The pitch for occupancy sensors is always the same slide: a heat map of a floor, some warm colours near the windows, and a number claiming a percentage of desks are empty. It is a compelling slide. It is also the point at which most sensor projects stop being useful, because nobody in the room has decided what they would do differently if the number were true.

Occupancy sensors in an office building are worth installing when a specific decision is waiting on a specific measurement. Installed without that, they produce a dashboard somebody opens twice.

Three questions worth measuring

Only three come up often enough to justify the hardware in a commercial building, and they are not the same question.

How much of this space is used, and when? Landlord-side, this drives cleaning schedules, HVAC run times and amenity sizing. Tenant-side it drives their space decisions, which eventually becomes your renewal conversation.

Is this room free right now? A live question with an immediate answer, which is a booking problem instead of an analytics one.

Did the booking happen? The gap between reservations and attendance is the most actionable number in the amenity programme, and the one nobody has.

What the data changes

Cleaning is the clearest case. A floor cleaned five nights a week regardless of whether forty people or four were on it is a contract written for a building that no longer exists. Occupancy data converts that from a negotiation about rates into a conversation about scope, which is a much better conversation to have.

Conditioning is the second. Running full HVAC for a floor at fifteen per cent occupancy on a Friday is expensive in a way that shows up in the operating statement without ever being attributed to anything.

Amenity sizing is the third, and the one that changes capital decisions. A conference centre that turns away requests at eleven and sits empty at three does not need to be larger. It needs different booking rules, and knowing that costs a lot less than building another room.

A sensor is only worth its install cost if somebody has already agreed what they will change when the number arrives. Otherwise it is a very precise way of confirming a suspicion nobody acts on.

Ghost bookings, the most useful number

Every building with bookable rooms has them: the recurring meeting that stopped happening in March and still holds the best room every Tuesday, the reservation made in case, the block held by someone who left.

Pairing a reservation with an occupancy signal turns this from a suspicion into a rule you can apply. If a room is booked and nobody enters within ten minutes, the booking releases. That single behaviour recovers more usable capacity in most buildings than any amount of additional space, and it does not need a heat map to work.

Badge data first, sensors second

Most buildings already hold a rough occupancy record and never look at it. Access control logs every entry, by floor, with a timestamp, going back as far as retention allows.

It is a blunt instrument. It counts entries instead of presence, misses people who tailgate, and says nothing about which room anyone sat in. What it does give you, free, this week, is the shape of the week: which days are busy, when the peak arrives, whether Friday looks like Tuesday, and how that has changed over two years.

For most of the decisions on the list above, that shape is enough to act on. Buy sensors when badge data has taken you as far as it can and a specific question remains, not before, because the second dataset is much easier to justify once the first one has already changed something.

The privacy question, answered honestly

Tenants will ask, and the answer has to be better than "it is anonymous".

The technologies differ in what they can see, and that difference matters. Passive infrared and thermal sensors count bodies and cannot identify anyone. Camera-based counters can, in principle, whatever the vendor's processing claims. Badge and Wi-Fi data are identifiable by construction, because they are tied to a person's credential or device.

The defensible position for a landlord is to measure spaces instead of people: counts and durations at the room or zone level, no individual tracking, no desk-level attribution of a named person, and a plain-language note in the tenant handbook saying exactly that. Any programme that would embarrass you if described accurately to a tenant should be redesigned instead of explained better.

Where sensors are usually oversold

Desk-level occupancy in a landlord-run building is the clearest example. It answers a tenant's question with the landlord's money, generates the most privacy friction, and requires the densest and most expensive deployment.

Predictive maintenance from occupancy is another. Vendors will connect footfall to equipment wear, and the connection is real, but the interval you would derive from it is a refinement most buildings cannot yet act on because their preventive schedule is not being followed reliably in the first place.

Start small, in the places where the decision is already waiting

Meeting rooms and shared amenity spaces first: cheap to instrument, immediate booking value, no privacy argument, and the data has an obvious owner. Then common areas and entrances, which give you building-level patterns for cleaning and conditioning without touching tenant space.

Tenant floors last, and only in partnership with the tenant, funded in a way that reflects who gets the benefit. A landlord instrumenting a tenant's floor without them is buying a difficult conversation for a number they cannot act on alone.

What it costs to run, not to buy

The hardware quote is the easy part. Three ongoing costs turn up later.

Batteries, which sounds trivial until a five-year battery life across two hundred sensors becomes a recurring maintenance route with its own schedule. Network coverage in the places sensors sit, which is usually the corridor ceiling and the basement where coverage is worst. And somebody's time to read the output, which is the cost most often left out entirely and the one that determines whether the project produces anything.

What to ask a sensor vendor

What exactly is captured, and at what granularity? Counts per zone is a different product from identities per desk, whatever the marketing says.
Where does the data go and who else can see it? Their cloud, your tenancy, both. Ask what happens to it if you cancel.
What is the battery life in this deployment? Not the datasheet number under ideal conditions.
Does it push to our systems, or only to their dashboard? A dashboard nobody opens is the default failure mode of this category.
What happens when a sensor stops reporting? Silent failure produces confident graphs built on missing data, which is worse than no graph.

Where BuildingHub fits

BuildingHub integrates with Density and Butlr, so occupancy signal arrives next to the things it should change instead of in a separate dashboard: bookings in Amenity Hub can be released when a room goes unused, and utilisation sits with the reservation records instead of beside them. The full integration list is on the integrations page. Related: what amenity software should do.

Occupancy sensors: what the data changes