How to Fairly Divide Weeks in a Shared Holiday Home: 5 Models Compared

Why peak weeks cause almost every argument

A holiday home has fifty-two weeks, but a shared one really only has about eight that anyone fights over: the school summer holidays, Christmas, New Year, Easter and the long weekends. Everything else tends to sort itself out with a message in the group chat.

This is worth naming explicitly, because it changes what you need to solve. You do not need a scheduling system for the whole year. You need an agreed procedure for a handful of contested periods, and a simple way to book the other forty-four weeks without ceremony.

The failure mode is almost always the same. Nobody decides on a procedure, so the default takes over: whoever books first, wins. That quietly rewards whoever checks the calendar most often, or whose job makes it easiest to plan eleven months ahead — and the family that loses out twice in a row stops believing the arrangement is fair.

Model 1 — Rotating first choice

Each year, one party picks first, then the next, and so on. The order rotates: whoever chose first this year chooses last next year. With four families on a four-year cycle, everyone gets one first pick, one second, one third and one last.

This is the most widely used model among families sharing a house, and for good reason: it is easy to explain, obviously even over time, and needs no arithmetic.

• Best for: three to six parties who use the house regularly and plan to keep it for years.
• Weakness: it is only fair across the full cycle. The family that draws last place in year one has to take it on faith that year four will come.
• Make it work: write the rotation order down for the whole cycle at the start, so nobody has to relitigate it annually.

Model 2 — The draft (snake order)

A variation on rotation, borrowed from sports drafts. Instead of one pick each, parties take turns choosing weeks in order — and then the order reverses for the second round. With three families the order runs A, B, C, then C, B, A, then A, B, C again.

The reversal is the whole point: it compensates the party who picked last in the first round by letting them pick twice in a row at the turn.

• Best for: groups where each party takes several weeks a year, not just one.
• Weakness: needs everyone available at roughly the same time to make their picks, or a clear deadline per turn.
• Make it work: set a response window — 72 hours is common — after which the turn passes to the next party and can be picked up later.

Model 3 — Points or budget

Every party gets the same number of points per year. Weeks are priced: the first week of August might cost 10 points, a wet week in November 1. Each family spends its budget how it likes — one big peak week, or a scattering of cheap ones.

This is the most flexible model, and the one that best handles genuinely different preferences. The family with school-age children and the retired couple who prefer quiet shoulder seasons stop competing with each other entirely.

• Best for: groups whose needs differ a lot, or where some parties care much more about specific weeks than others.
• Weakness: someone has to price the weeks, and the pricing itself can become the argument.
• Make it work: price the calendar once, review it every two or three years rather than annually, and let unspent points roll over at most one year.

Model 4 — Ballot or lottery

Contested weeks are drawn at random. Simple, unimpeachably neutral, and requires no negotiation — which is exactly why some groups like it and others hate it.

The pure version has an obvious flaw: chance has no memory, and one family can lose three years running. The fix is a weighted ballot, where anyone who missed out last year gets extra entries this year.

• Best for: larger groups, or ones where relationships are strained enough that any human decision would be suspected of bias.
• Weakness: unweighted, it can produce genuinely unfair runs.
• Make it work: always weight it by who lost last time, and draw in front of everyone — a video call is enough.

Model 5 — Fixed split

Each party simply owns the same weeks every year. The Müllers always have the first half of August; the Schmidts always have Christmas.

It sounds rigid, and it is, but it removes the annual negotiation entirely and lets families plan years ahead. For groups whose preferences genuinely do not overlap, it is the least-effort answer available.

• Best for: two or three parties with naturally different preferences, or houses where one family lives much closer.
• Weakness: it hard-codes whatever inequality existed when you set it up, and it cannot absorb life changes — a new baby, a new job, a divorce.
• Make it work: agree a review point (every three years, say) where anyone can ask to renegotiate without it being taken personally.

Which model should you pick?

There is no universally best answer, but there is a reliable way to choose: match the model to how similar your preferences are and how much administration you will tolerate.

• Preferences barely overlap → fixed split. You are solving a problem you do not have.
• Preferences overlap and you want minimum admin → rotating first choice.
• Several weeks each per year → draft with a response deadline.
• Very different needs and appetites → points.
• Trust is low, or the group is large → weighted ballot.

Whichever you choose, two rules matter more than the choice itself. Write it down, and define what happens when someone misses their turn — an undefined deadline is how a good system quietly reverts to first-come-first-served.

Putting it into practice

A model on paper still needs somewhere to live. Most groups start with a shared calendar, which shows who is staying when but has no opinion about who was entitled to book — anyone with edit access can simply take a week, and nothing warns anyone.

Keyra was built for this: stays are requested and approved rather than written straight in, and the Pro plan automates a rotation for the periods you mark as contested. Each owner gets a turn in an order that rotates every year, with a response deadline that passes the turn on if nobody replies. Every other week on the calendar keeps working normally — you only formalise the weeks that actually cause arguments.