Demand is not the question in Bali. Arrivals hit a record in 2025, and the island remains one of the most searched leisure destinations on earth. The question that actually determines whether a villa makes money is what happens on the supply side, and that number is rarely put next to the arrivals figure in the same sentence: more than eighty thousand active short-term rental listings across the major platforms, at an average occupancy of roughly 47 percent. The exact count shifts between checks as the underlying database updates; AirDNA showed 84,264 active listings when checked directly on 7 July 2026, and a separate check around the same period showed a different but comparably sized figure. The precise number is less important than the order of magnitude and the occupancy rate, both of which have held steady.
Read those two numbers together and the picture changes. Demand grew. Supply grew faster. The average listing in Bali is now vacant more nights than it is occupied.
Why an average this low is not a warning sign for every villa
An average blends everything: the unlicensed one-bedroom with borrowed furniture and the professionally operated, correctly licensed four-bedroom with real photography and dynamic pricing. Those two properties are not competing in the same market, even though they sit in the same dataset. The 47 percent average is dragged down by a long tail of supply that should probably not exist as an investment in the first place: illegal listings that carry growing distribution risk under the licensing enforcement now underway, properties with no real operating discipline behind them, and villas bought for lifestyle reasons and rented out as an afterthought.
Supply stopped being the differentiator around the time supply stopped being scarce. What differentiates a villa now is whether it was built to a standard and run by people who treat occupancy as a discipline, not an accident.
What actually separates the top and bottom of that range
Three things, consistently, in what we see on the ground: licensing status, because platforms are now technically able to check it and unlicensed supply is starting to disappear from search results rather than merely rank lower; design and build quality, because guests can tell the difference in photographs before they ever arrive; and active operations, meaning real pricing management, real guest communication, and real maintenance, rather than a listing that was set up once and left alone.
The reframe this data supports
More than eighty thousand listings sounds like a bad time to invest. For a generic villa, that is probably true. For an asset that is correctly zoned, well designed, properly licensed and run with real discipline, it is mainly a sign that the market is becoming an operator's market: a genuinely well-built, correctly licensed, professionally operated asset stops competing on price and starts competing on the fact that most of its nominal competition should not really be counted as competition at all.
Harmonie's three villas average more than 80 percent occupancy in that same oversupplied market, and the collection is forecast to close 2026 24 to 40 percent above 2025. That gap between one listing among tens of thousands and a specifically designed, professionally run one is not accidental, and we explain exactly how in a separate note on our operating model.
Real numbers do not travel in public. In a serious first conversation, we open the Harmonie model: real occupancy, real rate, revenue, operating cost lines, and what we would do differently.
Talk to Françoislast reviewed: july 2026
Related: What PP 28/2025 means for villa owners · Understand the Harmonie proof
Sources: AirDNA, Bali market overview (84,264 active listings, 47% occupancy, $148.8 ADR, $13.3K annual revenue per listing), checked directly 7 July 2026; Badan Pusat Statistik (BPS) Bali, 2025 foreign arrivals. Figures move monthly on AirDNA and should be reconfirmed for any investment decision.