Hybrid Office Space Utilization: Why Average Occupancy Leads to Poor Office Sizing

In brief

• CBRE’s 2026 benchmarking data puts the global office occupancy rate at 111%, meaning more employees are allocated to buildings than there are physical seats, yet average daily utilization sits at only 53% because attendance concentrates on two or three midweek days rather than spreading across five.
• Roughly seven in ten organizations name Tuesday as their single highest-attendance day, according to CBRE, while Leesman’s larger employee-level panel finds a flatter midweek pattern with Thursday narrowly ahead; both datasets agree that Monday and Friday attendance remains marginal.
• A 2025 peer-reviewed study from Aalto University found that ignoring “passive occupancy” (desks left occupied by belongings after a person has stepped away) can understate true desk unavailability by close to double, a gap that directly distorts right-sizing decisions.
• In Italy, CBRE’s Head of Leasing, Fabio Mantegazza, describes a market consolidating around four days in the office and one remote day, a heavier in-office pattern than the global three-day standard, which changes how peak-day crowding plays out locally.
• Sizing an office to its weekly average attendance produces a floor that is uncomfortably full on peak days and visibly empty on the rest, an outcome that wastes costs and damages the in-office experience CRE teams are trying to protect.

Short answer

Office occupancy has become a story of two numbers, both pointing away from the weekly average. Globally, CBRE’s 2026 benchmarking puts allocated occupancy at 111% of physical seats, while average daily utilization is only 53%; the gap is attendance concentrated into two or three midweek days rather than spread evenly across five. Sizing decisions built on the average of five days understate how full an office gets on its peak day and overstate how full it is on the rest. The organizations narrowing that gap size around peak-day percentiles, measure actual rather than assigned desk use, and treat Monday and Friday capacity as intentionally smaller than Tuesday through Thursday.

Definition

Utilization rate, occupancy rate, and peak utilization measure three different things, and confusing them is where most sizing errors start. Occupancy rate compares the number of people assigned to a space against the number of seats available; CBRE’s 111% figure means organizations have allocated more employees to a building than it has desks, which only works because not everyone comes in on the same day. Utilization rate measures how many of the available seats are actually occupied on an average day; a separate, lower figure (53% globally in CBRE’s 2026 data). Peak utilization measures occupancy on the single busiest day, the number that determines whether people can actually find a desk, a room or a quiet corner when they show up (80% globally, and also 80% in EMEA). A workspace can carry a low average utilization rate and still feel overcrowded, because the average hides the peak.

Why this matters differently for landlords, CRE teams and advisors

For the corporate real estate manager: average-based sizing (headcount divided by an assumed attendance rate) systematically under-builds for peak days and over-builds for the rest of the week. The fix is sizing to a peak percentile of measured attendance, typically the 85th or 90th percentile day, rather than to the mean, and separating individual desks from collaboration, support and amenity space, which now needs to flex independently of desk count.

For the institutional landlord: multi-tenant buildings that can pool variable, shared peak demand across several occupier organizations, rather than betting an entire floor plate on one company’s attendance curve, are structurally better positioned for a market where 96% of occupiers now run a formal attendance policy but almost none of them need the office full every day.

For the advisor/capital markets specialist: “how many desks per employee” is no longer a single ratio to quote across clients; it is a function of the client’s targeted attendance policy, its enforcement gap (CBRE finds 70% of employees are in the office less than leadership expects), and whether the client actually measures passive occupancy or only raw sensor presence.

The gap between allocation, use and peak, quantified

CBRE’s Adaptive Spaces 2026 benchmarking, based on client portfolios covering roughly 303 million square feet globally, tracks occupancy rate, utilization rate and peak utilization as three separate lines, and 2025 marks the first year peak utilization (80% globally) exceeded the 65% target CBRE has tracked since the pandemic. Global occupancy allocation reached 111% of physical seats, up from 101% in 2024, while design density tightened to 190 square feet per seat and people density to 158 square feet per person, both down year over year. Utilization has become the most closely tracked occupancy metric among CRE teams, with 83% tracking it in 2025, and 87% of organizations now set an explicit utilization target, nearly half of them aiming for 76% to 85%.

Where that peak actually lands in the week depends on which dataset is doing the measuring, and the two most authoritative ones disagree in an instructive way. CBRE asks organizations to name their single highest-attendance day at a policy level: 73% name Tuesday, 23% Wednesday, only 3% Thursday, with Monday and Friday close to zero. Leesman’s Rhythms Report 2025, built on a panel of more than 100,000 individual employees describing their own habitual office days, finds a flatter distribution instead: Thursday leads at 41%, Tuesday follows at 39%, Wednesday at 37%, Monday at 25% and Friday at 21%, with close to a third of respondents reporting no fixed routine at all. The two figures are not measuring the same thing. CBRE’s is an organizational designation of a single dominant day; Leesman’s is a distribution of individual habitual attendance across a much larger and more varied panel. Read together rather than against each other, they agree on the point that matters for sizing: no single weekday carries anywhere near five-day average attendance, Monday and Friday are consistently the lightest, and the midweek block from Tuesday to Thursday is where real capacity pressure sits, whichever single day within it happens to peak for a given organization.

Italy adds a layer that a globally averaged figure erases. In an interview with Limitless Workspace, Fabio Mantegazza, Head of Leasing Italy at CBRE, describes the Italian market as consolidating around a “4+1” model, four days in the office and one remote, a materially heavier in-office pattern than the “3+ days” standard that 66% of organizations report globally. Guido Ferraresi of DILS, in a separate interview, adds that smaller Italian occupiers below roughly 5,000 square meters are in several cases increasing their footprint rather than shrinking it, driven partly by a more structured four-out-of-five-day return and a need for more collaboration and client-facing space. A sizing model imported unchanged from a market where three-day attendance dominates will misjudge peak-day pressure in a market where four days is closer to the norm.

Why the average hides the real sizing problem

The instinct to size an office by dividing headcount by an assumed attendance rate treats attendance as if it were smooth across the week. It is not, and the smoothing itself is the error. An organization that plans a floor around 53% average utilization will build a space that is comfortable on Monday and Friday and overcrowded on the peak midweek day, precisely the days CBRE’s data shows CRE teams are already worried about, with average peak utilization at 80%, well above the 65% comfort target. The correct planning input is not the average; it is a peak percentile of measured attendance, typically the 85th or 90th percentile day, which absorbs normal week-to-week variation without building permanent capacity for the rare all-hands day.

A second, less visible distortion comes from how “occupied” is measured in the first place. A 2025 peer-reviewed study from Aalto University, published in the Journal of Corporate Real Estate, examined desk-sensor data across ten office sites in Finland, Norway and Sweden and introduced the concept of passive occupancy: a desk that looks empty to a presence sensor because the person has stepped away, but is not actually available because their belongings are still on it. Without accounting for a clean-desk policy time delay, sensor-based utilization data understated true desk unavailability by an average of 67% to 111% depending on the site studied; adding a 120-minute delay window nearly doubled measured utilization on average, an increase of 86.4%. The practical consequence for sizing is direct: an organization that cuts desks based on raw sensor occupancy, without correcting for passive occupancy, risks removing far more capacity than the actual attendance pattern can absorb, and only finds out once the floor feels short of desks on an ordinary Tuesday.

The composition of the floor is shifting alongside the count of desks. CBRE’s companion 2026 design report shows individual “Me” space falling from 56% of the floor in 2021 to 35% in 2025, while collaboration, support and amenity space have all grown; the activity-based model that 47% of respondents expect to define the future office already allocates closer to 40/20/15/25 across those same four categories. A sizing conversation that still starts and ends with “how many desks” only addresses the shrinking part of the floor plate.

What this means for how CRE teams should actually size a floor

Sizing to a peak percentile rather than a weekly average changes the planning conversation before it changes the floor plate. It requires attendance data granular enough to calculate a percentile in the first place, meaning badge, Wi-Fi or sensor data collected over several months, not a single site visit or a policy document. It requires separating the sizing of assigned or bookable desks from that of collaboration, support, and amenity space, since CBRE’s data show these are moving in different directions and at different speeds. It requires correcting raw sensor occupancy for passive occupancy before using the corrected figure to justify a cut, following the Aalto University finding that this correction can nearly double the measured figure. And it requires treating the gap between a targeted policy, a communicated policy, and an enforced policy as a planning input rather than an HR footnote, since CBRE finds that 70% of organizations already report employees attending less than leadership expects, up from 61% in 2024, a gap that will only widen the case for over-building if it is ignored.

Organizations that have not yet measured the gap in their own portfolio can start with a structured self-assessment rather than a full sensor deployment. The Workspace Alignment Index, Limitless Workspace’s diagnostic framework, assesses the fit among cost, space, and actual usage across four dimensions and serves as a practical first step before committing to a resizing decision.

A multi-tenant answer to the same problem

Multi-tenant business centers structured around activity-based principles offer one practical answer to the peak-day sizing problem, because they pool variable demand across several occupier organizations rather than betting an entire floor plate on one company’s attendance curve. Stella Santa Giulia, the LEED Gold-certified business center that Stella33 operates within the Spark Business District in Milan’s Santa Giulia regeneration area, allocates roughly half of its surface to shared common areas, event space, and meeting rooms rather than assigned desks, which lets attendance from several tenant companies flex against a shared capacity buffer instead of each one carrying its own peak-day margin.

The same logic applies to any organization evaluating flexible or serviced space in Milan as part of a hybrid footprint: shared, activity-based capacity absorbs uneven midweek demand more efficiently than a single-tenant floor sized to its own average, and platforms such as UfficiMilano.com list flexible and serviced office options across the city built on comparable multi-tenant models.

Sizing approaches compared

Frequently asked questions