Office days and desk sharing: cutting space without wrecking the week

Many companies have made the same decision in recent years: shrink their office space and switch to shared desks, leaning on remote work. The savings are real, and significant — in major cities, a workstation costs between €6,000 and €12,000 a year, overheads and services included. Going from a hundred desks to seventy represents several hundred thousand euros a year.

Then comes the Monday morning when forty people are looking for thirty desks, when two teams that were supposed to work together came in on different days, and when someone who spent fifty minutes commuting ends up spending their afternoon in a meeting room on a video call. The real-estate savings are real; they're paid for elsewhere, in wasted time and frustration.

The problem is almost never the number of desks. It's the distribution of attendance, and the fact that it was never actually organised.

The imposed-days rule: why it so often fails

The reflexive answer is to impose days: "team A on Monday and Tuesday, team B on Wednesday and Thursday." It has the merit of simplicity, and it fails regularly, for three reasons.

It ignores family arrangements that are already in place. A remote-work day isn't a preference: it's often childcare, help for a parent, a medical appointment. Changing that day doesn't require a calendar tweak, it requires rebuilding an entire family arrangement. It's the leading cause of departures after an attendance reorganisation — and it's also the one nobody writes down in their exit interview.

It concentrates everyone on the same days. Tuesday and Thursday are everyone's preferred days, for commuting and rhythm reasons. A distribution imposed by team produces a packed Tuesday and a deserted Friday, meaning office space is sized for the peak and paid for all week.

It ignores what makes coming in actually useful. Coming into the office to sit through six video calls with people located elsewhere serves no purpose and durably undermines the rule's credibility. An employee who makes the commute needs to find something there that home doesn't offer: their project colleagues, a workshop, an issue that's hard to resolve remotely.

What you need to know before deciding anything

Three pieces of information are missing from almost every leadership team that decides on an attendance policy. They can be gathered in a week.

Actual occupancy, day by day, over two months. Not reported occupancy: entry counts, bookings actually honoured, or a simple visual headcount at 11am and 3pm. Most companies discover a twenty-to-thirty-point gap between what they believed and reality.

Real working pairs and trios. Who actually needs whom, physically, to get things done? This isn't the org chart. It's often cross-functional pairs — someone in support and someone in development, a salesperson and an analyst — who never cross paths because their days don't line up.

Everyone's fixed personal constraints. This is the most decisive piece of information, and the hardest to get honestly, because nobody wants to expose their private life in front of the team.

This last piece of data collection deserves to be done separately, without anyone seeing anyone else's answers. In a meeting, heavy constraints stay unspoken and preferences get voiced as though they were requirements: you'll end up building a schedule on false information.

OuiCanPlan exists precisely for this: you describe the situation in a few lines, each employee answers via a link in two minutes, confidentially, and you get back the attendance arrangements that are genuinely compatible with everyone's constraints, along with a breakdown of who is excluded by what when no perfect solution exists. You make the call with full knowledge of the facts. The results are real, with no advertising.

Four attendance models, and who they suit

A single team day. Each team has one shared, mandatory day of attendance, and the rest is free. It's the simplest and most effective model for organisations under a hundred people: a single constraint to remember, a single day to protect, and a natural spread if team days are staggered across the week.

A shared core plus free choice. Two identical mandatory days for everyone — for company-wide cohesion — with the rest left free. Beware: this model saturates two days and empties the other three. It only makes sense if the office space was sized for the peak, meaning the real-estate savings being pursued are modest.

A quota with no fixed day. Everyone must be present three days a week, at their own discretion, with a daily desk cap. The most comfortable for employees, the riskiest for collaboration: without a synchronisation mechanism, people who need to work together never cross paths. Best reserved for highly autonomous roles.

Event-based attendance. No recurring day; people come in for a workshop, a review, welcoming a new hire, a difficult meeting. Excellent for highly distributed teams, demanding in terms of coordination, and best avoided if the company has junior staff: you learn your job by sitting next to someone.

In all four cases, one shared condition applies: everyone's presence must be visible to everyone else, in advance. Without that, every trip to the office becomes a gamble.

New hires and juniors: the blind spot

An attendance policy designed by experienced employees almost always forgets those who don't yet have an internal network. A junior working remotely three days a week doesn't progress at the same pace: they don't overhear hallway conversations, don't dare interrupt with a thirty-second question, and take six months to understand what a deskmate would have explained in two weeks.

Two rules address most of this. The first three months should be mostly in-person, with a mentor present on the same days — not just assigned, actually there. And every new hire should have a designated buddy whose attendance days line up with theirs.

This isn't a favour done for the junior: it's what prevents a failed hire that costs twelve months of salary.

Reducing office space without hurting day-to-day life

The real-estate savings are legitimate. They become counterproductive when calculated on an average instead of a peak, and when they eliminate the spaces that make coming in bearable.

Three principles hold up:

Size for the second-busiest day, not the average. An average occupancy of 60% with Tuesday at 90% means desks are missing on Tuesday, every Tuesday. Sizing on the average is the leading cause of flex-office failure.

Keep more closed rooms than you think you need. In a shared office, video calls are the dominant activity, and a call taken in an open-plan space disturbs six people. Budget one small room or booth for every eight desks, not for every twenty.

Keep a personal storage space. A locker. It's negligible to buy, and it's what distinguishes an employee working flex-office from one who's camping out.

The real savings calculation subtracts, from the real-estate gain, the cost of time lost looking for a desk, meetings moved for lack of a room, and the turnover it causes. A well-sized flex-office setup remains very profitable; one undersized by 15% isn't profitable at all anymore.

Announcing the rule: four lines, and a reason

An attendance policy should be announced as a decision, with its reasoning included. The absence of a justification is what turns an acceptable rule into a source of ongoing conflict.

State what's expected, why, what remains free, and when the rule will be reviewed. For example: "Starting October 1st, each team has one shared day of attendance, set by the team itself. It's the only way to guarantee available desks every day while letting everyone keep their current remote-work days. All other days remain free. We'll review this in January."

The last sentence matters as much as the others: a rule announced as reviewable gets accepted, a rule announced as permanent gets challenged within the first week.

The three metrics to track

• Daily occupancy rate, to spot saturation before it becomes the topic of every conversation.
• The number of "pointless" office days: ask once a quarter, with a single question, how many times the person came in without finding it worthwhile. Two or more per month signals a poorly calibrated rule.
• The co-presence rate of working pairs: the only metric that tells you whether attendance actually serves a purpose. If people come in without ever crossing paths with those they work with, the policy has missed its target, whatever the overall occupancy rate looks like.

In summary

Flex-office and remote work are two sound economic decisions that fail for the same reason: they're decided on a spreadsheet, based on an average, without ever asking employees what's actually constrained in their week.

A confidential two-minute survey per person, sizing based on the peak rather than the average, one shared day per team, and a rule announced with its reasoning: that's what separates a lasting real-estate saving from one that gets reversed a year later, after three departures nobody saw coming.