What is an occupancy confidence score?

How it works

An occupancy reading says a desk, room, floor or building was in use at a given time. A confidence score sits beside it and says how much evidence supports it. When a booking, a check-in, network presence and a badge swipe all point the same way, confidence is high. When only one source reports, or sources conflict, confidence is low.

There is no industry-standard formula. A simple version, shown below, counts how many of the sources that cover a space agree with the reading. More advanced models weight each source by how reliable it has proven at that site, since a sensor on a desk is stronger evidence for that desk than a building badge swipe.

The score is most useful in aggregate. A floor where most readings have high confidence can support decisions. A site where most readings rest on one source tells the team to add a signal, raise check-in adoption or fix a data feed before relying on it.

Formula

Confidence score = (Sources that agree with the reading ÷ Sources covering the space) × 100

Sources that agree with the reading: Signals for the same space and time that report the same state, in use or empty.
Sources covering the space: Signals that report on that same space at the same level of detail and time, such as a booking, a check-in and a desk sensor for one desk. A building badge entry covers a building, not a desk.

Worked example · illustrative figures

A desk on Tuesday morning is covered by three desk-level sources: a booking, a check-in and a desk sensor.
The check-in and the sensor both show the desk in use. There was no booking, so the booking disagrees.
2 ÷ 3 = 0.67.

Confidence score: 67%, because two of the three desk-level sources agree. The score measures agreement between sources; it is not the probability that the desk was in use.

How to measure it

Decide which sources count for each type of space. A building badge reader covers a building, not a desk.
Score each reading, then report the share of readings at high, medium and low confidence by site and floor.
Track the score over time. A falling score usually means a broken data feed or falling check-in adoption.

Common mistakes

Counting sources that are not independent, such as a booking and the automatic check-in it created, as two votes.
Reporting occupancy figures without their confidence, so a weak site looks as solid as a strong one.
Treating low confidence as low occupancy. It means the evidence is thin, not that the space is empty.

Frequently asked questions

Why does occupancy data need a confidence score?

Occupancy figures drive expensive decisions, such as releasing a floor or renewing a lease. A confidence score shows which figures rest on several independent signals and which rest on one, so teams act on solid data and fix weak data first.

How do you improve occupancy confidence?

Add an independent signal where coverage is thin, such as office network presence where only badges exist. Raise check-in adoption and repair broken data feeds. Confidence rises when more sources cover each space and they agree.

Is there a standard occupancy confidence formula?

No. A simple version divides the number of sources that agree with a reading by the number of sources covering that space. Some models also weight sources by how reliable they have been at each site. Whatever the method, write it down so every site is scored the same way.

Multi-source occupancy (unified occupancy model) One occupancy picture built by combining Wi-Fi, badge, sensor and booking data, so no single source has to be complete or perfect.
Presence data (sensorless occupancy) The record of who was actually in the building and when, from signals you already have: network, badges and check-ins.
Occupancy rate The share of available desks, rooms or floor area in use, at a point in time or on average over a period.
Occupancy monitoring Continuous measurement of how many people use each building, floor, room and desk, and when, from presence signals or sensors.
Badge data Access-control swipe records showing who entered a building or floor and when, often used to measure attendance.
Occupancy sensor A device that detects whether a desk, room or zone is in use, or counts the people in it, usually anonymously.

Kadence editorial team · Last reviewed October 1, 2026