Nine checks to run before your occupancy number goes in front of leadership

There is a particular kind of meeting that most Heads of Workplace have sat in.

You have brought a number. Utilisation, occupancy, however your organisation phrases it. Someone senior looks at it, pauses, and asks where it came from. And the honest answer — the one you do not give in the room — is that it came from an export, which came from a system, which measures something slightly different from the other system that also produces a number.

This is not a competence problem. It is a structural one. Most workplace teams are not short of building data; they are short of a settled way to turn it into a figure that holds up. As Nate Colle at Metrikus puts it: "The challenge today isn't collecting more data. Most organisations already have plenty of it. The difficult part is making sense of that data altogether."

So before your next number goes in front of anyone, here are nine checks worth running against it. They take about twenty minutes. Most teams find at least two that change the answer.

1. Where did the number actually come from?

Badge data, booking data and sensor data measure three different things: who entered the building, what people intended to do, and what actually got used.

They are not interchangeable, and the gap between them is not small. A badge figure counts the person who came in for a 20-minute coffee exactly the same as the person who sat at a desk for eight hours. If your utilisation number is really an attendance number, it will read high — and the decision it supports will be sized for people who were barely there.

2. What's the denominator?

"68% occupied" against what? Desks, seats, square metres, or headcount?
A utilisation figure without its divisor named beside it is not a metric; it is a guess. Two teams in the same organisation can produce wildly different percentages from identical sensor data purely by dividing by different things. Write the denominator next to the number every single time.

3. Peak or average?

An average hides the day that hurts.

If you are running 42% on average but 91% on a Tuesday, you do not have an underused building. You have a badly distributed one — and the fix for that is completely different. Averages argue for giving space back. Peaks argue for redistributing it. Reporting only the average makes the first argument for you.

4. Which days are in it?

A five-day average across a three-day-attendance workforce will almost always look like underuse.

If your organisation has settled into Tuesday–Thursday, a Monday-to-Friday denominator builds two guaranteed-quiet days into every figure you produce. Report attendance-day occupancy separately. Otherwise you will find yourself arguing for a decision the data does not actually support.

5. How many weeks?

One quiet week is not a trend. One busy Tuesday is not consistent utilisation.

Four weeks is the minimum you would want to draw a conclusion from, and a full quarter is what holds up when someone challenges it. Seasonality, project cycles, school holidays and a single large offsite can all move a month. The shorter your window, the more confidently you will be wrong.

6. Put your booking system next to your sensor data - desks and rooms. 

You will find desks held all day by someone who never came in, and rooms that show as fully committed and sit physically empty.

This is the most common surprise when organisations connect two data sources that previously sat apart. Desk bookings matter most, because they are usually what the occupancy figure itself is built from: every unreleased desk gets counted as a person.

Halving a ghost booking rate is usually the cheapest capacity a workplace team will ever recover. It costs a behaviour change rather than a building.

7. Does every floor behave the same?

A building-level average with one floor at 85% and three at 30% comes out around 44% — a number that describes none of the four floors.

Building averages are comfortable because they are simple. They are also where the useful signal goes to die. Usually there is a reason one floor behaves differently: a team anchors there, or the layout suits a particular kind of work. That reason is the insight. The average is not.

8. Does the number match what people say?

If the data says 40% and employees say it is rammed, the data is not necessarily wrong. It is answering a different question.

Often, the complaint is about meeting rooms and the measurement is about desks. The mismatch between what the numbers report and what the building feels like is not a reason to distrust the data — it is a prompt to check whether you are measuring the thing people are actually experiencing.

9. Could you rebuild it next month, the same way?

If it took three days of spreadsheet work, you do not have a metric. You have a one-off report that will probably disagree with the next one.

This is the check that matters most over time. A number you cannot reproduce cannot be tracked, and a number that cannot be tracked cannot support a decision anyone revisits. If reproducing it depends on one person remembering what they did, that is a risk rather than a reporting process.

What to do with the answers

If you cleared all nine, your number is in better shape than most — write the methodology down so the next one matches.

If two or three gave you pause, that is the normal result, and it is a more useful outcome than a clean pass. The point of the exercise is not a perfect figure. It is knowing which parts of your number you would defend and which you would caveat, before anyone else asks.

The three that cause the most trouble are the first, the fifth and the seventh: where the number came from, how long a window it covers, and whether a building-level average is hiding floors that behave nothing like it.

Want to put your occupancy data to the test?

If these checks have raised questions about the accuracy of your occupancy figures, it may be time to take a closer look at what your workplace data is telling you.

The Metrikus Occupancy Audit helps you understand how effectively your space is being used, identify gaps in your current data and uncover opportunities to make better-informed workplace decisions.

Get your free Occupancy Audit and find out what your workplace data is really telling you.


Produced by Metrikus with Nate Colle. October 2026.

×

Book a demo

Book a demo
BLOG_POST