Data-Driven Workplaces: Measuring Real Office Occupancy

Every quarter a version of the same conversation happens between a CFO and a head of workplace. The CFO has a lease renewal on the desk and a number in mind. The head of workplace has a feeling that Tuesdays are packed and Fridays are dead. Neither has a number they would defend in front of the board.

Office space is usually the second-largest cost after payroll, and it is the one most companies still manage on anecdote. This article is about replacing the anecdote: what "occupancy" actually means, why the numbers most companies have are wrong, and how out-of-the-box reporting from a booking system gives leadership a figure they can act on.

Three numbers that are not the same thing

The first mistake is treating booking data as occupancy data. Three distinct measurements matter, and they diverge in exactly the ways that lead to bad decisions.

Bookings are intent. A desk booked for Tuesday means someone planned to be there when they made the booking. Bookings overstate demand by the no-show rate, which in offices without check-in is commonly between 15% and 30% on a given day.

Occupancy is presence. A desk is occupied when someone is at it. This is the number the lease decision depends on, and it is the hardest to measure without either sensors or a check-in step.

Utilisation is occupancy over capacity, across time. Seventy checked-in desks out of one hundred on Tuesday is 70% utilisation for that day. Averaged over a month, by weekday, by floor and by team, it is the shape of how the office is actually used.

A workplace team that reports bookings as occupancy will renew a lease it could have shrunk. A team that reports monthly average utilisation without the weekday breakdown will shrink an office that is genuinely full two days a week.

Why the data you already have is not enough

Most companies have some occupancy signal already. Each one is misleading on its own.

• Badge swipes count people entering the building, not where they sit or for how long. A person who badges in, attends a meeting and leaves counts the same as one at a desk for nine hours. Multi-tenant buildings and tailgating make it worse.
• Wi-Fi association counts devices, double-counts phones and laptops, and knows nothing about desks.
• Manual counts are accurate for the hour they were taken and expensive to repeat.
• Desk sensors are accurate and expensive, and they answer "is there a warm body here" without knowing who booked, who did not show up, or which team.

Bookings with check-in sit in a useful middle: cheap, per desk, per person, per day, and tied to intent. The check-in is what turns intent into presence. We explained the mechanism in the article on auto-release; here we look at what the resulting data can do.

The reports leadership actually asks for

Desk & Park produces its reports from the same objects people book: every desk, parking spot and meeting room on the map records bookings, check-ins, no-shows and releases. From that, the board report answers the questions that come up in the lease conversation, without exporting anything.

Peak-day utilisation

Not the monthly average. The question that sizes an office is "how full are we on our busiest day, and how often does that happen?" The weekly utilisation view shows each weekday side by side, so a floor that averages 45% but hits 85% every Tuesday and Wednesday is visible as what it is: a two-day-a-week bottleneck.

No-show rate

The share of bookings never checked in, by weekday, floor and space type. This is both a data-quality metric (how much your booking data overstates demand) and a behaviour metric (which teams book and do not come). It typically falls sharply once auto-release is on, and the report shows the drop.

Parking occupancy

Reported next to desk occupancy for the same day, because a car park that fills at 8:40 while the floor sits at 60% is a parking problem, not a space problem. Rightsizing the office without this number is how companies end up with fewer desks and the same lobby arguments. See why parking is the harder problem.

Utilisation by team and department

Which teams come in, on which days, and whether they cluster. This turns a vague "sales likes Tuesdays" into a neighbourhood plan: put teams that overlap on the same floor and teams that alternate on the same desks.

Meeting-room usage

Rooms booked versus rooms checked in at the kiosk, and the average slot length. Meeting rooms are the space type most often over-provisioned, and the one where a booked-but-empty room is most visible to everyone walking past.

From numbers to decisions

Data is only useful if it changes a decision. These are the decisions the reports are built to support.

Rightsizing the lease. With peak-day utilisation by floor over three months, the conversation with the CFO becomes arithmetic. If two of four floors never exceed 55% on their busiest day, the office can lose a floor without anyone losing a desk on Tuesday, provided the desk ratio and the booking rules are adjusted together.

Setting the desk ratio. The ratio of people to desks is a policy that should follow the measured peak, not the other way round. A company at 70% peak utilisation with a 1:1 ratio can move to 1.4:1 and give the space back; a company at 95% with 1.5:1 needs more desks or a spread-out attendance policy.

Choosing which days to anchor. If the data shows two overloaded days and three empty ones, the fix may be an attendance pattern, not more space. The utilisation-by-team view shows which teams could move a day without losing overlap with the teams they work with.

Deciding what to build. More phone booths or more meeting rooms? The room report answers with actual usage, not the number of complaints.

Making the data trustworthy

Reports are only as good as the check-in behind them. Three practices keep the numbers honest:

1. Check-in on the object. QR stickers on desks and spots, scanned where the person sits. A check-in from the sofa at home is not presence, and a system that allows it is measuring intent twice.
2. Auto-release on. No-shows must be released and recorded as no-shows, or they silently inflate both bookings and, if the report counts bookings, occupancy.
3. A stable time window. Report on the same window every month (for example, Monday to Friday, working hours per location) so the numbers compare. Desk & Park applies each location's working days, holidays and time zone, so a bank holiday in one office does not read as a collapse in attendance.

Out of the box, not out of a data warehouse

Many organisations plan to "build the dashboard later" in a BI tool. Later rarely comes, because the join between the badge system, the booking tool and the HR directory is nobody's job. Reporting that comes with the booking system, over the same objects people book, on the same screen the admin already uses, is the version that gets looked at every month.

Board members and managers in Desk & Park have their own reporting view, narrowed to what they are allowed to see: a manager gets their teams, a board member gets the organisation. The C-level question, "how full is the office really," gets an answer with a date range and a breakdown instead of a feeling.

If your occupancy figure today is a booking count or a badge count, the first step is smaller than a data project: put the office on a map, turn on check-in, and let the reports run for a month. The lease conversation will be different next quarter.