Most organisations that buy an occupancy monitoring system never verify whether it actually works. They install sensors, connect dashboards, and start making six-figure real-estate decisions based on numbers nobody has audited. That is the equivalent of running your finance department on a spreadsheet you've never reconciled. If your occupancy data is off by even 10%, a decision to consolidate two floors into one could leave 80 people without desks on a Tuesday morning — and the facilities team wearing the blame.
So before you ask what your occupancy monitoring system tells you, ask a harder question: how do you measure the system itself? Here is a practical framework built from real deployments in offices, universities and public buildings — the environments where measurement errors are most expensive and least forgiven.
Accuracy is the foundation everything else stands on, yet it is the metric vendors are most vague about. A serious provider will contractually commit to a minimum accuracy figure. At Vemco Group, which has been in people counting and occupancy analytics since 2005, the contractual minimum is 96%, and typical performance reaches 98–99% when conditions such as lighting, layout and visitor behaviour allow. Notice the honesty in that phrasing: any vendor promising a flat 99% regardless of environment either hasn't deployed in a glass-walled atrium with harsh backlight, or is hoping you won't check.
To validate accuracy yourself, run a manual audit. Pick two or three entrances, count people physically (or via recorded footage where policy allows) for defined windows — a busy morning arrival period and a quiet mid-afternoon — and compare against system counts for the same timestamps. Do this at commissioning, then again quarterly. Environments change: a new coffee cart near an entrance, a rearranged reception, seasonal light angles. All of these can shift sensor performance, and only periodic audits catch the drift.
Here is something implementers learn quickly and buyers learn too late: raw counts lie in staffed buildings. A university library entrance might register 40 movements an hour, but if 12 of those are security staff, cleaners and librarians crossing the threshold repeatedly, your utilisation figures are inflated before you've made a single decision. Modern AI sensors with staff exclusion solve this — Vemco's systems can filter out staff from occupancy counts — but you should test the exclusion, not just tick the feature box. Ask a known staff member to walk the entrance five times during your audit window and confirm the count doesn't move.
One practitioner observation from years of commissioning: the most common source of "bad data" complaints isn't the sensor at all — it's an unmonitored secondary exit. People enter through the counted main door and leave through a fire-exit-turned-shortcut, so live occupancy figures creep upward all day and never reset properly. Before blaming the technology, walk the building and map every real-world entry and exit path, including the ones the architectural drawings don't admit people use.
Historical reporting tolerates a few minutes of latency. Live occupancy management does not. If your occupancy monitoring system triggers alerts at a predefined limit — for capacity compliance, safety thresholds or crowd management in public buildings — you need to measure the time between the threshold being crossed physically and the alert reaching the person who must act. Test it: stage a controlled crossing of the limit and time the alert. Anything over a minute or two turns a safety control into a historical record.
For universities and public institutions with statutory capacity limits, this latency test belongs in your acceptance criteria, not in a post-incident review.
Once you trust the data, measure the system by the quality of decisions it enables. The metrics that matter to facility managers and workplace leaders are rarely raw headcounts. They are:
The hardest and most valuable measurement is financial return, and it only becomes possible when occupancy data stops living in its own dashboard. Buildings running HVAC on fixed schedules condition empty floors every evening and every quiet Friday. When occupancy data feeds directly into HVAC, BMS and security systems — the role VemFusion plays in Vemco deployments — ventilation and climate control follow actual presence instead of assumptions. Offices, universities and public buildings use exactly this approach to cut the waste that always-on systems generate by default.
To measure this properly, establish an energy baseline before integration: metered consumption per zone, per schedule, for at least one comparable seasonal period. Then compare post-integration consumption against it. Without the baseline, you'll have a plausible story instead of a defensible number — and finance directors fund numbers, not stories.
An occupancy monitoring system is only as credible as the last time someone tested it. Organisations that treat measurement as an ongoing discipline — audited accuracy, verified exclusions, timed alerts, baselined savings — end up with data their leadership actually trusts. Those that don't end up defending dashboards nobody believes.
If you want help defining acceptance criteria for a new deployment, or auditing the accuracy and ROI of an occupancy monitoring system you already run, talk to the Vemco Group team at vemcogroup.com/contact-us — bring your floor plans and your energy bills, and we'll show you exactly what to measure first.