Occupied versus paid-for space - what unused space really costs

The biggest cost decision in an existing building portfolio is not the new build. It is the space you already have - how much of it you need, how much to keep, convert or give up.

It is surprisingly often made on two foundations that were never meant for it: access data, which says who entered the building, and surveys, which say how people rate their own use. Neither says which space was actually used.

This article is about that gap between the space you pay for and the space that is occupied. It builds on Measuring occupancy, which explains how a headcount comes about and why it drifts, and on Comparing sites, which covered fair comparison between buildings.

A booking is intent, presence is measurement

The room is booked but empty. The floor is officially full but half occupied. The booking system reports Friday utilisation nobody has ever seen in the building.

That is not a fault of the booking system. It does what it was built for: it manages intentions. A booking says someone meant to come. It does not say they came, or how long they stayed.

Access data only half closes the gap. A badge at the main entrance shows someone was in the building - not whether they sat on the third floor all day or spent two hours in a ground-floor meeting room and then went home to work. A statement about space needs a measurement on the space.

Measurement happens per zone: a floor, a wing, a canteen, a group of meeting rooms. Zones can be named the way the building is organised - by floor, department or room group - so that facilities, HR and security read from the same source.

The sum hiding in the existing stock

How large the lever can be is shown by a simple thought. Imagine an area of a building that is barely used from Thursday afternoon onwards. It is still heated, ventilated, lit and cleaned until Monday morning - for people who are not there.

The logic behind it is unspectacular: the decision to close an area at certain times is not new. What is new is that it can rest on a measured usage curve instead of the impression that Thursdays get quieter.

That is the normal case in existing stock: the saving is there, it just goes unfound because nobody measures the use.

Peak against mean: how much reserve a space needs

Whoever sizes space faces a contradiction. Plan for the average and there is not enough room on Tuesday morning. Plan for the peak and the space stands half empty four days out of five.

The answer lies in a number that almost never appears in space reports: the gap between peak and mean, per zone and per weekday. A worked example for an office zone with 120 places:

Three conclusions follow from it, and from no booking report: the zone never fills, even in the busiest hour of the week. Friday is not a weaker Tuesday but a different day - and a candidate for asking whether the zone needs to open on Fridays at all. And the reserve you need for Tuesday is smaller than the gap between capacity and mean would suggest.

How the peak factor per zone is calculated is covered in Measuring dwell time. For space planning two views are added: the weekday pattern per zone, because office, canteen and meeting rooms have different weeks, and a twelve-month seasonal map, because a measurement in October does not describe the same space as one in July.

So every space decision needs a series of measurements that includes at least a holiday period. Counts are kept for 730 days for exactly that reason.

The floor nobody misses

A usage curve only becomes a decision in a second step, and that step is often skipped. Closing a zone at certain times does not close its users. They go elsewhere, and the question is whether elsewhere can take them.

So a space decision that holds runs in four steps:

1. Measure until the curve holds. At least four weeks, better two months, and across a holiday period if possible. Before that, every week looks like an exception.
2. Identify the candidate. The zone and time window with the lowest peak against capacity - not the lowest mean. A zone with a low mean and a high peak on Thursday afternoon is no candidate for Thursday.
3. Close it on trial and watch the other zones. What matters is not that the closed zone stays empty but that the peak in the receiving zones stays below their capacity. Occupancy per zone shows that within the same week.
4. Only then switch permanently - and let the building services follow the new pattern instead of maintaining a timetable by hand.

The third step separates a saving from a displacement. A closure that merely shifts what used to be spread out saves energy in one zone and makes three others less pleasant to be in.

The canteen at 12:10

At 12:10 there are 80 people at the counter, at 13:15 the canteen is empty. Everybody knows the pattern, nobody can steer it, because occupancy is not visible where the decision is made - at the desk, before getting up.

The canteen is therefore the textbook case of occupancy being used twice. Once by those who run the space: occupancy per area - counter, seating, terrace - as a percentage, with a site status from safe to full and a notice to operations when an area reaches capacity. And from the weekday patterns, planning for counter and cleaning staff.

Once by those who use the space: a display on the intranet or at the entrance showing whether there is room right now. How such a display has to be built to hang in a public space is covered in Measuring occupancy. The effect is simple: people spread themselves out, and the peak flattens without anyone writing an instruction.

One detail decides trust in that display: a zone without data is never shown as “empty”. Anyone shown an empty canteen at 12:05 who walks into a full one will not look a second time.

Building services that follow occupancy

The air conditioning runs for an empty floor, the lights burn in an empty corridor. Occupancy is known - the building services just never hear about it.

The connection is a call to the building management system: if a zone's occupancy crosses a threshold upwards or downwards within a given time window, a message goes to an address the building services accept. Every call is logged, and a rule can be checked beforehand in a test run.

One point matters more than any threshold, and it belongs in every integration meeting: the occupancy of a zone without current data is passed on as unknown, not as zero people - technically as null, not 0.

The difference sounds like a programmer's detail and is a safety issue. A building system that receives “0 people” consistently turns off lights and ventilation. If it also receives “0” when only the sensor has failed, an occupied floor sits in the dark. The receiving side must be able to tell “unknown” from “empty” - and that only works if it receives them differently.

Cleaning by passages, not by the clock

Cleaning to a fixed schedule is either too much or too little. Complaints come from the heavily used facilities, costs from all of them - and the contractor bills by plan, not by need.

How a cumulative usage counter works is covered in Passenger counting in public transport, where it drives escalators and toilets in stations. In an office building the default is a message after 1,000 passages with a 60-minute lock-out; the counter restarts after sending.

The real lever is in the contract. Because every trigger is in the history with time and delivery status, a cleaning service can be agreed by triggers instead of a flat fee - with proof of use per facility that both sides can read.

Safety in the evening

In workshops, common rooms and gyms late in the evening, the risk is not the full room but the one person there alone. Cameras are mostly not allowed there, and a count without a notification changes nothing.

The threshold in the “too few” direction is described in Passenger counting in public transport using the example of a waiting area. In a building a typical rule reads: gym between 21:00 and 06:00 local time, notify security as soon as only one person is left. Security is then called specifically rather than by patrol schedule - and the duty of care can be evidenced without any personal data arising anywhere.

Universities and halls of residence: the same questions, a different calendar

On a campus a second rhythm runs over the week: teaching term, exam period, semester break. A library that needs every seat in June stands half empty in August, and a lecture hall that overflows on Tuesdays is often locked on Fridays. That is what the twelve-month seasonal map is for - on a campus it is not an extra but the main view.

The questions for a university's administration are the same as in an office building: align energy and maintenance with the usage profile, and justify investment in space - especially once study and working habits have shifted.

For students a different figure matters: whether a study seat is still free at 2 pm in exam week. A public display at the entrance or on the intranet answers that before anyone crosses the campus. How it needs to be built is covered under the canteen above and in Measuring occupancy.

Compare buildings by use, not by floor area

The portfolio report compares buildings by area and cost. Hardly anyone compares them by use, because counting definitions differ from building to building or do not exist at all.

With uniform measurement the report gains a usage column made of measurements. How a fair comparison between buildings is calculated - which denominator, which coverage, when to show no number - is covered in Comparing sites. For buildings the same applies, with one difference: buildings with different operating hours belong in different groups, otherwise the comparison measures the calendar instead of use.

What belongs in the tender

1. Is use measured per space - or estimated from access and bookings?
2. Can zones be named after the building's organisation, and does every department read from the same source?
3. Are there peak and mean per zone and weekday, and a seasonal view over twelve months?
4. How is a zone without current data passed to building services - as empty or as unknown?
5. Can a cleaning service be evidenced by triggers?
6. How long are counts kept, and is that enough for a comparison over a year?

How we do it

The ANALYSIT Counting System measures occupancy per zone live and reports peak and mean per weekday, with peak factor, weekday pattern, a twelve-month seasonal map and dwell time per zone. Zones can be named as floor, department or room group; counts are kept for 730 days and can be exported as CSV and XLSX.

For canteens and shared spaces there is utilisation per area with site status, a public display without login and a notice to operations when capacity is reached. The building management system receives calls when occupancy crosses a threshold upwards or downwards in the chosen time window, logged and testable beforehand; a zone without current data is passed on as unknown, not as zero people.

Added to that are the cumulative usage counter for cleaning by passages, the threshold towards the minimum for safety in the evening, and comparison across buildings with median, percentiles and coverage.

If you have a space decision coming up and today's basis is bookings and surveys, talk to us. Four weeks of measurement usually already answer which day is the weakest.