Field notes · Buying & procurement

What do occupancy sensors actually cost? A buyer's guide to the whole bill

Occupancy sensor cost is mostly not the sensor. Hardware, installation, licences, calibration and governance: the full cost structure by sensor type, and the questions that keep a project honest.

A laptop on a desk showing an analytics dashboard with charts, the software layer whose licence model drives most of the real cost of an occupancy sensing project
The sensor is the cheapest thing in this picture. The licence model behind the dashboard is where the money goes.

Nobody publishes real prices in this market, including, mostly, us: our pricing is per zone and scoped per deployment, and this post is not a rate card. What we can do is show you the cost structure, because the pattern we see in procurement is consistent: buyers compare sticker prices on hardware, then discover the actual budget in lines four through nine. Here is the whole bill, and the questions that expose it before you sign anything.

01 · THE NINE LINES OF THE REAL BILLThe nine lines of the real bill

  1. Sensor hardware, per zone. The number on the quote. Varies by an order of magnitude across sensor categories, and it is the least decisive line on the page.
  2. Installation. Ceiling access, mounting height, cabling runs, lift hire for atriums, after-hours work in occupied buildings. Routinely rivals the hardware cost; occasionally exceeds it.
  3. Power and network. PoE is the civilised answer, one cable for both, but it assumes switch capacity and a routed VLAN. Segregating sensors from the corporate network is not optional in a modern security review.
  4. Software licence. Per device, per zone, per square metre or per site, monthly or annual. Over a five-year horizon this line usually beats every other line on the bill. The licence trigger matters more than the licence price: per-device pricing punishes you for wide coverage; per-zone pricing means one bill whether a zone needs one sensor or three.
  5. Calibration and validation. A count nobody has verified is a rumour. Ground-truth auditing on your site, at your densities, is a real cost, and we argue it should be in every pilot, including ours.
  6. Integration engineering. Getting counts into the BMS, the dashboard nobody asked for out, and the systems you already run in. Budget days, not hours, for BMS point mapping in older buildings.
  7. Governance. Privacy impact assessment, legal review, staff consultation. Here architecture moves money: a stored-footage system carries a footage-governance program for its whole life; an anonymous-by-architecture system makes this a short review, which is a genuine cost line saved, not just a comfort.
  8. Maintenance and recalibration. Sensors get bumped, repainted around, refitted around. Zones change when furniture does. Ask what a re-zone costs, because your floor plan will not survive five years unchanged.
  9. Exit. Data export, contract minimums, hardware ownership. The cheapest system to enter is sometimes the most expensive to leave.
Hands sorting paper forms next to a phone calculator, the budgeting work of pricing an occupancy sensing project across hardware, installation, licences and governance
Lines four through nine are where the project budget actually lives. The sensor quote is line one.

02 · COST PROFILES BY SENSOR CATEGORYCost profiles by sensor category

We compared the categories on capability and privacy in do occupancy sensors use cameras?; here is the same taxonomy priced as a structure:

  • PIR and motion sensors. Cheapest hardware by far, near-zero governance. But they detect motion, not people: no counts, no density, no dwell. Cheap per unit, expensive per answer, because "is someone in there" is the only question they answer.
  • Doorway beam and thermal counters. Moderate hardware, low governance. Count crossings at a threshold, so zone occupancy is inferred from ins minus outs, and the inference drifts with every missed crossing until it is re-anchored. Fine for one door; brittle for open plans, atriums and multi-entry zones.
  • Wi-Fi and Bluetooth counting. Seductive because the access points are already on the ceiling, so it looks free. The costs surface later: device-to-person ratios that swing with policy, dwell inflation from parked phones, and a privacy analysis that is worse than it looks because device identifiers persist. We wrote up why MAC counting breaks.
  • Camera-based edge AI. The highest hardware line and the richest output: true zone counts, density, dwell, queues, from ceiling height, in real time. The decisive cost question inside this category is architectural: on-device processing with no stored footage keeps the governance line short; anything that records shifts real money into line seven for the life of the system.

03 · THE FIVE-YEAR ARITHMETIC THAT KEEPS QUOTES HONESTThe five-year arithmetic that keeps quotes honest

Run every proposal through the same frame: total five-year cost, divided by zones covered, divided by five. That one number, cost per zone per year, absorbs the licence trigger, the maintenance reality and the amortised install, and it makes radically different pricing models comparable on one line. Two patterns emerge immediately when you do this:

Hardware is bought once. Licences, governance and recalibration are bought every year. The five-year view is the only honest one.

First, hardware differences shrink: a pricier sensor amortised over five years is dollars per zone per month. Second, licence and governance differences compound: they recur, scale with coverage, and never amortise. The expensive-looking system with one bill per zone and a governance review that fits on a page is frequently the cheapest thing on the shortlist by year three.

An aerial view of a dense city skyline at dusk, the portfolio scale at which occupancy sensing costs compound across dozens of buildings and hundreds of zones
At portfolio scale every recurring line compounds. The licence trigger you accept for one floor is the one you will pay across every building.

04 · TEN QUESTIONS TO PUT ON EVERY VENDOR'S TABLETen questions to put on every vendor's table

Ours included. A vendor who answers these quickly is quoting a system; one who cannot is quoting a sensor.

  1. What is the licence trigger: device, zone, area or site, and what happens when I add coverage?
  2. What is the all-in, five-year cost for one floor of six zones, installed?
  3. How is accuracy validated on my site, and is a ground-truth audit included in the pilot?
  4. What does a re-zone cost when the floor plan changes?
  5. What data leaves the building, exactly, and how do I verify that on the network?
  6. What is the privacy governance load: what will my legal team need to review, once and annually?
  7. What does the BMS integration require from my side?
  8. Is the count consumable without your dashboard, API-first, if my team wants the raw feed?
  9. What breaks if your company disappears: does the hardware keep counting locally?
  10. What are the contract minimums, and what does export look like at exit?

Where we land in this structure is no secret: per-zone pricing, anonymous by architecture so line seven stays short, ground-truth validation in the pilot, and the sensor speaking standard protocols so line six stays sane. If you want the five-year number for your actual floor, scope it with us, and bring the ten questions.

If your building can't answer one of these questions yet, we should talk.