Bali Luxury Villas: ADR and Occupancy, Measured

What do ADR and occupancy actually measure for Bali luxury villas?#

Bali's published luxury figures describe hotels, and a principal pricing a villa has to translate them before use. For 2025, Horwath HTL, working with the Bali Hotels Association and C9 Hotelworks, reported island-wide hotel occupancy of 73.2% and an ADR of IDR 2.35 million, equal to USD 150 at an average rate of IDR 16,705 to the dollar (measured, Bali Hotels Association member data). Its top rate band, hotels charging above USD 501 a night, recorded USD 796 ADR at 49.4% occupancy. These are measured numbers. They are hotel numbers.

Two definitions carry the whole piece. Average daily rate (ADR) is room revenue divided by rooms sold, so it is silent on the nights that went unsold. Occupancy is rooms sold divided by rooms available. Multiply the two and the result is RevPAR, revenue per available room, the only one of the three that lets a full house at a low rate be compared with a half-empty house at a high rate. A villa that sells half its nights at a high rate can post the same RevPAR as a hotel that sells most of its nights at a lower one.

Indonesia's statistics office measures something narrower again. BPS, Badan Pusat Statistik, publishes the TPK, Tingkat Penghunian Kamar, which is the room occupancy rate, for star-classified hotels and for non-star hotels. The monthly Bali release carries no ADR. A private villa listed on a booking platform sits in neither group, so BPS cannot tell the allocator how full Bali's luxury villas are.

The sources below are ranked by authority, not by how close they sit to a villa. BPS comes first because it is a state statistical office that counts classified hotels. Horwath HTL comes second because it publishes ADR by rate band and by submarket, from a member sample. Platform-derived estimates come last, because they are the nearest thing to villa data and the weakest in method. Every number in this article carries its grade, and where a number could not be verified it was left out.

How seasonal is star-hotel occupancy in Bali, month by month?#

BPS star-hotel occupancy in Bali moved between 46.61% in March 2025 and 69.54% in August 2025, a swing of 22.9 percentage points. The chart below gives the twelve months of 2025 as BPS reported them. January is derived and not read directly, because the February reading of 51.62% came with an 8.66-point monthly fall. The simple unweighted mean of the twelve months is 60.6% (derived, Victaura Research arithmetic), close to the 60.7% annual figure for 2025 that GoodStats reports from BPS data.

The trough sits in the first quarter and the peak in the middle of the year. February and March are the weak months. March 2025 was 6.10 points below March 2024, which puts March 2024 at 52.71% (derived). From May the series climbs, passes 67% in July and peaks in August, then eases to 57.97% in November before December recovers to 60.88%. BPS, as reported in the press, tied the August 2025 reading in part to sports and arts events, a reminder that a single month carries event noise. Horwath HTL separately attributes the July and August peak in RevPAR to Australian winter demand and European summer travel.

Year on year, the comparisons available are mixed, and the full year is slightly softer. September 2025 at 68.17% was 1.83 points above September 2024 at 66.34%. February 2025 was 3.65 points below February 2024, and March was down 6.10 points. For the full year, GoodStats reports BPS data putting Bali's star-hotel occupancy at 60.7% in 2025 against 62.36% in 2024. The allocator should read 2025 as slightly softer than 2024, directionally, and not as a trend.

The Horwath HTL series reads about twelve points higher, and the gap is itself a finding. Its Bali Hotels Association data put July 2025 occupancy at 85.9% and March at 55.3%, with an annual average of 73.2%, against the 60.6% BPS mean. Horwath does not reconcile the two series, and neither does this article. Coverage differences are the first suspect, since BPS counts all classified hotels while the association data come from reporting members, but that is a hypothesis. A villa underwriter should anchor on the lower series and treat the higher one as an upper reference for professionally managed stock.

How do ADR and occupancy change as the rate band rises in Bali?#

In Bali's hotel market, from Rate D upward, occupancy falls as the rate band rises, and ADR is what holds RevPAR up. Horwath HTL splits 2025 performance into six bands by ADR. Rate A, above USD 501, ran 49.4% occupancy at USD 796 ADR. Rate B, USD 201 to 500, ran 69.2% at USD 286. Rate C, USD 141 to 200, ran 74.4% at USD 172. Rate D, USD 81 to 140, ran 78.9% at USD 106 (all measured, Bali Hotels Association data via Horwath HTL, 2025). Below Rate D the pattern breaks: Rate E, USD 51 to 80, ran 74.2% and Rate F, below USD 50, ran 68.1%.

The top band lost occupancy and gained rate in 2025. Rate A occupancy fell from 53.2% to 49.4%. Its USD ADR rose 3.9% and its IDR ADR rose 8.0%, so IDR RevPAR was flat at +0.1%, while USD RevPAR slipped 3.7% to USD 393. Horwath reads this as pricing power with softer demand depth, and that is also the reading here. A luxury asset that sells half its rooms is a different risk from a mid-market asset that sells four-fifths of them.

Rate B shows where resistance appeared. Its occupancy fell from 73.9% to 69.2%, and RevPAR fell 2.5% in IDR and 6.2% in USD, which Horwath describes as rate resistance after two years of strong growth. An upper-mid villa shares that exposure in direction, though the size of the effect for a villa is not measured anywhere in the public record.

The pipeline is top-weighted, which matters more than the current print. Of 5,641 tracked pipeline rooms across 45 hotels, Rate A accounts for 1,752 rooms and Rate B for 1,373, together 55% of the total (derived from Horwath HTL's pipeline table). A mature destination normally shows a mid-scale-heavy distribution, and Horwath notes that Bali does not. New luxury rooms will compete for the same premium guests whose occupancy already slipped in 2025.

Currency is a second variable the allocator carries. In 2025 Bali's IDR ADR rose 2% while the USD ADR fell 2%, which Horwath attributes to currency dynamics and exchange rate movements, at an average rate of IDR 16,705 to the dollar. A villa costed in rupiah and sold in dollars has an exposure that the hotel average hides, and it moves the result in either direction.

How do Uluwatu, Canggu, Ubud and Seminyak compare on ADR and occupancy?#

Horwath HTL's Jimbaran and Uluwatu submarket posted the highest ADR on the island in 2025 at IDR 4.8 million, on 66.0% occupancy. That is Horwath's own grouping, which merges the two areas, and the figures cover hotels and not villas. IDR RevPAR was IDR 3.1 million, the highest of the ten submarkets it tracks, with ADR down 1.3% and occupancy up 0.4%. Horwath calls this relative stability at the top end (measured, 2025).

Ubud priced its way through a softer year. Horwath records occupancy easing 4.8% to 68.6%, ADR rising 11.0% to IDR 3.6 million and RevPAR rising 5.6% to IDR 2.5 million. Seminyak ran 74.9% occupancy at IDR 2.4 million ADR, with RevPAR down 4.4% to IDR 1.8 million. Nusa Dua led the island on occupancy at 79.2%, helped by group and conference demand according to the report.

Canggu and Tabanan carry the weakest evidence. Horwath records 60.0% occupancy and IDR 2.2 million ADR, and warns that the submarket is newly tracked in 2025, that the sample base grew through the year and that year-on-year comparison is weak. Its 59.7% ADR jump and 19.5% occupancy fall should not be read as market moves. A principal who sees those two numbers quoted without the caveat is looking at a sampling artefact, and the grade for this submarket is directional.

Length of stay differs by area, and it changes what an ADR means. Horwath reports average stays of 2.4 days in Canggu and Tabanan, 2.5 in Ubud, 3.1 in Nusa Dua, 3.2 in Seminyak and 4 in Legian, with no figure given for Uluwatu. Short stays mean more turnovers per occupied night, and so more cleaning and laundry cost for each night sold. The next article in this series, on the nightly rate board as a comparables set, takes up how to read rates across areas with such different stay patterns.

For the Uluwatu peninsula the public record is thin, and the site matters more than the submarket. Horwath's grouping spans two distinct micro-markets, so it is a regional read and not a site read. Readers who want the site context for one specific location can open the Uluwatu villa development project page, which sets out the site and the scheme and states no return figures.

What is published for Gili Air, and what is not?#

For Gili Air the best published occupancy figure is 41.6% for 2024, from an industry consultancy, and Victaura Research could verify no public luxury ADR series for the island. C9 Hotelworks, in an article dated 11 August 2025, put overall Gili Islands hotel occupancy at 42% for 2024, up 4.4 points on 2023, with Gili Trawangan at 47.7%, Gili Air at 41.6% and Gili Meno at 36.9% (industry estimate, 2024).

The inventory is large and mostly small-format. The same source counts 364 properties and 4,150 keys across the three islands, with Gili Trawangan holding 166 properties and 2,320 keys, or 56% of the keys, and Gili Air holding 31% of the inventory. Visitors totalled 773,148 in 2024, 26% above 2019, with Gili Air taking about a quarter of them (industry estimate).

Island occupancy sits well below the Bali hotel averages, and the stay pattern explains part of it. C9 reports an average stay of two to three nights and a primary age group of 20 to 40. It estimates international-grade midscale to luxury properties at 50% to 70% annual occupancy, and gives no method for that range. The allocator should treat the range as the weakest grade in this article, directional at best.

The honest reading is that Gili Air has no public ADR series for villas. Rates there have to be built from advertised listings and dated observations, which is a data-collection task and not a published statistic. A 2024 island average of 41.6% also cannot be set against Bali's 2025 BPS months without noting that the years, the sources and the samples all differ. How occupancy converts into net owner revenue once management terms apply is covered in the third article of this series, on villa management economics.

Why is a hotel's ADR not a villa's ADR?#

A hotel ADR is a per-room price and a villa rate is a per-unit price, so the two numbers are not the same object. A luxury villa is typically sold as a unit with several bedrooms, a pool and a staff team. Its nightly rate rolls all of that into one price, while a hotel ADR divides room revenue by rooms sold. Setting a three-bedroom villa's nightly rate against a hotel ADR overstates the villa unless the allocator divides by bedrooms or compares revenue per available night on both sides (directional, a matter of definition).

The revenue mix differs. Hotel RevPAR covers room revenue only, while a hotel's total revenue also includes food, beverage, spa and events. A villa's revenue is mostly accommodation plus fees for extras. Horwath's Bali figures cannot say how much of a hotel's profit came from outside the room, so the ADR comparison is like for like on rooms and unlike on profit.

Occupancy is defined differently, and platforms blur it. A hotel's available rooms are the rooms in its inventory. A villa's available nights depend on whether owner-blocked and maintenance nights sit in the denominator. Platform vendors estimate this from booking calendars, as Airbtics states in its own methodology, so a blocked night can be read as unavailable or as unsold. The choice can move occupancy materially and is rarely disclosed.

Villa supply sits outside the formal count. Horwath's outlook says the steady flow of new villas and apartment-style units adds competitive accommodation supply outside the formal hotel sector. It also reports that the provincial government is tightening oversight of short-term rentals. Villa supply therefore competes with the hotel numbers while sitting outside them.

The correction is a method, not a multiplier. The sections below set it out. No single factor converts a hotel ADR into a villa ADR, and a quoted single factor should be treated as an assumption and not as a measurement.

What does short-term-rental platform data say about Bali villas?#

Platform-derived data for Bali in 2025 shows 64.0% occupancy and an ADR of IDR 1,477,708, with revenue per listing falling. Airbtics, which tracks active listings weekly and estimates figures from booking calendars, reports annual revenue per listing of IDR 348.6 million in 2025, down 10.94% from IDR 391.4 million in 2024 (industry estimate, modelled from calendars, all property types).

That ADR converts to roughly USD 88 at Horwath's 2025 average rate of IDR 16,705. The conversion is derived, Victaura Research arithmetic across two different sources, and it is shown only to size the gap. Against Horwath's USD 150 island-wide hotel ADR, the all-grades platform average sits well below the hotel average, which is consistent with a sample dominated by smaller, lower-priced units (directional).

Supply grew faster than revenue. Airbtics counts 8,919 net additions to active listings in 2025, a 29.4% supply expansion and the largest absolute increase among the seven Indonesian markets it analysed, and it classes Bali as declining. Revenue per listing fell as supply rose. For a luxury villa the relevant question is whether new supply competes in the same tier, and a platform average cannot answer it.

None of this is a luxury ADR. The sample runs from single rooms to large estates, and the report does not publish a luxury cut. A luxury villa ADR has to be built from comparable listings with dates and a stated method. Figures that circulate in short videos or on agency pages are not data, and none of them appears in this article. A luxury villa ADR for Bali quoted without a dated listing set is a number the public record does not contain.

How is a hotel figure corrected into a villa figure?#

The correction is a six-step bridge from hotel RevPAR to villa revenue per available night, and each step carries its own grade. The steps are directional. The allocator should replace each with a measured input before relying on the result, and should write down which input is weakest.

Step one is to anchor on RevPAR and not on ADR. Take the Horwath RevPAR for the submarket or the rate band, such as IDR 3.1 million for Jimbaran and Uluwatu in 2025, because it already embeds occupancy. Horwath does not publish the two cuts together. Then convert it to a per-key basis using the villa's bedroom count, so that the comparison is per room on both sides.

Step two is to rebuild occupancy for the villa and stress it on the trough. Use the BPS low months and not the annual mean for the downside case. March 2025 at 46.61% is the lower reference, and the 22.9-point swing across the year shows how far a single average can mislead. Count owner-use and maintenance nights in the denominator only as the management contract defines them.

Step three is to strip what a hotel earns outside the room and add what a villa earns outside the rate. Cleaning fees, pool and staff charges and platform commission all sit between the guest's price and the owner's revenue. Those are contract terms, and they decide the owner's share far more than the headline rate does.

Step four is to test the result against supply and compliance, and step five against currency. Check how many rooms and villas are entering the same tier in the same submarket, using Horwath's pipeline table, and confirm the licensing position at the date of reading. Then rerun the case with the rupiah moving in both directions. Step six is the write-down. Name the single input with the weakest grade, say so in the first paragraph of the memo, and do not let a clean spreadsheet hide it.

What supply and regulation sit behind Bali's luxury numbers?#

Horwath HTL tracks a hotel pipeline of 5,641 rooms across 45 hotels in Bali, a net rise of 284 rooms on the 5,357 rooms at the end of 2025. That is modest at the aggregate level. The clustering is not modest. Canggu holds 1,480 pipeline rooms, Jimbaran and Uluwatu 1,263 and Ubud 1,044, together 3,787 rooms or 67% of the total (measured, Horwath HTL tracked pipeline, March 2026; the share is derived).

Villa and branded-residence supply is growing outside those hotel counts. In its branded residences section, Horwath HTL, citing C9 Hotelworks market research, counts more than seventy hospitality-managed real estate developments actively on sale, in a report dated March 2026. Canggu and Berawa hold about 40% of that supply and Uluwatu 21%. Villas are about 18% of the total, up from 13% a year earlier, and about 84% of villa developments have fewer than thirty villas (industry estimate).

Tenure is shifting, which changes who the buyer and the renter are. Leasehold terms of 25 to 35 years dominate. Freehold offerings rose from about 12% of supply in 2025 to about 23% in 2026, driven by launches in Seseh, Pererenan and Nyanyi, while the report says other regions typically show 36% to 50% freehold inventory and Canggu and Berawa remain predominantly leasehold (industry estimate).

Compliance is the regulatory fact the allocator must check at the date of reading. Horwath reports that the provincial government set a mandate requiring all short-term rental accommodation to demonstrate full legal compliance by 31 March 2026, with listings that fail to verify registration through the national digital framework facing removal from major booking platforms. It also states that foreign individuals cannot directly hold the tourism operating licences required for short-term rentals, so foreign investors need a PT PMA structure, which it says carries a minimum committed investment above IDR 10 billion. Victaura Research has not verified the enforcement outcome after that date, and the allocator should confirm the position with counsel before using any rental projection.

What is weak in this evidence, honestly disclosed?#

The first weakness is that no series measures a luxury villa. BPS counts classified hotels, Horwath HTL reports hotel performance from a member sample, and the platform estimate covers all property types. Every number above is a proxy for a villa, and the correction in the previous section is a method and not a measurement.

The second is that the hotel sources disagree. The BPS mean of 60.6% for 2025 and Horwath's 73.2% differ by about twelve points on the same island and the same year, and the reason is not reconciled in either source. The allocator should carry both and name which one an assumption leans on.

The third is series construction. A 24-month BPS series was the intended table. The BPS website could not be opened for this check, so the table shows twelve months of 2025 from the monthly releases, with January derived from the February reading and its reported 8.66-point fall. The February, March, April, June, July, August, September and December readings were re-checked against press reports of the BPS releases, and May matches the June reading less its reported 6.56-point rise. The October and November levels could not be re-checked on an open page beyond the reported 6.60-point fall between them, and should be re-verified against the BPS tables on use.

The fourth is vintage and sample. The Gili figures are for 2024 and come from a consultancy, the 50% to 70% range for midscale to luxury properties has no stated method, Canggu and Tabanan are newly tracked, and the platform numbers are modelled from calendars. Each is labelled where it appears.

The fifth is what no number here says. Nothing in this article is a forecast, a return, a yield or a price for any project, and no figure from the market data should be read across to a specific asset without a site-level comparables set.

Skin in the game disclosure. Victaura, through its parent Greystone B.V. (Netherlands), develops projects on Lake Como (Italy), in Nungwi, Zanzibar (Tanzania), on Gili Air and in Uluwatu, Bali (Indonesia), and holds an off-plan capital position on Al Marjan Island, Ras Al Khaimah (UAE). Readers should assume commentary may be influenced by, or benefit, these positions. This document is classified as marketing material under MiFID II Article 24(3). It is not investment advice.