Time off, on-call and cover: building a team schedule without favouritism

In most teams, summer leave planning gets settled in March, in the space of a week, and it gets settled badly. The first people to submit their dates get what they want; the last people inherit whatever's left; and an unwritten rule eventually takes hold, where seniority or having children grants priority. Nobody decided it, nobody owns it, and everyone has internalized it.

The leave schedule is probably the most flammable topic in team life. It touches family life, a sense of fairness, and the unspoken balance of power between colleagues. It therefore deserves the one thing that reliably calms this kind of topic: an explicit rule, announced before collection begins, applied without exception.

Why "first come, first served" is the worst system

It's the default rule in most organizations, precisely because nobody chose it. It has three serious flaws.

It rewards administrative availability, not need. Whoever checks their email first on Monday morning wins. Whoever was in client meetings all week loses. That criterion has nothing to do with fairness.

It creates a race that pushes everyone to submit their dates as early as possible — before they actually know — and then to change them later. The schedule becomes unstable at exactly the moment it should be settling down.

It makes conflict invisible. Two people who wanted the same week never talk to each other; one simply learns it's already taken. A negotiation that would have taken three minutes never happens.

Set the framework before asking for anything

Before any collection begins, the manager settles and announces four things. This announcement is worth half the work.

The minimum coverage level, by period and by skill. "Two people out of five present in August" is a framework. "We'll figure it out" is not. Break it down by critical skill: if only one person knows how to close the monthly books, the constraint applies to them, not to headcount overall.

Peak and blocked periods, with their reasons. Month-end close, back-to-school for an education-software company, a seasonal peak. A justified restriction is respected; an unexplained one is contested.

The tie-breaking rule, announced beforehand, not after the fact. Three defensible rules exist: year-over-year rotation (whoever had August last year gives it up this year), an open lottery, or priority given to school-holiday constraints, owned explicitly as a company choice and compensated elsewhere for those who don't benefit from it. What's not defensible is failing to decide and letting the rule form on its own.

The response deadline, and what happens to people who don't respond. Without this, you'll be chasing people down four times.

An imperfect rule announced in advance is always better received than a perfect rule applied after the fact.

Collecting: what you actually need to ask

A shared spreadsheet where everyone writes in their weeks produces three well-known problems: early entries influence later ones, everyone sees their colleagues' choices and either self-censors or reacts to them, and you only get one piece of information — the dates — when arbitration actually requires five.

The questions that genuinely let you build a schedule:

1. Your preferred weeks, ranked, with at least two options.
2. The weeks that are genuinely impossible, and why if you want to say — a rental already booked, a wedding, a custody schedule fixed by court order. The difference between "I'd prefer" and "I can't" is the single most valuable piece of information in the process.
3. Your flexibility: would you accept shifting by a week if it unblocks the schedule? In exchange for what?
4. Your booking constraints: by what date must you confirm a rental or a ticket? Some people have a real deadline, others don't.
5. What you know how to do that few others do, to assess coverage by skill rather than by headcount.
6. Your availability for on-call or standby duty over the period, and your absolute blackout dates.

These answers should be confidential. Someone who has to explain in front of the whole team why they're set on the second week of August ends up giving up asking for it. The same person, answering privately, tells the truth — and it's that truth that lets you build a schedule that holds.

OuiCanPlan handles exactly this step: you describe the period and your coverage constraints in a few lines, each person answers via a link in two or three minutes without seeing anyone else's answers, and you get back a readable summary — who wants what, who can't, where the collisions are, which combinations meet your coverage target. You decide based on data, not on pressure from whoever pushed hardest last. Les résultats sont réels et sans publicité.

Building the schedule: a four-pass method

First pass: place the hard blocks. Everything locked in — a custody ruling, a wedding, a non-refundable rental, medical treatment — goes in first and doesn't move. This is usually about 20% of requests, and it shrinks the problem space enormously.

Second pass: place requests with no conflict. A good half of the wishes don't collide with anything. Confirm them immediately and say so: half the team is settled within twenty-four hours, and the tension eases.

Third pass: resolve the collisions. For every over-requested week, apply the announced rule. Before deciding, put the people involved in touch with each other: in half the cases, they'll find an arrangement you wouldn't have thought of, because one of them had flexibility they hadn't bothered to mention.

Fourth pass: check actual coverage, skill by skill, week by week. This is when you discover that three people are out in week 33 and neither of the remaining two knows how to handle supplier emergencies. The solution isn't to refuse someone's leave, it's to train someone before summer — but you have to have caught it back in April.

The special case of on-call and standby duty

Leave gets divided up once a year; on-call rotations come around constantly. They follow different rules.

Rotation should be mechanical and visible across twelve months. A published annual calendar beats twelve monthly negotiations. Everyone can see their weeks ahead of time and plan their life around them.

Hardship periods should be counted separately. Christmas, New Year's, long-weekend bridges and summer holiday weekends don't balance out against ordinary weeks. Keep a separate tally of hardship periods, over several years, and circulate it. It's the only way to stop the same people always ending up with December 24th.

Swaps between colleagues should be allowed and logged. Two people trading weeks helps everyone; the problem isn't the swap, it's the unrecorded swap that nobody remembers when it's time to tally things up.

Compensation must be known upfront: time off in lieu, a bonus, a rest day. On-call duty whose payoff gets negotiated after the fact reliably breeds resentment.

The most common mistakes

• Collecting with no framework. Asking for "your preferences" without stating the coverage target implicitly promises that everything will be granted.
• Deciding without explaining. An unexplained refusal feels like personal arbitrariness. The same refusal, explained in two sentences, is almost always accepted.
• Forgetting part-timers and interns. Their schedules follow different constraints, and they're routinely left out of coverage calculations.
• Writing nothing down. With no record of who got what and when, next year's rotation is impossible to justify, and it's always whoever pushes hardest who wins.
• Handling the schedule alone. Leave, illness, departures happen: always name a second person with the same information and the same access.

What you gain from doing this

A schedule built on an announced rule, confidential collection and a written record takes half a day of work, versus two or three weeks of diffuse negotiation under the default system. It produces fewer conflicts, fewer last-minute changes, and above all it dismantles the widespread — and often well-founded — belief that the same people always get what they want.

This is a topic where perceived fairness matters almost as much as actual fairness. A visible rule, applied without exception and documented from year to year, delivers both at once.

The timeline for collecting requests

When you ask matters as much as how. Three timelines work in most organizations.

Summer is handled in February or early March. Any earlier, people don't know yet; any later, rentals are booked and positions have hardened. Allow two weeks for responses, one week for arbitration, and a firm publication before the end of March.

End-of-year holidays are handled in October. It's the most contentious period of the year, because Christmas and New Year's don't substitute for each other in blended families. A strict two-year rotation, written down and available to consult, settles almost everything.

Long weekends are handled as early as January, for the whole year. There are few of them, they're known well in advance, and they divide up very cleanly by rotation.

In all three cases, the same sequence applies: framework announced, confidential time-limited collection, decisions published along with their reasoning, schedule locked. What's locked can of course still be swapped between colleagues afterward — as long as the swap gets recorded somewhere.

What to keep from one year to the next

A very simple table, kept by the manager and accessible to the team, is enough: who got which period, who gave one up, who took which unpleasant on-call shift. Three columns, one row per person per year.

This document does two things. It makes rotation possible, since everyone knows who had what. And it defuses the most painful conversation in team management — the one where someone claims it's always the same people who get what they want: with a record, the question is settled in thirty seconds, either way. Without one, it never gets settled.