Occupancy Assessment

Reported behaviour and observed behaviour diverge — reliably, in predictable directions. People over-report presence, under-report absence, and describe space use as they intend it rather than as it happens. The Occupancy Assessment supplies the observational counterpart: actual presence, utilisation by setting type, peak patterns, sharing behaviour, avoidance and overload. Alone, it explains nothing about why. Paired with reported demand, it is the triangulation engine of the entire diagnosis — the method that converts two partial pictures into evidence.

• During Diagnose, fielded in parallel with the Dynamics Assessment so reported and observed data cover the same period
• Over a minimum representative cycle — long enough to capture rhythm, never a two-week snapshot sold as behaviour
• Before any capacity modelling: Sharing Ratio and Calibration are only as good as this baseline
• Continuously in Recalibrate as the standing behavioural evidence stream

• Project team configuring measurement scope with FM and IT
• Data protection officer approving the measurement architecture before deployment
• Works council where applicable — observational measurement without consultation destroys the trust the strategy needs
• Strategist defining which questions the measurement must answer (from the triangulation register)

• Floor plans and space inventory from Data Requests, coded by setting type
• Existing data sources for cross-validation: badge, booking, Wi-Fi, sensor — each with known limitations documented
• Triangulation register from the Dynamics Assessment: the reported claims requiring observational testing
• Privacy configuration: aggregate measurement, no individual tracking, cohort thresholds on all reporting

1. Define the measurement questions first, instrumentation second — measuring everything measurable produces data lakes, not evidence.
2. Configure coverage across setting types: desks, meeting rooms, collaboration areas, quiet and focus spaces, social space, support space; include environmental conditions where relevant to the questions.
3. Establish the measurement period across a representative cycle, capturing weekly rhythm, month-end effects and known seasonal patterns.
4. Measure utilisation and occupancy as distinct metrics: a floor at 45% average occupancy with 95% Tuesday peaks is one finding, not two contradictory ones.
5. Analyse time-based patterns: daily curves, weekly rhythm, duration-of-use per setting, booking-versus-presence gaps for reservable spaces.
6. Detect behavioural signatures: avoidance (consistently empty settings), overload (queuing, overflow into wrong settings), sharing friction, meeting-room hoarding.
7. Run the triangulation: reported presence versus observed presence, reported crowding versus measured density, claimed collaboration versus collaboration-space use — per team where thresholds permit.
8. Report divergences as questions for Targeted Insights, not as verdicts.

• The booked-but-empty pattern: high reservation rates with low measured presence in meeting rooms
• Avoided settings: spaces the concept assumed valuable that behaviour has voted against
• Peak concentration: average utilisation that hides unusable Tuesdays and dead Fridays
• Overflow behaviour: focus work in cafés, calls in stairwells — demand the current supply cannot place
• Reported–observed gaps that are systematic per team rather than random — those gaps have causes worth finding

• Occupancy and utilisation baseline by setting type, time and (threshold permitting) team
• Peak-pattern and rhythm analysis
• Behavioural signature register: avoidance, overload, hoarding, overflow
• Reported-versus-observed divergence analysis
• Question set for Targeted Insights

The baseline feeds Taxonomy Spatial Benchmarking (supply performance), Sharing Ratio (peak and rhythm assumptions), and Calibration (attendance assumptions). Divergences scope Targeted Insights. Behavioural signatures inform the Kit of Parts — avoided settings are design evidence. In Recalibrate, continuous occupancy data measures whether the delivered concept behaves as modelled.

Occupancy data describes what; it never explains why. An empty quiet room is acoustics, culture, location, furniture or policy — five different strategies. Deciding which divergences matter, what period counts as representative, and when a pattern is behaviour rather than noise requires analytical judgement. The platform computes the patterns; the strategist decides what they mean.

• Selling booking data or a two-week study as behavioural evidence
• Reporting average utilisation without peak structure — the single most consequential analytical error in the field
• Measuring without questions, producing dashboards instead of answers
• Treating observed behaviour as preference when it is adaptation to current constraints
• Deploying sensors without works-council consultation and privacy architecture, poisoning trust for every later method

A pharmaceutical company reported chronic meeting-room shortage; booking data showed 89% reservation. Measured presence showed 41% of booked slots fully unused and median actual meeting size of 3.2 in rooms averaging 8 seats. The shortage was real — but it was a shortage of small rooms and a surplus of hoarding, not a shortage of rooms. The Kit of Parts answer cost a fraction of the requested expansion.