How we check what is on this site

This page exists so that a number on this site can be argued with specifically rather than trusted vaguely.

Three classes of number, and how to tell them apart

	What it means	How it looks
Quoted	A regulation says it, or the company that charges it published it on its own public rate card. The source is linked at the foot of the page with the date it was read.	No marker.
Reported	The figure circulates in the trade, but the party charging it publishes no public rate card, so nobody outside their contract has read it. Booking.com and Agoda commission are the two on this site.	A reported marker.
Modelled	The model below produced it. Always a range.	A modelled marker.

If a figure would be none of the three, it does not get published. Pages under construction carry a visible placeholder instead, and the release gate refuses to pass a build that still contains one.

The model, input by input

1. Channel mix

Platform commission is not one number, because a villa is rarely on one platform. The model weights the published and reported rates across an assumed mix:

Channel	Share	Commission used	Class
Airbnb	45%	15.5%	Quoted — Airbnb publishes it
Booking.com	35%	15%reported	Reported
Agoda	10%	15–25%reported, midpoint used	Reported
Direct	10%	0%	Assumption

That weights out to a blended platform commission of 14.2%^modelled of booking revenue. The mix is an assumption, not a measurement. No primary source publishes a Bali-wide channel split. If your villa is 80% direct, your blended rate is much lower and the model will understate what you net; the estimator lets you change it.

2. Operating cost bands

Each band is a share of gross booking revenue. The fourth column decides whether a markup can be applied to it, which is the difference between two companies advertising the same percentage.

Cost	Tight	Loose	Who buys it
Housekeeping, laundry and guest consumables	6%	10%	A manager may buy this for you
Pool, garden and routine maintenance	4%	7%	A manager may buy this for you
Utilities: electricity, water, internet	4%	8%	Yours, never rebilled
Repairs, replacement and wear	3%	6%	Yours, never rebilled
Channel manager and payment processing	2%	4%	Usually the manager

The markup applied to services bought on an owner's behalf is taken as 10–25%^reported. Applied by some managers to cleaning, laundry and maintenance bought on the owner's behalf. No operator located during this build publishes a markup figure. The clause to read is whether third-party costs are rebilled “at cost” or “cost plus”; if the contract is silent, it is cost plus.

3. The order the deductions bite in

Order is not cosmetic. Two of these are charged on a larger base than most owners assume:

1. Accommodation tax first. It is charged on what the guest pays and it is never the owner's margin. Over a one-month booking it does not apply at all.
2. Platform commission on the booking value.
3. Management commission on booking revenue — not on what survives the platform. Every rate card read for this build is charged this way.
4. Channel manager and payment processing.
5. Operating costs, plus markup on the ones the manager buys for you, plus any staff payroll billed separately. A plan that excludes salaries is that plan plus a payroll.
6. Final income tax on the gross rent, not on the profit.

4. What the model produces

Output	Tight operation	Loose operation
Intermediary layer, share of gross	37.2%	42.5%
Multiple of the advertised commission	1.9×	2.1×
Reaches the owner, per 10,000 of gross	USD 3,220	USD 1,488
Total gross-to-net compression	68%	85%

The two compression figures on this site are different measurements, not rival estimates of the same thing. The intermediary layer (37–42%^modelled) is what the middlemen take. Total compression (68–85%^modelled) adds operating costs and Indonesian tax, which are real money leaving but are not anybody's commission. Quoting the larger number as if it were the fee is the mistake this site was built to stop, so it is not one we are going to make in the other direction.

Where the arithmetic exists twice

The estimator runs in a browser, so its arithmetic is written a second time in JavaScript. That is the one place a single source of truth cannot protect itself, and it is how published figures drift apart: the reference site in an adjacent vertical shows one statutory cap in its body copy and a different one in its own structured data on the same page, because a yearly uprating was applied in one place and not the other.

So the build extracts the real functions out of the shipped JavaScript, runs them against the Python over a grid of villa sizes, commission rates, channel mixes and stay lengths, and fails if any case disagrees. It is a build step, not a habit, because eyeballing it works exactly once.

Dates and corrections

Every page carries the date its rates and regulations were checked. This build checked them on 20 September 2026. Indonesian regional tax rates are set by each regency and change; platform fees change; published rate cards change without notice. A figure here with a date on it is a claim about that date and nothing more.

Found an error? Write to contact@balivillacare.com. Corrections are made on the page with the date of the correction, not silently.

Questions about the method

Why publish the model at all?

Because the alternative is what the nearest comparable site does: publish cost figures described as synthesized from public market data, with no sample size, no collection period and no source. A number you cannot argue with is a number you cannot check. Everything below can be disagreed with line by line, and if you think an input is wrong you can substitute your own and re-run it.

Is the estimator a valuation of my villa?

No. It is the published model applied to three inputs. It has not seen your villa, your photographs, your reviews, your calendar or your statements, and it cannot know your area's real achieved rate. Treat the output as a way of framing a question for a manager, not as a number to negotiate against.

Where do the operating cost bands come from?

They are modelled, not measured, and the page says so wherever they appear. No primary source publishes Bali-wide villa operating costs; anyone presenting one as measured has not said where they measured it. The bands are wide on purpose and the difference between the tight and loose end of them is larger than the difference between two management companies' headline rates, which is itself the most useful thing in this model.

Why is the management commission charged on booking revenue rather than on what is left after the platform?

Because that is what the rate cards say. Every published rate card read for this build states its percentage against booking revenue, rental income or the published rate. None of them charges on net. It matters: at a blended platform commission the difference is several percentage points of gross.

Why does the final income tax come off the gross rather than the profit?

Because PP No. 34 Tahun 2017 charges it on the gross rental amount and it is final, meaning it is not reduced by expenses and not recomputed at year end. This is the step that makes a weak year worse than merely flat, and it is routinely left out of yield figures quoted to buyers.

What is not in the model?

Purchase price, lease amortisation, furnishing and refurbishment cycles, insurance, exchange rate movement between the currency you spend in and the rupiah, and any income tax you owe in your own country of residence. The model answers one question — what happens to a booking between the guest and your account — and it is a mistake to read it as a return on capital.