How to split a vacation rental without an argument
Most of the difficulty in dividing a house is not the arithmetic. It is that the group starts negotiating after everyone already knows which room they want. Here is an order of operations that avoids most of the usual trouble, whether or not you use any particular tool.
1. Agree on the rule before anyone sees the rooms
This is the single highest-leverage step, and it is free.
Once people have seen the listing photos, every proposal about how to divide the rooms is also a proposal about who gets the balcony. Nobody can evaluate the rule on its merits any more, because everybody can see what the rule does to them personally. Suggestions start getting read as manoeuvres, and sometimes they are.
Decide the method while the house is still hypothetical — when you are choosing between listings, or the moment someone says “should we book it?” At that point people argue about what is fair in general, which is a conversation groups are good at.
If you take nothing else from this page: the rule is easy to agree on before the booking and hard afterwards.
2. List every bed, not every room
Count sleeping spots, not doors. A room with two twins is two beds that two different people can be assigned. A room with a queen is one bed for one person, or one bed for a couple.
Get this right, because the bed is also the thing that carries money. A bed holds a share of the trip total, and whoever sleeps in it pays that share, split between them if there are two. So a queen that sleeps two costs the same whether one person takes it or two do: alone it is the whole share, shared it is half each. That is what makes a solo attendee’s choice between a queen and a bunk a real one, and it is why a group that splits by head instead quietly has the bunk paying down the queen.
This distinction matters because the thing people are actually allocated is a place to sleep. A group that divides by room ends up negotiating a second time over who takes which bed inside the shared rooms, usually on arrival, usually while tired.
3. Describe the beds honestly, including the bad parts
Whatever method you use, people can only express preferences over things they know about. The descriptions are the entire information set, and vague ones produce meaningless rankings.
The details that reliably change how people rank a bed:
- Bathroom. Ensuite, or shares the hall bath, and with how many people.
- Floor and stairs. Ground floor versus third floor matters to more people than you would guess, and to some people it is decisive.
- A door that closes. An open loft or a pull-out in the living room is a different product from a bedroom, and should be labelled as one.
- Light and noise. Faces the road, above the kitchen, no window.
- Bed size, and whether it is bunks.
Write the unflattering ones down. A person who chooses the windowless room knowing it is windowless has made a real choice; a person who discovers it on arrival has a grievance, and they are right to.
“Shares the hall bath” tells you a bathroom is shared, not where it is. Nothing in a bed list captures adjacency — which room is next to which, how far the bathroom is, who you can hear through the wall. If those things matter to your group, they need to be said out loud; no ranking form will surface them.
4. Collect preferences privately
Everyone says which beds they would genuinely accept, in order, and how much they would pay above or below an even split for each.
Two rules make this work:
- Nobody sees anyone else’s numbers. Visible bids turn the exercise into a negotiation, and people start responding to each other instead of stating what they want. It is also what keeps strategic bidding impractical — the mechanism is not strategyproof, and privacy is most of what substitutes for that.
- Adjustments have to sum to zero. If you push one bed up $100, you have to push others down by $100 in total. This is what stops the form being “I would like to pay less for everything,” which is true of everyone and tells you nothing. You are stating relative value: how much more this bed is worth to you than that one.
The number to submit is the one that is actually true for you. It is tempting to bid tactically, but doing it well requires knowing what everyone else submitted, and you do not. Bidding your real values is also what makes the outcome defensible afterwards — you cannot object to a result computed from numbers you invented.
You are not bidding against your friends. You are answering one question: how much better is this bed than that bed, to you? If the honest answer is forty dollars better, write forty dollars. Who ends up where, and who pays what, is the program’s problem — and it solves it better the more truthful that number is.
5. Handle couples explicitly
Two people sharing a bed should be treated as one participant with one set of preferences and one price to divide however they like between them. Both of them have to confirm the pairing, from their own ballot. That is not a formality: a one-sided declaration would let anyone attach themselves to somebody else’s numbers and be valued as a pair on the strength of it.
Who is comfortable sharing with whom is settled by the group before anybody opens the form. It is not something an allocator can be told or should try to weigh.
6. Run it once, and publish the result
Compute the assignment and the prices, then show the whole group who got which bed at what price.
Publishing the result is not just courtesy — it is what makes the outcome checkable. If the method has the envy-free property, then anyone can look at the full list, compare their own bed and price against everyone else’s, and confirm that they would not rather swap. A result nobody can inspect has to be taken on trust, and trust is what you were trying to avoid spending.
Do not rerun it because somebody is unhappy with the outcome. Rerun it only if an input was wrong — a bed described incorrectly, a person who never submitted, someone dropping out. Re-rolling until the result is popular is the same as not having a method.
The awkward cases
Somebody joins after the allocation
Add them, and rerun the whole thing rather than slotting them into whatever is left. A late arrival given the leftover bed at the average price is the original problem in miniature.
Somebody drops out
Their share does not simply vanish — the house costs what it costs. Remove them and rerun; everyone’s price moves, and it should.
The price gap looks enormous
A wide gap between the best and worst bed is what a method looks like when it is working: the premium has to be big enough that the people who did not get the good room genuinely prefer the money. On a real trip we measured, the fair method produced a wider spread than the unfair one it replaced.
If the group finds the spread unacceptable anyway, that is a legitimate position, and it is worth stating it as its own rule up front — a cap, or equal bidding budgets — rather than discovering it when the numbers appear.
Someone refuses to submit
Chase it before you run the allocation, not after. A missing ballot is not neutral; the allocation runs over a smaller group and that person gets whatever is left. It is also the most common way this goes wrong in practice — not disagreement, just somebody who never got round to it.
Why splitting rent by room is harder than it looks — what goes wrong with even splits, rankings, and auctions, and why the obvious compromise between them does not work either.
Room Selector 5000 does steps 2 through 6. You list the beds, share one link, and everyone ranks privately.
Set up a trip