A Foundational Guide for Facility Management (FM-007) · By David Gray | DavidGrayProjects.com
Introduction: From Opinions to Evidence
Corporate real estate (CRE) and facility management (FM) decisions used to run on anecdote: "the floor feels full," "we need more conference rooms," "the renewal rate looks cheap." Those instincts are not always wrong. But leadership now expects evidence: clean data, defined metrics, and a clear line from each recommendation to its financial consequence.
This article is not about buying another dashboard. It is about building enough data discipline that a CFO can trust the answer when you recommend giving back a floor, renewing a lease, consolidating two sites, or funding a capital project. It builds on the workplace and technology themes in FM-002 and FM-003, and gives the lease framework in FM-005 the evidence base it needs.
The thesis is simple: the value of real estate data is measured by the quality of the decisions it changes, not by the volume of data collected. Everything below is organized around that idea.
1. Why the Bar for Evidence Has Risen
Four pressures have raised the standard of proof for space and cost decisions.
Hybrid volatility. Attendance varies by team, day of week, and season. A single average hides the peaks that actually drive capacity, and the troughs that drive cost per use.
Cost scrutiny. Occupancy is a large and sticky cost. Rent, operating expenses, services, and depreciation do not flex with attendance. Leaders want to know which portion of that cost is buying productive use.
Reporting obligations. Energy, water, and emissions disclosures need defined boundaries and auditable source data, not estimates assembled at year-end.
Workplace experience. Space that does not support how people work wastes money even when it is "fully utilized." A floor where every desk is booked but no one can find a quiet room for a call is not efficient; it is mis-designed.
The consequence is that FM and CRE leaders are increasingly asked to defend recommendations the way a finance team defends a forecast: with stated assumptions, traceable sources, and a range of outcomes rather than a single number.
2. The Core Datasets - and What Each Actually Measures
The most common analytical mistake is treating different data sources as interchangeable. They are not. Each one measures a different thing and carries a different bias.
| Dataset | What it measures | Common blind spot |
|---|---|---|
| Badge access | Building entry | Does not show where people sit or how long they stay; tailgating undercounts |
| Wi-Fi associations | Device presence by zone | Multiple devices per person; signal bleed between floors |
| Desk and room booking | Intent to use | No-shows and walk-ups distort the picture |
| Occupancy sensors | Physical presence at a desk or in a room | Coverage gaps, calibration drift, and privacy design |
| Surveys and service tickets | Experience and pain points | Response bias; loudest voices dominate |
| Lease abstracts, operating-expense reconciliations, general ledger | Contracted and actual occupancy cost | Inconsistent cost allocation across sites and business units |
| Building automation, submeters, utility bills | Energy and demand | Missing intervals; meters that do not map to spaces or tenants |
| CMMS, condition assessments | Asset health and maintenance effort | Poor asset hierarchy; incomplete work-order history |
Before any decision, document which source answers which question, and reconcile sources that disagree. If badge data says 60 people entered the building and sensors say 85 desks were occupied, one of those numbers is wrong or they are measuring different things. Resolve that before presenting either to an executive.
I learned this the uncomfortable way. On one portfolio review, our team walked into a steering meeting with three different occupancy figures for the same floor: one from security's badge reports, one from the booking tool the workplace team managed, and one from a sensor pilot. Each department defended its own number, and the meeting spent most of its time arguing about whose data was right instead of what to do with the floor. We went back, spent a few weeks reconciling the sources and writing down what each one actually measured, and the next meeting took a fraction of the time. The decision did not change much. The confidence in it did.
A simple source-to-question map
A practical discipline is to write down, for each decision, the primary and secondary source:
"How much workspace do we need?" Primary: sensor or validated booking data at percentile peaks. Secondary: badge data for building-level trend.
"What does each seat actually cost us?" Primary: lease abstracts and general ledger. Secondary: service contracts and depreciation schedules.
"Where are people unhappy with the space?" Primary: structured surveys and post-occupancy evaluations. Secondary: service tickets by category and location.
"Which buildings are inefficient to operate?" Primary: utility and submeter data, benchmarked. Secondary: building automation trend logs and CMMS work-order density.
When the primary source is missing or unreliable, say so. A recommendation built on a secondary source should carry a wider uncertainty range and, usually, a smaller commitment.
3. Metrics That Matter to Executives
Executives do not need fifty KPIs. They need a small set of well-defined measures that connect space to money, risk, and performance.
Cost per productive seat. Total occupancy cost divided by a consistently defined denominator. State whether you mean assigned seats, provided seats, or peak concurrent use; mixing them makes the metric meaningless.
Peak and percentile utilization. Report the distribution (for example, the 80th or 90th percentile day) rather than the mean. Validate that the observation period covers seasonal and business-cycle peaks.
Space efficiency by type. Compare provided capacity with observed use separately for workstations, focus rooms, and meeting rooms. A floor can be underused overall and still short of the room types people need.
Energy use intensity (EUI) and demand cost. Benchmark with established methods. ENERGY STAR Portfolio Manager, for example, reports weather-normalized source EUI, which adjusts for year-to-year weather and converts site energy to source energy for comparability [1]. Do not compare raw site EUI across climates or building types.
Project NPV, IRR, and payback. Keep modeled savings separate from measured and verified savings. For energy projects, use a recognized measurement and verification approach such as IPMVP, which defines savings against an adjusted baseline rather than a simple before-and-after comparison [2].
Carbon and comfort. Publish the calculation method and boundary (for example, location-based versus market-based Scope 2 under the GHG Protocol [3]) and the comfort thresholds you are managing to.
Every dashboard should end in a decision and a financial implication (the same principle I apply to data-driven project controls): what changes over the next 12 to 36 months if management acts, and what it costs if it does not.
Worked illustration: why the denominator matters
Illustrative example only. The figures below are hypothetical and chosen for easy arithmetic; they are not drawn from any actual portfolio.
Assume a leased office with a fully loaded annual occupancy cost of $3.0 million (rent, operating-expense pass-throughs, services, and depreciation of tenant improvements) and 600 workstations.
Cost per provided seat: $3.0M / 600 = $5,000 per seat per year.
If the 90th-percentile day shows 330 seats in use, cost per peak-occupied seat: $3.0M / 330 = about $9,100.
If the average day shows 240 seats in use, cost per average-occupied seat: $3.0M / 240 = $12,500.
All three numbers are "correct." They answer different questions. The first is useful for benchmarking fit-out density. The second tells you what capacity is actually buying at peak. The third is the number most likely to alarm a CFO, and the one most likely to be used to justify an aggressive cut. A recommendation that sizes the space to the average day would leave roughly 90 people without a seat on a peak day. Choose the denominator deliberately and disclose it.
4. Analytics Maturity - Honestly Assessed
Descriptive: What happened? Dashboards and heatmaps.
Diagnostic: Why did it happen? Drivers and competing explanations.
Predictive: What is likely next? Demand or equipment-risk forecasts with stated uncertainty.
Prescriptive: What should we do? Alternatives compared under explicit constraints.
Most portfolios do not need to reach stage four everywhere. Data quality, privacy constraints, and integration cost differ by site. Aim for the maturity level the decision requires, not the one a vendor roadmap suggests.
A useful test: if the decision would not change regardless of what the model predicts, the model is not worth building. A building with a lease expiring in nine months and a signed relocation does not need a predictive occupancy model. A portfolio of forty sites with staggered expiries over five years probably does.
Maturity self-assessment
| Question | If "no" |
|---|---|
| Do we have one agreed definition for each core metric? | Stay at descriptive; fix definitions first |
| Can we reconcile badge, booking, and sensor data for priority sites? | Do not build predictive demand models yet |
| Do our space and asset identifiers match across systems? | Invest in normalization before visualization |
| Have we compared a past forecast with what actually happened? | Treat all forecasts as scenarios, not predictions |
| Does any metric currently drive a budget or lease decision? | Connect one metric to one decision before adding more |
5. Five High-Value Use Cases
Hybrid rightsizing. Combine peak occupancy, desk and room demand, and employee feedback before reducing area. Remember the financial distinction: reconfiguring or removing desks does not reduce contracted rent. Savings arrive only through a lease event, sublease, or consolidation, which is why this analysis belongs in the lease timeline described in FM-005. Before a give-back, model the full cost: restoration obligations, sublease downtime and broker fees, and the capital cost of reconfiguring the space you keep.
Cleaning and soft services. Align service frequency with actual use while meeting hygiene standards and contractual service levels. Change the contract terms, not just the schedule, or the savings will not materialize. Watch for the hidden cost: if service reductions generate complaints and reactive call-outs, the net saving can disappear.
Energy and demand management. Use interval data and occupancy patterns to tighten HVAC schedules and setbacks, after confirming comfort, ventilation, and equipment constraints with engineering. Demand charges often respond to peak management (staggered start-up, pre-cooling) more than to total consumption reduction, so look at the tariff before choosing the measure.
Capital planning. Combine equipment condition, consequence of failure, energy performance, and replacement cost to rank investments. (FM-008 covers the deferred-maintenance side of this in detail.) A building scheduled for exit in three years should not receive the same renewal investment as a long-term hold.
Renewal versus relocation. Compare future occupancy demand, total occupancy cost, workforce access, business continuity, and schedule risk across options with the same data definitions. Relocation analyses often understate one-time costs (moves, IT, furniture, duplicate rent during transition) and overstate the speed at which a new space reaches stable use.
A note on numbers: published "typical savings" percentages vary widely by portfolio and method. Build your own baseline and present projected savings as a range with stated assumptions. Where this article does not cite a figure, that is deliberate.
Worked illustration: consolidation is a lease decision, not a furniture decision
Illustrative example only; hypothetical figures.
Suppose analysis shows two floors of a three-floor lease could be consolidated into one at the 90th-percentile day. Annual rent and operating costs are $1.1 million per floor.
Option A - reconfigure only. Remove workstations on one floor and convert it to collaboration space. Capital cost: $400,000. Rent saving: $0. Benefit: experience and capacity, not cost.
Option B - sublease one floor. Assume 9 months of downtime, a sublease rate at 70% of contract rent for the remaining four years, broker and legal fees of $150,000, and $250,000 to reconfigure the retained floors. Over four years the gross recovery is roughly 3.25 years x $1.1M x 70% = about $2.5 million, against about $400,000 in costs and the continuing primary-lease liability if the subtenant defaults.
Option C - negotiate an early surrender of one floor with the landlord. No sublease or subtenant-credit risk, but landlords typically require a termination payment, and the saving depends entirely on what the landlord will accept given its own re-leasing prospects.
Option D - wait for a contraction right or renewal, if the lease provides one. Lowest execution risk, but the saving is deferred and the organization keeps paying for space it does not need in the meantime.
The right answer depends on market conditions, credit risk, and timing. The point is that utilization data identifies the opportunity; the lease determines whether it becomes money. A presentation that jumps from a heatmap to "$1.1 million in annual savings" has skipped the most important part of the analysis.
I have sat through more than one presentation that made exactly that jump. In one case, a well-built utilization study showed a floor that was rarely more than a third full, and the savings slide simply multiplied the floor's rent by the remaining years on the lease. When we pulled the lease, the tenant had no contraction right, the sublease market in that submarket was soft, and restoration obligations on the floor were significant. The space was still worth addressing, but the honest number was a fraction of the headline, and it arrived years later than the slide implied. Getting that correction on the table early protected the credibility of everything else the team was recommending.
6. A Practical Data Stack
You do not need an enterprise data platform to start. You do need a disciplined sequence.
Ingest badge, booking, sensor, utility, CMMS, and lease data through authorized, documented access.
Normalize building and space identifiers, time zones, units, and floor-plan versions. This step usually takes longer than expected, and it is where most projects stall.
Store raw and transformed data separately, with access controls and retention rules.
Define every metric once, in a data dictionary that names the owner, formula, and source.
Visualize role-specific views: executives need decisions and trends; site managers need exceptions and work queues.
Automate refreshes, data-quality checks, and lineage so the numbers can be traced back when challenged.
The data dictionary is the product
If you build only one artifact, build the data dictionary. It is also the foundation of the boardroom-ready reporting and governance that executives expect from any program. A minimal entry for each metric should include:
| Field | Example |
|---|---|
| Metric name | Peak workstation utilization (P90) |
| Definition | 90th-percentile daily maximum of occupied workstations / provided workstations, business days only |
| Source system | Desk occupancy sensors; booking system as fallback |
| Exclusions | Public holidays; days with more than 10% sensor outage |
| Owner | Workplace analytics lead |
| Refresh | Weekly |
| Known limitations | Sensors not installed on floor 3 until Q2 |
This single table prevents most of the arguments that otherwise consume executive meetings.
7. Data Quality, Privacy, and Trust
Reconcile conflicting systems before presenting results. Look for booking no-shows, dead sensors, holiday weeks, and one-off events that skew results.
On privacy, design for trust from the start:
Report aggregated data wherever possible, and set minimum group sizes for any report that could identify individuals.
Involve legal, HR, and information security before deployment, not after the first complaint.
Recognize that privacy and labor rules differ by jurisdiction, and some (for example, rules governing biometric data) carry significant penalties.
Document the purpose of each dataset and do not repurpose occupancy data for individual performance management without a separate, explicit decision.
Occupancy sensors and connected building systems are also part of the operational-technology (OT) attack surface. NIST's guide to OT security explicitly covers building automation and physical access control systems [4]. Treat new sensor networks as you would any connected system: segmented, inventoried, patched, and owned.
8. Change Management
People will assume space analytics is surveillance unless you show otherwise. State the purpose (better space and services, not individual monitoring), publish the small set of KPIs you track, involve department leaders in interpreting the results, and validate proposed changes through pilots and feedback before committing capital.
Three practices make a measurable difference:
Share results back. If data shows a team's meeting rooms are always full, tell them what you are doing about it. Analytics that only ever lead to cuts will lose cooperation quickly.
Pilot before scaling. Reconfigure one floor, measure for a full business cycle, and publish what worked and what did not.
Give leaders a role. Department heads who help interpret their own data are far more likely to support the resulting decisions.
9. Roadmap: 90 / 180 / 365 Days
First 90 days. Baseline utilization, occupancy cost, and energy performance for priority sites. Assign metric owners and publish the data dictionary. Pilot one low-risk service adjustment and measure it.
By 180 days. Integrate the core sources. Build one rightsizing or reconfiguration business case tied to a real lease or capital date. Test condition-based maintenance on a small set of critical assets. Report results, including what did not work, to executives.
By 365 days. Scale the use cases that produced verified results. Connect analytics to the budget and capital planning cycle. Require the standard metrics in every lease and investment review.
10. Decision Gates: When Is the Data Good Enough?
Data is never perfect. The question is whether it is good enough for the size of the decision.
| Decision | Minimum evidence before committing |
|---|---|
| Adjust cleaning or service schedules | Four to eight weeks of reconciled occupancy data; contract review |
| Reconfigure a floor | One full business cycle of utilization by space type; user feedback |
| Sublease or give back space | Percentile peak analysis across seasons; headcount plan; lease and market review |
| Relocate or consolidate sites | All of the above, plus workforce-access analysis, full one-time cost model, and scenario ranges |
| Major energy capital project | Weather-normalized baseline; engineering study; M&V plan agreed before construction |
Scaling the evidence requirement to the size and reversibility of the decision keeps analytics from becoming either a rubber stamp or a reason never to act.
11. Five Slides for the CFO
Current portfolio performance against defined metrics.
The material gaps between space provided, demand observed, and cost.
Recommended interventions and the alternatives considered.
Investment, projected savings range, risks, and sensitivity.
Measurement plan, decision gates, and who is accountable.
Conclusion
Data should help CRE and FM leaders provide the right space, in the right place, for the work that actually happens, and prove the value in cost, experience, resilience, and environmental performance. Clean definitions and disciplined decisions matter more than sophisticated dashboards. The leaders who earn credibility are the ones who disclose their denominators, separate modeled from verified savings, and connect every insight to a lease, budget, or capital decision.
Next in the series, FM-008 looks at the cost of the decisions organizations defer.
Related from Albers Management:
Data-Driven Project Controls: Turning Information into Insight – Albers Management
Reporting & Governance: Building Boardroom-Ready Confidence – Albers Management
Sources
ENERGY STAR Portfolio Manager Technical Reference: Climate and Weather; Source Energy Technical Reference. energystar.gov; Source Energy (PDF)
Efficiency Valuation Organization, International Performance Measurement and Verification Protocol (IPMVP). evo-world.org
GHG Protocol, Scope 2 Guidance (location-based and market-based methods). ghgprotocol.org
NIST SP 800-82 Rev. 3, Guide to Operational Technology (OT) Security (2023). csrc.nist.gov
Worked examples labeled "illustrative" use hypothetical figures to explain methods.
About the Author
David Gray is a capital delivery strategist, owner's representative, and founder of DavidGrayProjects.com. With over two decades of experience helping organizations bring complex projects to life, from data centers and healthcare facilities to higher-ed campuses, David blends practical delivery with forward-thinking strategy.
He writes about project controls, capital planning, and real estate development to help leaders deliver smarter, faster, and more sustainably.
📩 Connect on LinkedIn | 🌐 Explore More at DavidGrayProjects.com | 🌐 Explore More at Albers Management