Occupancy monitoring in public spaces — libraries, council offices, university campuses, transport hubs, NHS estates — is pitched as a route to 20-30% energy savings and rightsized property portfolios, and the numbers stack up when the deployment is disciplined. The problem is that most first-time buyers skip the discipline. They procure a building-wide sensor blanket, generate dashboards nobody acts on, and quietly write the pilot off after 18 months. The RICS 2023 Public Sector Property Report found that only 22% of local authority estates had usable occupancy data, despite years of investment in smart building kit. This piece walks through what occupancy monitoring genuinely unlocks in public spaces, why the blanket-deployment reflex fails, and the single-zone adoption strategy that produces payback inside a year.
What the savings actually look like when the maths is honest
The headline claim — that occupancy monitoring can cut building energy use by 20-30% — comes from a specific mechanism: matching HVAC, lighting, and cleaning schedules to real occupancy rather than assumed occupancy. The Carbon Trust’s 2022 analysis of UK non-domestic buildings found that HVAC alone accounts for around 40% of operational energy in offices, and that between 15% and 25% of that is wasted heating or cooling under-occupied zones during nominal working hours. In public buildings the waste is often worse. A council customer service centre scheduled for 08:00-18:00 occupancy typically sees genuine footfall concentrated in a four-hour window; a university lecture theatre booked for 90 minutes is frequently vacated after 40. Occupancy sensors expose that gap.
The second saving is rightsizing. The UK Government Property Agency’s 2023 State of the Estate report noted average workspace utilisation across the central government estate sat at 39%, meaning more than half the leased footprint was structurally underused. Rightsizing decisions — consolidating floors, sub-letting wings, deferring lease renewals — depend on defensible occupancy data, not on badge swipes or manager estimates. A single credible utilisation figure has, in several documented cases, unlocked seven-figure property savings. Occupancy monitoring at engineering consultancy Hilson Moran identified a peak occupancy of just 71% across the office, which is precisely the kind of number that converts a boardroom argument about desk allocation into a lease decision.
The third saving, less discussed, is operational: cleaning contracts, security patrols, and reception staffing scaled to genuine footfall rather than worst-case schedules. Verdantix’s 2024 Smart Building Technologies benchmark put combined soft-services savings from occupancy-informed scheduling at 8-14% for large public sector estates.
Why building-wide rollouts keep failing
The dominant procurement pattern in the public sector is the blanket deployment: an estate manager, under pressure from a net-zero deadline, buys 500 sensors, wires them into a dashboard, and waits for insights to arrive. They don’t. What arrives is a heatmap that everyone agrees is interesting and nobody knows how to act on. The BEIS Public Sector Decarbonisation Scheme evaluation published in early 2024 found that of 340 monitored capital projects, only 31% could demonstrate operational behaviour change traceable to the data being collected.
The failure mode is not technical. Modern PIR, thermal, and radar occupancy sensors from vendors including Disruptive Technologies, VergeSense, Butlr, and Irisys are all accurate enough for the job. The failure is that no specific decision was written down before the sensors were bought. Without a decision the data must trigger — close floor 3, reduce cleaning frequency in the east wing, shift the reception rota, renegotiate the HVAC contract — the dashboard becomes decorative. This is the pattern the Cabinet Office’s Government Property Function has been quietly pushing back against in its 2024 guidance, which now recommends “decision-first” utilisation projects rather than estate-wide instrumentation.
There is a legitimate objection here: building-wide deployments do capture cross-zone circulation patterns that single-zone pilots miss, and for mature estates already fluent in acting on utilisation data, the wider net is the right net. But most public sector buyers are not mature estates. They are first-time buyers, and the evidence is that first-time buyers who skip the single-zone discipline stall on dashboards.
The single-zone strategy that actually produces payback
The strategy that consistently produces payback inside 12 months looks the same across the case studies worth reading: pick one high-traffic zone, name the specific decision the data must trigger, instrument only that zone, and hold the project to that decision. The zone should be somewhere the operational cost of getting occupancy wrong is high — a main entrance hall, a shared meeting suite, a set of consultation rooms, a lecture block. The decision should be specific enough to be falsifiable: “if peak occupancy is below X for six consecutive weeks, we consolidate”, not “we will understand our space better”.
A UK construction firm’s HQ deployment of 120 occupancy sensors and 65 indoor air quality monitors delivered a reported 249% ROI with payback in under 12 months, and the reason the numbers worked was that the deployment was scoped around named decisions on floor consolidation and HVAC scheduling before procurement began. The sensor count sounds large, but it was concentrated in a defined zone with defined questions attached — not sprayed across the estate.
For public sector buyers, the practical sequence is: (1) identify the zone where a rightsizing or scheduling decision is already on the table but stuck for lack of evidence; (2) write the decision criteria down, with thresholds and a review date; (3) specify sensors sized to that zone, ideally with configurable monitoring areas so the zone boundary can be adjusted without re-installing hardware; (4) commit to acting on the data at the review date, or to a documented reason not to. Tools like DIREK‘s SpaceLens are among the platforms that support configurable zone boundaries, which matters because public space usage patterns shift as services are reorganised.
A note on procurement: EN 16798-1 and CIBSE TM40 give useful reference frameworks for defining occupancy-linked ventilation and thermal targets, and tying the pilot to one of these standards makes the business case defensible to a Section 151 officer in a way that a vendor whitepaper never will.
What public space buyers should watch for
Three signals separate a serious occupancy pilot from a dashboard purchase. First, privacy posture: PIR and radar sensors that count presence without identifying individuals are appropriate for public spaces; camera-based systems raise Data Protection Impact Assessment obligations under UK GDPR that most councils are not resourced to handle. Second, integration path: sensor data that cannot flow into the existing BMS or CAFM system will not change behaviour, because the operational teams live in those systems. Third, contract structure: sensor-as-a-service pricing that includes analytics and quarterly review sessions produces more decisions than one-off capital purchases, because it forces someone external to keep asking what the data means.
The Public Sector Decarbonisation Scheme, now in its Phase 3b window with £1.17bn allocated, explicitly permits occupancy monitoring within eligible measures, but only where it is linked to demonstrable operational savings — another reason the decision-first framing matters at the funding application stage, not just at the deployment stage.
Occupancy monitoring unlocks genuine savings in public spaces — energy, property, soft services — but the savings are extracted by discipline, not by density of sensors. The buyers who succeed name the decision first and instrument one zone against it; the buyers who fail buy the blanket and hope the insights arrive on their own. This week, the useful exercise for any public sector estate lead considering an occupancy pilot is to pick one high-traffic zone and write down, in a single sentence, the decision its data must trigger before any quote is requested.