Meeting room portfolios are often provisioned using peak-demand assumptions and booking data that overestimate actual use. This study aims to develop and test an evidence-based framework for optimizing meeting room capacity and configuration using minute-level occupancy sensing and analytics.
Seat-mounted sensors recorded minute-by-minute occupancy in 17 meeting rooms over 10 working days, yielding 802 detected meetings. Service-level sizing at the 95th percentile (P95) and Full Downward Substitution (FDS) produced an FDS service-level room configuration. K-means clustering identified meeting archetypes, which were used to parameterize a collision-aware Monte Carlo simulation that evaluated 486 candidate configurations over 10,000 iterations.
The FDS service-level configuration met the aggregate P95 target but still resulted in a 5.71% failure rate, underscoring the importance of temporal overlap. Clustering revealed two dominant archetypes (light and heavy), and the data indicated a material need for single-person focus rooms (28% of meetings). In the studied meeting room portfolio, the selected configuration (1 Focus [F], 4 Small [S], 5 Medium [M], 3 Medium-plus [MP], 1 Large [L] and 1 Extra-large [XL]) reduced the active meeting area by 22.8% while lowering failures to approximately 3%.
Minute-level sensing enables corporate real estate (CRE) teams to set explicit service targets, quantify tradeoffs between space and failure risk and validate redesign decisions under uncertainty. The framework offers an implementable blueprint for meeting room redesign and governance.
The study integrates service-level sizing, substitution-based room-configuration design, clustering and simulation-based validation into a single workflow, providing an actionable link between occupancy analytics and portfolio design. This integrated approach enables CRE teams to move beyond measurement to prescriptive action, addressing a gap left by prior studies that stop at utilization reporting.
