source data for the quarterly preliminary GDP estimates and similar figures. It covers about 13,000 enterprises and about 25,000 establishments: companies with capital of ¥100 million or more are surveyed in full, while other establishments are sampled by stratified sampling by industry and employee size. Its main survey items are just two — sales (revenue) and number of employees — but it is effectively the only statistic that tracks nationwide service-industry revenue on a monthly basis.

The “75 Accommodations” series used in this article is contained in the survey’s time-series table <Sales (revenue) by industry of business activity (medium group) (monthly)>. A point to note here is the “by industry of business activity” method of aggregation. In the enterprise survey, when a single enterprise operates multiple businesses, it reports sales separately for each business activity, and the statistic classifies “industry” by business activity. In other words, the hotel-business revenue of a railway company or real estate company goes into the accommodation industry column as accommodation activity, not under the enterprise’s main industry. In the establishment survey, by contrast, establishments are classified by their main type of business activity. Because establishment-based tables count differently, this article uses the business-activity-based time series throughout.

There are three further caveats. First is the note in the table title that “values through December 2024 are adjusted values.” Figures for 2024 and earlier are results of the former Survey of Service Industries adjusted for changes arising from updated population information, sample rotation, etc., and are constructed to connect directly with the figures of the new survey from January 2025 onward. YoY changes for each month of 2025 are therefore ratios against the adjusted 2024 values. Second is the difference between preliminary (p) and final figures. In principle, preliminary figures are released in late in the second month after the reference month, and final figures in late in the fifth month after, at which point the numbers are replaced. The August figures here are preliminary, and the June and July 2026 figures are also still at the preliminary stage. In fact, the May figure was revised from ¥664.6 billion in the September preliminary release to ¥682.7 billion in the final release on October 23. Third, revenue here is not limited to room charges. It is the sales (revenue) of enterprises and establishments engaged in the accommodation business, and includes income from ancillary departments such as restaurants, banquets, weddings, retail and bathing facilities. It must not be read as “room revenue.”

Monthly Revenue YoY — 20 Consecutive Months of Growth from January 2025 to August 2026

Now let us look at the revenue trend. Accommodation industry revenue exceeded the same month of the prior year for 20 straight months from January 2025 through August 2026. Full-year 2025 revenue was ¥7.23275 trillion, up 6.4% from 2024 on an adjusted basis (¥6.79825 trillion). By month, August’s ¥702.35 billion was the largest, followed by November at ¥680.33 billion, October at ¥669.64 billion and December at ¥640.81 billion. This seasonal pattern — the autumn travel season and the year-end period being the largest after the summer peak — broadly matches the guest-count statistics.

Monthly revenue in 2026 moved as follows: January ¥554.01 billion (+4.6%), February ¥548.28 billion (+7.2%), March ¥670.46 billion (+9.4%), April ¥614.07 billion (+1.3%), May ¥682.71 billion (+11.3%), June ¥528.98 billion (+1.9%), July ¥618.45 billion (+8.0%) and August ¥718.85 billion (+2.3%). August set a new single-month record since the survey began, but its growth rate was smaller than May’s or July’s. The January–August cumulative total was ¥4.9358 trillion, up 5.8% year on year. However, the June–August figures are preliminary, and as with May (preliminary ¥664.6 billion → final ¥682.7 billion), revisions exceeding ¥10 billion in the final release remain possible.

Source: Compiled by HotelBank Editorial Team from Statistics Bureau of Japan, “Monthly Survey on Service Industries” (75 Accommodations, business-activity basis; 2024 adjusted values; June–August 2026 preliminary). Chart unit: ¥100 million.

Source: Compiled by HotelBank Editorial Team from Statistics Bureau of Japan, “Monthly Survey on Service Industries” and Japan Tourism Agency, “Overnight Travel Statistics Survey”. Release stage is shown as “revenue / guest-nights”. Revenue per guest-night = revenue ÷ guest-nights (includes food and beverage etc., so it is not a room rate).

Overlaying Volume and Value — Guest-Nights Down 8 Months in a Row, Revenue Up 8 Months in a Row

Overlay the guest-count statistic, and the picture of the 2026 lodging market changes. According to the Japan Tourism Agency’s Overnight Travel Statistics Survey, total guest-nights fell short of the prior year for eight consecutive months, from 45.46 million in January 2026 (−6.5%) to 62.68 million in August (−6.7%). Some months saw steep declines — April −7.2%, June −8.4% — and the January–August cumulative total was 413.40 million guest-nights, down 5.0% year on year. By segment, August guest-nights by Japanese residents were 50.04 million (−5.8%) and by foreign visitors 12.64 million (−10.3%), meaning both domestic and inbound travelers declined. Meanwhile, accommodation industry revenue in the same August rose 2.3%, and 5.8% on a January–August cumulative basis. Two lines run side by side in different directions: shrinking in volume, expanding in value.

The reason is that 2026’s revenue growth is being supported not by volume but by the amount per guest-night. The “price factor” — the portion of revenue growth that cannot be explained by changes in volume, i.e., the YoY change in revenue per guest-night — varied from month to month in 2025, between +0.2% (August) and +12.4% (April). In 2026, however, it has stayed at or near double digits in almost every month: January +11.9%, March +11.8%, May +15.0%, July +13.3% and August +9.7%. The Japan Tourism Agency itself notes that the YoY figures for guest-nights may include the effect of changing the stratification criterion from “number of employees” to “number of guest rooms” starting with the January 2026 survey, so the roughly −5% decline cannot be taken at face value as a contraction in demand. Even so, the pattern of value growing well ahead of volume has been consistent since autumn 2025.

Source: Compiled by HotelBank Editorial Team from Statistics Bureau of Japan, “Monthly Survey on Service Industries” and Japan Tourism Agency, “Overnight Travel Statistics Survey”. June–August 2026 revenue is preliminary; August 2026 guest-nights are first preliminary figures.

The Japan Tourism Agency’s room occupancy rate points the same way. Room occupancy for all properties was 64.4% in August 2026 and 61.0% in July, both below the same months of the prior year in the agency’s statistics. If revenue alone is rising while occupancy falls and guest numbers decline, then either rates are rising, the share of high-spend properties and guest segments is expanding, or non-lodging income is growing. Government statistics on value and volume alone cannot isolate which of these is the main driver.

Revenue per Guest-Night Is ¥11,469 — From a 2025 Annual Average of ¥10,940 to ¥11,940 in 2026

Lining up revenue per guest-night (revenue divided by guest-nights) as a time series reveals a step up in level. The annual average was ¥10,315 in 2024 and ¥10,940 in 2025 (+6.1%); on a monthly basis it reached ¥11,927 in November 2025 and ¥11,654 in December, climbing above ¥11,500 toward year-end. In 2026 it settled in the ¥12,000 range in the first half — January ¥12,186, March ¥12,321, April ¥12,505, May ¥12,575 — and although it dipped seasonally in summer (June ¥11,546, July ¥11,458, August ¥11,469), it held +7.9% to +15.0% year on year in every month from January to August. The January–August cumulative figure was ¥11,940, up 11.3% year on year.

However, this number is not a “room rate.” The numerator includes non-lodging income such as food and beverage, banquets, retail and bathing facilities, and the denominator is guest-nights, so for two guests sharing a room the per-room amount would be close to double. In addition, the guest-count statistic covers all lodging facilities, including simple lodgings and company/organization accommodation facilities, whereas the revenue statistic covers enterprises with capital of ¥100 million or more in full and others by sample. The two populations overlap but are not identical, so revenue per guest-night should be read strictly as “a ratio of two statistics” — an indicator whose direction and magnitude of change matter more than its level.

Source: Compiled by HotelBank Editorial Team from Statistics Bureau of Japan, “Monthly Survey on Service Industries” and Japan Tourism Agency, “Overnight Travel Statistics Survey” (revenue ÷ guest-nights). 2024 revenue figures are adjusted values.

Why Our Estimated Settled ADR Moves Differently — Median vs. Total, Per Room vs. Per Guest-Night, Lodging vs. All Revenue

Is this growth in revenue per guest-night, then, the same thing as the room-rate increase individual hotels experience? Looking at our estimated settled ADR (settled rates estimated from public prices on major booking sites, finalized months only) across 12 prefectures, the median YoY change of the 12 prefectures ranged from +3.1% to +14.6% in January–May 2026 — the same direction as the government statistics — but turned negative in summer: −3.7% in June, −3.1% in July and −4.6% in August. In August, only Okinawa (+2.7%) was above the prior year, while major tourist destinations were uniformly negative: Tokyo −9.2%, Kanagawa −7.5%, Kyoto −8.4% and Hokkaido −5.1%. Osaka has posted −22% to −33% since April, reflecting the rebound from the prior year’s Expo period. In the finalized September data, the median recovered to +4.1%, with 11 of the 12 prefectures back above the prior year. A cross-check of volume and rate YoY across all 47 prefectures as of July (guest-nights on a first-preliminary basis; −4.7% in the second preliminary release) is presented as a four-quadrant volume × rate map in Japan July 2026: Guest-Nights -3.5%, ADR -4.4% Across 47 Prefectures.

Revenue per guest-night rose 9.7% in August, while estimated settled ADR in the same August fell 4.6% at the median. Rather than one of the numbers being wrong, the gap arises because they measure different things. There are four plausible factors. First, our area ADR is the “median” of covered properties — the level of a typical property in the area — so the median barely moves no matter how much prices at high-rate properties rise. Revenue, by contrast, is a “total,” so revenue per guest-night rises simply when the share of high-spend properties or guest segments expands, even if individual properties’ prices do not change. Second, revenue includes food and beverage, banquets and retail. Third is the difference in denominator — per room vs. per guest-night: if the number of guests per room falls, the amount per guest-night rises. Fourth, our coverage is limited to properties listed on major booking sites with confirmed activity (verified property types: business, city, resort, ryokan and capsule), and does not include properties that do not use booking sites or company/organization accommodation facilities.

On the other hand, for January–May 2026 the median estimated settled ADR also rose, by +3% to +15%, so revenue growth in that period can be explained by both “rate increases at typical properties” and “compositional change.” The natural reading of why the two diverged in summer is that, at least among typical properties on major booking sites, rates were adjusted downward, while total market revenue was supported by the high-rate end of the mix and by ancillary income. There is no evidence to establish causation, but this is consistent with what operators feel: “total market revenue is growing, yet our own rates are hard to raise.”

Source: Compiled by MetroEngines Research and HotelBank Editorial Team. YoY change in estimated settled ADR (finalized months, median across all property types); 8,086 properties covered in total across the 12 prefectures in August 2026.

Source: Compiled by MetroEngines Research and HotelBank Editorial Team. ADR is each prefecture’s estimated settled ADR (all property types, median, tax-excluded equivalent, finalized months). Osaka’s YoY is a comparison against the 2025 Expo period. Aichi’s September figure includes the Asian Games period (September 19 – October 4).

How to Read the Autumn — October and November Were the Largest Months After August Last Year; Next Releases on October 30 and November 24

To sum up so far, the 2026 lodging market is in a phase where “volume is below the prior year, value is above it.” To judge whether this pattern continues into autumn, it helps first to understand the prior year’s seasonal structure. In 2025, monthly revenue peaked at ¥702.35 billion in August, followed by November at ¥680.33 billion and October at ¥669.64 billion. October and November combined came to ¥1.34997 trillion, or 18.7% of the annual total, and the four months of September–December totaled ¥2.56547 trillion, 35.5% of the year. In guest-nights as well, October 2025 recorded 59.73 million, second only to August, making autumn the year’s second peak in both volume and value.

Our September estimated settled ADR, already finalized, was above the prior year in 11 of 12 prefectures, with the median at +4.1%, recovering from the summer declines. Tokyo (+1.9%), Kyoto (+1.1%), Kanagawa (+5.6%) and Hokkaido (+4.5%), all below the prior year in summer, turned positive, while Okinawa rose 10.0% and Fukuoka 9.8%. Narrowing to business hotels, September was above the prior year in 44 of 47 prefectures; we separate calendar effects from underlying strength in Business Hotel ADR Sep 2026: 44 Prefectures Up YoY, Median +7.9%. This includes special factors — September’s five-day holiday period and the Asian Games in Aichi — but if rates at typical properties are recovering in autumn, it would push revenue-based market size above the prior year in October–November as well. Whether the roughly −5% decline on the volume side continues is a separate question, however, and there is no guarantee that the price factor will keep outweighing the volume factor.

The upcoming release schedule is as follows. The Japan Tourism Agency is scheduled to release the second preliminary figures for August 2026 and the first preliminary figures for September 2026 of the Overnight Travel Statistics Survey on October 30, 2026, at which point the August figure of 62.68 million guest-nights (−6.7%) used in this article will be revised (we measured revisions from the first to the second preliminary release over 15 months at a median of 1.8% and a maximum of 3.4% for guest-nights in Japan Lodging Stats: Preliminary Figures Move 3.4%, Occupancy 1.0pt). For the Statistics Bureau’s Monthly Survey on Service Industries, the e-Stat release calendar lists the next release at 8:30 a.m. on November 24, 2026; under the release rules (preliminary figures in late in the second month after the reference month, final figures in late in the fifth month after), this corresponds to the preliminary figures for September 2026 and the final figures for June. It will also be necessary to check in which direction the June–August preliminary figures are revised in the final release.

Source: Compiled by HotelBank Editorial Team from Japan Tourism Agency, “Overnight Travel Statistics Survey (August 2026, first preliminary)” and the e-Stat release calendar (Monthly Survey on Service Industries).

Summary

In August 2026, accommodation industry revenue was ¥718.85 billion (+2.3%), guest-nights were 62.68 million (−6.7%), and revenue per guest-night — the ratio of the two — was ¥11,469 (+9.7%). On a January–August cumulative basis, revenue rose 5.8%, guest-nights fell 5.0% and revenue per guest-night rose 11.3%: the 2026 lodging market is expanding as value makes up for the decline in volume. However, this “price factor” does not mean that room rates at typical properties simply rose. The median of our estimated settled ADR fell year on year in June–August, and what lifted total market revenue may have been the high-rate end of the mix and non-lodging income. For autumn, given that October and November were the largest months after August last year and that September’s estimated settled ADR returned above the prior year, there is room for revenue to exceed the prior year. But June–August revenue figures are preliminary and August guest-nights are first preliminary figures, so the numbers should be read on the assumption that they will move with the October 30 and November 24 releases.

Frequently Asked Questions

Q. Where can I find statistics on accommodation industry revenue?
Results of the Statistics Bureau of Japan’s “Monthly Survey on Service Industries” are available on e-Stat. This article used “75 Accommodations” in <Time series> Sales (revenue) by industry of business activity (medium group) (monthly), in millions of yen. Preliminary figures are released late in the second month after the reference month, and final figures late in the fifth month after.

Q. Is revenue ÷ guest-nights the same as the room rate (ADR)?
No. Revenue includes non-lodging income such as food and beverage, banquets and retail, and the denominator is guest-nights rather than rooms. The two statistics also do not cover the same population. This article calls it “revenue per guest-night” and treats it as an indicator whose direction and magnitude of change matter more than its level.

Q. How much do preliminary figures change?
Accommodation industry revenue for May 2026 was revised from ¥664.6 billion in the preliminary release to ¥682.7 billion in the final release. The Japan Tourism Agency’s guest-night figures are also revised from the first preliminary release (effective response rate 36.5%) to the second preliminary release (around 45.1%). Figures at the preliminary stage should be read on the assumption that they will be revised.