Most return-to-office policies are followed and still fail. Someone comes in on Tuesday because the policy asks for three days a week, and the two colleagues they needed are working from home that day. Nobody broke a rule and everyone hit their three days, but the days did not line up, and the trip was wasted.

None of that shows up in an attendance report, which is part of why the problem persists so quietly. JLL's 2026 occupancy benchmark, covering 84 organizations and 716 million square feet of office space, found that 62% of organizations now require a fixed number of office days, up from 49% a year earlier. Utilization still sits at 56% against a target of 74%. The rules tightened faster than the offices filled.

The number of days is the least important decision you will make. What matters more is who picks the days, because that shapes how people feel about the policy, and how much the office can actually hold once they pick them.

This guide is written for HR and workplace teams at mid-sized companies, along with the IT people who usually end up owning the tooling.

What does return to office mean in 2026?

Return to office is not one model, and most arguments about whether it works come from people comparing different ones.

Return to office is not one model, and most arguments about whether it works come from people comparing different ones.

Full mandate. A fixed number of days for everyone, usually four or five. Works when the work itself depends on being in one place. Breaks when roles differ in how much collaboration they need and the policy ignores that.

Structured hybrid. The company sets a minimum, teams pick the days. Works when teams depend on each other and need to overlap. Breaks when nobody coordinates, so flexible quietly becomes random.

Role-based presence. Requirements differ by function. Works when you can explain the differences. Breaks when the logic starts to look like a pecking order.

Fully flexible with guidelines. Expectations are shared, attendance is not tracked. Works when teams are small and already coordinate well. Breaks when teams outgrow informal coordination.

Structured hybrid is now the default. JLL recorded its largest single-year change in three years of benchmark data: employees in the office three to four days a week rose to 55%, up 19 percentage points, and fully remote work dropped to 10% from 18%.

What separates these models in practice is not the number of days but how much coordination sits behind them. A three-day requirement where nobody knows who else is coming makes for a worse office day than a two-day model where a team arrives together, and people work this out within about two weeks.

Do return-to-office mandates work?

Mandates move attendance and not much else.

The attendance part is well documented. JLL found that 64% of organizations that brought in a mandate saw utilization go up. But there is more to the story that numbers on the dashboard cannot tell.

Strict mandates are unpopular. A study published in Nature followed 1,612 employees at Trip.com, half of whom worked from home two days a week while the other half stayed in the office full time. Over six months, the group with two home days was about a third less likely to quit, and their managers rated their work no differently. The flexibility kept people without costing output.

The most common reason for a mandate is innovation, and the assumption that good ideas need people in the same room. Fraunhofer IAO tested this in 2026 with more than 2,000 people and found no straight line between office attendance and innovation. Among employees who came in often, some were highly innovative and some were not. What separated them was the environment in the office itself. The innovative ones rated their space as genuinely suited to creative work and to running into people from other departments. The others came in just as much and found nothing there worth the trip.

It is not enough to bring people into the office on the same day. If there is no room for a messy whiteboard session and no reason for two departments to cross paths, you have bought attendance and nothing else.

The only thing a mandate does reliably is get people into the building. Some people will quit over it. And the people who stay only benefit if the office is well-suited for the work they do.

Why the European picture looks different

Reading the headlines, you would think everyone is being called back. In Germany that has not happened.

The ifo Institute tracks this every month. In February 2026, 24.3% of employees worked at least part of the week from home. That figure has barely moved since 2022. It peaked at 32.3% in early 2021 and dropped to 23.4% in the summer of 2024. The recalls that make the news are individual companies rather than a shift across the economy.

The split by industry says more than the average. In IT services, 76.4% work from home at least part of the week. In consulting it is 67.6%. In manufacturing it is 15.4%. If you work in HR at a software company, the people around you are not going to behave like the national number.

European offices are also not filling up the way offices elsewhere are. JLL found that EMEA is the only region where utilization has fallen since 2023, from 58% to 55%. Every other region went up.

That has consequences for space. Hines points out that because people cluster in the office midweek, companies cannot give up as much space as their average occupancy suggests. You still lease the building for all five days. London has cut floorspace per worker by 13.8% since 2019, while Berlin and Amsterdam cut less, because hybrid was already normal there before 2020.

So the American playbook does not transfer. European companies are starting from a different place, and a policy built for a workforce that never left the office will not fit one that never fully came back.

How many days should employees be in the office?

Around two, if you want a number. But picking the number first is how most policies go wrong.

The best evidence comes from a study by Fraunhofer IAO and the Techniker Krankenkasse, which tracked two years of actual productivity data from around 11,000 employees. Working from home turned out to be about 20% more productive for tasks people do on their own. Overall productivity was a different story. It rose with home office up to a point and then started falling once people spent more than about 60% of their time away from the office, which works out to roughly two office days a week.

The reason is that the exchange that happens in the office has no substitute. People stop picking things up from each other, and then they are missing information they need to do the work at home. It gets worse as it goes: the fewer people who come in, the less there is to be gained by coming in.

That figure is an average from one organization, so do not treat it as a target. What it does tell you is that the tipping point exists and sits lower than most mandates assume.

The more useful way to arrive at a number is backwards. Start with what your teams actually need to do together: the meetings that go badly on video, the onboarding that only works sitting next to someone, the projects where two departments have to solve something in the same room. Add up how much of that there is, then check what your office can hold on the days it would happen.

Announce the number first and you spend the next year defending a figure you picked for no reason.

Who should decide which days employees come in?

The teams themselves, and this is the part most companies get wrong.

Gallup asked employees who sets their hybrid schedule and how they feel about it. Where the team decides together, 90% call the policy fair and 55% say it helps them work with each other. Where leadership decides, 73% call it fair and 41% say it helps collaboration. Only 11% of employees work under a team-set schedule. Nearly a third work under one handed down from the top.

The best-performing option is the least used one.

Letting people choose individually does not solve it either. That scores well on fairness, at 88%, and poorly on collaboration, at 45%. Everyone picks the day that suits them, and the office fills with people who have nothing to do with each other. You end up with the commute and none of the benefit.

The reason team-set schedules work is that they are the only level where anyone knows what the overlap is for. A CEO cannot know that the design team needs Wednesdays because that is when the sprint review runs. The design team knows.

This is where anchor days come in: fixed days each week when a team is in the office together. Leadership sets the frame, so the minimum number of days, the core hours, the fairness rules. Teams pick which days inside it. That way you get the consistency an organization needs and the coordination a team needs, without one overriding the other.

Yoffix supports this directly. Hybrid policies can be configured at company, department, or team level, with anchor days set per team, so the company minimum and the team's choice live in the same system rather than in a policy PDF nobody reads and a chat message nobody remembers. Teams see who is coming in on which day before they commit to their own.

What makes an office day worth the commute?

Demand is not spread across the week. It piles up in the middle of it.

HubStar tracked occupancy across 173 buildings in 13 countries and found Tuesday is the busiest day almost everywhere, at 58.6% average occupancy against 34.5% on Friday. In London, Tuesdays reached 75%. Your capacity problem and your empty office problem are the same building on different mornings.

That matters because of what people come in for. Gallup asked hybrid employees what makes a day in the office worthwhile: 55% said building relationships with colleagues, 44% said talking to their manager face to face, 43% said working on things together. Nobody comes in for a desk. They come in for the people at the other desks, which is why an office at 35% capacity feels like a wasted trip even when there is plenty of room.

The space itself is often set up for the wrong work. JLL found companies cutting private offices while adding phone booths, up 41%, focus rooms, up 30%, and small meeting rooms, up 29%. That is the shape of the correction, and most offices are still mid-way through it. People come in to collaborate and then spend the day taking video calls at an open desk, which is worse than doing it at home.

Two things follow. The office needs enough capacity to absorb the peak days without people hunting for somewhere to sit. And it needs the right kinds of space for the work that brought people in.

Yoffix is built for this. Anchor days are set per team, so you can stagger them across the week and see the effect on peak load before you commit. Team zones can be assigned to different teams on different days, which means the same area serves marketing on Monday and Tuesday and IT on Wednesday and Thursday instead of sitting half empty for one of them. People see who is coming in before they decide, book a desk near their team rather than wherever is free, and reserve a room for the conversation they came to have. Where the commute is the friction, parking can be booked the same way.

Can an employer require office attendance in Germany?

Yes, but not on your own terms if there is a Betriebsrat.

The employer decides whether mobile work exists at all and how much of it there is. That part is not shared. What is shared is how it works in practice, which is where most return-to-office plans actually live. Under § 87 Abs. 1 Nr. 14 BetrVG the Betriebsrat co-determines the arrangement of mobile work, and that covers fixed presence days, which days of the week they fall on, and how requests get approved.

The LAG München made this concrete in August 2023 when it stopped an employer from ordering four presence days a month, because doing so departed from an existing Betriebsvereinbarung. There is a second constraint worth knowing: moving someone's place of work can count as a Versetzung under § 99 BetrVG, which needs the Betriebsrat's consent separately.

In practice this means the fastest route in Germany is a negotiated one, and the structure that gets negotiated is usually team-level rather than company-wide. Which is the model the research points to anyway. The legal constraint pushes you toward the better design.

Start that conversation at the beginning of the process rather than presenting a finished policy for approval. And bring numbers to it. Yoffix reporting can be anonymized, with configurable visibility and retention rules, so occupancy data can inform the discussion instead of becoming the subject of it.

How do you measure whether a return-to-office policy is working?

Attendance tells you people showed up. It does not tell you the day was worth it.

Four things are worth measuring instead.

Team overlap. How often a team is in the office at the same time, rather than how often its members are in the office at all. This is the closest thing to a direct measure of what people say they come in for. A team hitting three days a week individually but never overlapping is a policy that has failed while looking fine on paper.

Anchor day adherence. How reliably people make the days their team agreed on. If adherence is low, the days are wrong or something else is in the way, and either is fixable. Overall attendance can stay flat while this falls apart underneath.

Peak day headroom. How much space is left at the busiest hour of the busiest day. Once you are running out of desks or rooms at the peak, people stop trusting that they will have somewhere to work, and they stop coming.

Utilization against a target you actually believe in. JLL notes that the gap between real and target utilization narrowed partly because companies lowered their targets, which is a legitimate way to close it. An office built for a workforce that no longer exists is not underused. It is oversized.

Alongside these, ask people. Fraunhofer's productivity research found the useful early signals are things like knowledge transfer, mutual support, and whether teams still feel like teams, all of which shift before output does. Regular short surveys catch that while you can still act on it.

Yoffix reports on overlap, adherence, and utilization by floor, zone, team, and day, with weekly summaries for team leads. Absence data syncs from Personio, HiBob, BambooHR, HRworks, and SAP SuccessFactors, so the occupancy figures reflect who was actually available rather than counting people on holiday as no-shows.

How do you build a return-to-office plan?

Working through it in order:

Find out what your teams actually need each other for.

Not a survey asking whether people like the office. A conversation with team leads about which work goes badly on video, where onboarding stalls, and which projects need two departments in a room.

Check what your building can hold.

Desks and rooms at the busiest hour of the busiest day, against how many people a given policy would send in. If the numbers do not work, you know before you announce anything rather than after.

Set the frame, then hand over the days.

Leadership decides the minimum and the fairness rules. Teams pick their anchor days inside that, and you stagger them across the week rather than letting everyone land on Tuesday.

In Germany, talk to the Betriebsrat at the start.

The arrangement is co-determined, so it will be negotiated either way. Doing it early costs less than redoing a finished policy.

Fix what people find when they arrive.

Somewhere to think, somewhere to talk, and a reason for teams who never speak to cross paths. Presence without this buys attendance and nothing more.

Then watch overlap and adherence rather than headcount, and expect to adjust. The teams that got their days wrong will tell you within a month.

Closing

The companies where this works are not the ones with the strictest policy or the loosest one. They are the ones where the office can hold the people it asks for, and where the people affected by the schedule had a say in setting it.

Everything else is arithmetic about days.

If you’re currently planning or struggling with RTO, we’ve put together a practical return-to-office guide and checklist that walks you through proven strategies and pitfalls to avoid.