Holiday Home Investment: Break-Even Night Count

Is a holiday home a good investment? Start with the break-even night count#

A holiday home covers its running costs once the nights it rents, multiplied by what each night leaves after variable costs, equal its fixed annual costs. That number of nights is the break-even night count. It is a count of nights, not a yield. The principal who asks whether a holiday let is a good investment gets a better answer from this count than from any headline return, because the count shows which cost drives the result.

The question usually arrives as a search phrase, and the phrase hides the method. People type 'is a holiday let a good investment' or, in German, 'ab wann lohnt sich eine Ferienimmobilie', which means 'from when does a holiday property pay off'. Both ask for a threshold. A threshold is something an allocator can compute, test and challenge. A promised return is not.

This note builds the threshold in four steps. It writes the formula, prices each cost line, sets the sellable nights against what official statistics measure, and runs one illustrative case. The case belongs to no real project. Every figure in it is an assumption, labelled as such. The output is a way of reading a holiday home, and nothing in it is a return, a forecast or personal advice.

The method has limits, and the note states them before the arithmetic. Break-even before tax and financing is not profit. A property can clear its break-even and still lose to a bond. It can also sit below break-even and serve its owner well. The count separates two questions that sales material blends, whether the asset pays its way and whether the owner wants the asset.

How is the break-even night count written?#

The break-even night count is the fixed annual cost divided by the net contribution of one rented night. In symbols, N* = F / (A x (1 - p - m) - c / L - v). A is the average nightly rate set for the house, before the platform and manager shares and excluding any tax the guest pays on top. p is the platform fee as a share of that rate. m is the manager's share. c is the cost of one turnover, L is the average length of stay in nights, and v is the variable cost of running the house for one occupied night. F is everything that is owed whether or not a guest arrives.

Each term answers one question, and the terms do not mix. Fees scale with revenue. Cleaning scales with the number of stays, so it is spread over the length of stay. Utilities used by guests scale with occupied nights. Insurance, local property taxes and the maintenance reserve scale with nothing. Keeping the four families apart is what lets a reader see which one moves the result when the rate or the stay length changes.

Break-even occupancy follows from the count once the sellable nights are fixed. The sellable nights are 365 less the nights the owner keeps and the nights the house is closed for works. Divide N* by that figure and the result is the occupancy the house must reach to cover its costs. That percentage, and not the count alone, is what can be set against an official occupancy statistic, provided the statistic measures the same thing.

The count excludes three items on purpose. It excludes income tax on the net result, because the rules depend on the owner's residence. It excludes the cost of capital, because financing depends on the buyer. And it excludes the change in the property's value, because that is a separate bet. An owner who wants a full return adds these lines afterwards. The break-even night count stays the operating floor underneath them.

What do platforms and managers take from each night?#

Platform fees are the first deduction, and they are stated in public. As of 13 October 2026, Airbnb's host guidance describes a single 15.5% service fee paid by the host, with exceptions, and 16% for listings in Brazil and Mexico. The same page says Airbnb moved from a split fee to a single fee structure in 2026. For the formula, p is therefore a published number for one platform and a market-specific number for the others.

Other platforms publish less, and the allocator should not fill the gap with an average. Booking.com commissions are agreed per property and per market. A thread in the Booking.com partner community about a rate increase for Greek properties shows the practical point. A commission is a contract term that can change, and optional visibility features can change it. The figure to use is the commission in the property's own Booking.com contract, which the seller or the manager can produce.

A manager's share is usually the largest single deduction, and it is a negotiated term, not a statistic. Management agreements are commonly priced as a share of rental revenue, sometimes with a fixed element and with maintenance charges passed through (directional, no official average exists). The illustrative case below uses a round assumption of 20% and then tests the case without it. Any real calculation replaces that figure with the clause in the signed agreement.

Fees apply to the whole booking, which is why p and m belong inside the bracket. If the platform takes 15.5% and the manager takes 20% of the same rate, the owner keeps 64.5% before any cost line. That holds when both shares are taken from the same gross rate. Where the agreement charges the manager's share on revenue after the platform fee, the owner keeps more, and m is adjusted to match. A reader who compares gross rates between two houses without this step is comparing numbers the owner never receives.

Which costs depend on stays, and which on nights?#

Cleaning and laundry follow the number of stays, not the number of nights. A turnover costs the same whether the guest stays two nights or ten. The formula therefore divides the cost of a turnover by the average length of stay. In the illustrative case a turnover costs 90 and the average stay is four nights, which puts 22.5 on each rented night. At two nights the same turnover weighs 45 per night.

Length of stay is the hidden lever, and official platform data can only partly speak to it. Eurostat's headline series reports nights, not stays. The Estonian statistics office separates nights rented from guest nights: in its definition, a family of four staying three nights counts as three nights rented and twelve guest nights. Neither unit counts stays, so a reader who wants L must find the number of stays in another source or take it from the manager's own records.

Consumables and utilities split into a base and a variable part. Electricity for pool pumps, standing water charges and internet are owed with an empty house. Air conditioning, extra laundry, welcome supplies and guest water use rise with occupancy. The illustrative case puts 12 per occupied night in the variable bucket and leaves the standing part in F. The split is a judgement, and the count is sensitive to it mainly when occupancy is low.

The honest way to fill these lines is with the house's own invoices. A year of utility bills and turnover invoices converts assumptions into measured inputs. Until then they are assumptions, and the case below says so. A resort house with a pool and air conditioning will carry a heavier variable line than a mountain apartment. The formula treats both the same way and lets the data differ.

What fixed costs does a holiday home carry every year?#

Fixed costs are the part of the formula that does not care about the guest. They include insurance, the local property taxes owed by the owner, the maintenance reserve, the standing utility charges, accounting, licence or registration fees where they apply, and any fixed element of a management contract. Each one is a line in a ledger, and the sum is F. A principal who cannot list the lines has not yet priced the asset.

Maintenance is the line most often understated. Pools, roofs, air conditioning and furniture age faster under rental use than under owner use. A reserve set as a share of replacement value is a more honest entry than the cost of last year's repairs. The illustrative case includes the reserve inside the fixed 14,000. It does not name a percentage, because no official source gives one for resort houses.

Registration is becoming a fixed condition of renting at all. Regulation (EU) 2024/1028, published in the Official Journal on 29 April 2024, requires a Member State that imposes data transmission on online short-term rental platforms to establish or maintain a registration procedure for units in the areas where that requirement applies (Article 4(2)). Where a registration procedure applies, Article 7(1)(b) requires the platform to ensure that the host has provided a registration number before allowing the offer, and that the number is displayed in the listing. For the count, a unit without a number may have no sellable nights.

Local taxes on guests and local taxes on owners are different lines. A stay tax collected from the guest and passed to the municipality is a pass-through. A property tax or a tax on the rental income is the owner's. The note gives no rates, because they differ by municipality and change by decree. The reader takes them from the authority that levies them, and enters them in F or in the tax step that follows the count.

How many nights can an owner actually sell?#

The sellable nights are fewer than 365, and the gap is the owner's own use. An owner who keeps 30 nights and closes the house for works for none sells 335. If the same owner keeps 60 nights, the number is 305. The break-even count does not change, because fixed costs and the contribution per night are the same. The occupancy needed on the nights left rises.

Owner nights are a cost, and the cost is the rental they displace. A night kept in the peak week gives up a night that would have earned the highest rate in the year. The article on owner nights against rental nights sets out how to price that displacement. Here it is enough to say that the formula removes owner nights from the denominator, and the principal decides how heavily they weigh.

Seasons make the average rate a weak input. A house that rents at a high rate in six weeks and at a low rate in the rest of the year has an average that describes neither. The formula can be run per season, with its own A, L and sellable nights, and the seasonal contributions added up against F. This is more work and gives a more honest picture. A single annual average hides the weeks that carry the year.

The second-home logic and the rental logic are two different tests. A house can fail the rental test and pass the second-home test, if the owner uses it and the rental income only offsets running costs. The piece on rental yield in prime resort markets reads the yield side. The count reads the floor under it. Neither replaces the owner's own reason for holding the house.

What do Eurostat's platform statistics measure?#

Eurostat's platform series counts guest nights, and a guest night is not an occupied night. For the third quarter of 2025, Eurostat reported 398.1 million nights spent by guests in EU short-term rentals booked through Airbnb, Booking or Expedia, 8.7% above the third quarter of 2024 (measured, official, experimental). Two guests in a house for one night are two guest nights, and one occupied night.

The series is useful for direction and weak for any single house. It shows that platform-booked demand in the EU kept growing through 2025 and into 2026. Press coverage of Eurostat's annual figure puts 2025 at 951.6 million guest nights, up 11.4% on 2024 (directional, press-reported). A press report dated 3 July 2026 gives 144.3 million for the first quarter of 2026, up 9.7% on a year earlier (directional, press-reported). Eurostat's own monthly article puts April to June 2026 at 258.8 million, up 5.3% on the second quarter of 2025 (measured, official, experimental). These are booked nights, with no count of nights offered, so they cannot yield an occupancy rate.

Coverage changed, and the note should say how. Eurostat's release states that Tripadvisor announced its exit from this business in late 2024 and no longer provides data, so the series now rests on three platforms. The data are provided directly by the platforms that share them with Eurostat. Bookings made directly with the owner, or through other channels, do not appear. A house that sells half of its nights direct is half invisible in this series.

Eurostat's occupancy statistic is a different instrument with a different population. Its net occupancy rate of bed places divides overnight stays by the bed places on offer, excluding extra beds, and by the days they were available, according to Eurostat's glossary entry on the net occupancy rate. Statistics Finland, which reports to Eurostat, separates it from a gross rate that divides by every bed place on every calendar day. Eurostat's article on hotels records net bed-place occupancy ranging in December 2019 from 22.3% in Greece, an estimate, to 48.0% in Spain (measured, data extracted June 2020). The spread is between countries, in one month and in hotels, so it is no benchmark for a holiday house. The method used to grade such numbers is in grading a number.

How should Zanzibar's bed-occupancy figure be read?#

Zanzibar's Office of the Chief Government Statistician publishes bed-space figures, and they measure beds, not houses. For January 2026, The Citizen reported on 11 February 2026, citing the office, 913,911 bed spaces available, of which an estimated 800,124 were sold, a bed occupancy rate of 87.5% (an official estimate, directional and press-reported, single month). The primary release was not opened for this note, which is a gap in the sourcing.

A bed occupancy rate and a house occupancy rate can differ widely. A villa with eight bed spaces rented to a couple sells one night of the house and, in a count of beds, a fraction of its capacity. Whether the office counts sold beds as capacity sold or as guests housed is a point to check in its methodology, which this note did not obtain. Until that is read, the 87.5% is a statement about the establishments surveyed and not about any single unit.

One month cannot stand for a year. The Citizen's report for that month also cites 100,216 international visitors, with Europe at 69.5% of them (directional, press-reported, same office). A twelve-month series would show lower months, and a break-even occupancy has to be set against the whole year or run season by season. Using a single high month as the benchmark would flatter any house, whatever the formula says.

Mainland Tanzania is measured by a different office. The National Bureau of Statistics runs a monthly hotel survey for the mainland, for example its Hotel Statistics release for December 2020, which carries its own occupancy estimate. Zanzibar's series belongs to a separate statistical office. A reader must not blend them. The wider point on arrivals, beds and demand is in arrivals are not demand.

A worked case, declared illustrative#

The case below is invented, and no real house or project stands behind it. The average rate is 300 a night. The platform fee is 15.5%, the figure Airbnb states. The manager takes 20%, an assumption. A turnover costs 90 and the average stay is four nights. Variable cost per occupied night is 12. Fixed annual cost is 14,000. The owner keeps 30 nights. Currency is unspecified, since the point is the shape of the arithmetic.

Step one is the contribution of a rented night. The owner keeps 64.5% of 300, which is 193.5. Cleaning takes 22.5 a night, being 90 divided by four. Variable cost takes 12. The contribution is 159 a night. Step two divides the fixed 14,000 by 159, which gives 88.1, or 88 nights to the nearest night. Step three divides 88 by the 335 sellable nights, which gives a break-even occupancy of 26.3%.

The table varies one input at a time, except the turnover row, which changes two. Removing the manager lowers the count to 64 nights, before any cost of the work the manager did. Cutting the average rate to 220 raises it to 130. Raising fixed costs to 20,000 lifts it to 126. A turnover of 150 spread over three-night stays lifts it to 106. In these steps the rate and the fixed costs move the count most. The steps differ in size, so the ranking holds for these steps only.

The table is arithmetic, not evidence about any market. It shows how the formula behaves. It does not say that 300 is a realistic rate anywhere, that 20% is a usual management share or that 14,000 is a usual cost base. A principal who runs the formula replaces every input with a quote, an invoice or a contract clause, and then the output means something about that house.

What are the weak points of this method, honestly disclosed?#

The first weakness is that every input is an estimate until it is a document. The platform fee is published, but the average rate, the length of stay and the manager's share are not. A break-even computed from forecasts inherits their error, and the error compounds. Nothing in this note reduces it. The method moves the uncertainty onto named inputs, which is its only claim.

The second is that official occupancy statistics do not match a single house. Eurostat's platform series counts guest nights across three platforms and has no denominator. Its occupancy series covers establishments, a mostly hotel population. Zanzibar's series counts bed spaces, and the version the author could read was a one-month figure relayed by the press. None of them gives the occupancy of a private villa, and the note does not pretend otherwise.

The third is that the break-even ignores time and risk. Fixed costs paid in a quiet year and income received in a busy one are not the same cash flow. A house that reaches break-even in October has run a cumulative shortfall, met from reserves, for nine months. The count says nothing about vacancy shocks, regulatory change or a platform fee that moves, as Airbnb's did in 2026.

The fourth is that the note covers neither tax nor financing, and gives no advice. Tax depends on the owner and the jurisdiction. Financing depends on the lender. Both can decide whether a house is worth owning, and both sit outside the count. Readers who need them should take them from the competent authority and a qualified adviser, not from a research note.

The fifth is that the sources carry different grades. The Eurostat platform series is measured and official but experimental. The Zanzibar figure is an official estimate read through the press, so it is directional and press-reported. The 2025 annual total and the first quarter of 2026 are press reports of Eurostat releases, while the second quarter of 2026 comes from Eurostat's own article. The manager's share, the rate and the cost base are assumptions. The grades are shown beside each number for that reason.

How should an allocator read the break-even night count?#

The result is a floor, and the allocator reads it as a distance. If the break-even occupancy is 26% and a comparable official statistic for similar units sits well above that, the house has headroom on paper. If it is 45% and the comparable figure is 50%, the headroom is thin. The comparison is only valid if the statistic measures the same thing. None of the statistics in this note does, so the closer benchmark is the twelve-month booking calendar and payout statements of comparable units, requested from the manager or the seller.

Three questions follow from the count. Which input, if wrong by a fifth, moves N* most? Which official statistic is closest to the house's own occupancy, and what exactly does it measure? How many nights does the owner keep, and what do they cost? These can be answered with documents. They are the questions a lender, an auditor or a partner would ask in any case.

What the count never gives is a promise. It does not say the nights will be rented. A projected occupancy is a separate claim with its own evidence, and the difference between the floor and the projection is the risk the owner carries. Financing adds a further layer, read in rates and prime property, and it sits outside the count.

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.