What is energy benchmarking?

The two metrics

Energy use intensity (EUI) is total annual energy divided by gross floor area, expressed in kBtu per square foot per year. It is the workhorse metric and it comes in two forms:

• Site EUI — the energy delivered to the building, in the units it arrives in, converted to kBtu.
• Source EUI — site energy plus the generation and transmission losses upstream of the meter. Source EUI is what ENERGY STAR scores on, because it prevents a building from appearing efficient purely by using a fuel with high upstream losses.

The ENERGY STAR score is a 1-to-100 percentile against a national survey of comparable buildings, adjusted for size, occupancy, operating hours, climate and use type. A score of 50 is median; 75 or above is the certification threshold. It exists for around 20-plus property types and not for the rest.

What it takes to produce one

The data requirement is far lighter than the industry implies, and this is worth stating plainly:

• Twelve consecutive calendar months of whole-building consumption
• Every fuel — electricity, gas, steam, chilled water, fuel oil, propane
• Whole building, including tenant-metered load
• Monthly utility bills are sufficient. No interval data. No submetering.

Any vendor claiming submetering is required for an ENERGY STAR score is misrepresenting the requirement. The genuine gap in multi-tenant property is tenant-metered load missing from the whole-building total, and that is a real submetering case — but it is a coverage problem, not a resolution problem. See what is submetering.

Energy intensity beyond square footage

EUI normalises to floor area because that is what buildings have in common. For industrial facilities it is frequently the wrong denominator: a plant running three shifts has double the EUI of the same plant running one, and neither number says anything about efficiency.

Plants normalise to production: kWh per case, per pound, per part, per ton. That metric is far more useful and considerably harder to produce, because it requires line kWh and a production counter on the same clock, plus order and SKU context from an MES or ERP system. That last leg is where most deployments stop — it is an integration into a system the energy team does not own.

Where the deadlines and penalties are

More than 40 US jurisdictions now require benchmarking, building performance standards, or both. Two of them — Washington State and Philadelphia — require a documented ongoing operations-and-maintenance or energy-management programme rather than just a reported number.

Jurisdiction	Penalty	Timing
NYC Local Law 97	$268 per tCO2e over the limit, annually	Next step-down is 2030–2034, with a projected 57% of buildings over cap, against under 10% today
Boston BERDO	$234/tCO2e alternative compliance. Penalties split by size: ≥35,000 sq ft or ≥35 units → $1,000/day standards, $300/day reporting; 20,000–34,999 sq ft or 15–34 units → $300/day standards, $150/day reporting	2026 extended reporting deadline was 15 August 2026
Washington DC BEPS	—	Cycle 1 evaluation year began 1 January 2026; Cycle 2 begins 2028; CY2025 data was due 1 May 2026
Philadelphia BEPP	$2,000 missed-deadline fine	Tune-up reports for buildings over 200,000 sq ft due 30 September 2026
Washington State CEBS	up to $5,000 plus $1.50/sq ft annually	Tier 1, over 220,000 sq ft, passed 1 June 2026
Denver	$0.70 per kBtu unmet	—

Correcting a widespread error: Local Law 97 does not tighten in 2027. The next step-down is 2030. The 2030 figure is both accurate and more consequential, because the share of buildings over cap is projected to move from under 10% to about 57%.

Worked example

A 220,000 sq ft office building using 4,800,000 kWh of electricity and 12,000 therms of natural gas in a calendar year.

• Electricity: 4,800,000 kWh × 3.412 kBtu/kWh = 16,377,600 kBtu
• Gas: 12,000 therms × 100 kBtu/therm = 1,200,000 kBtu
• Total: 17,577,600 kBtu
• Site EUI = 17,577,600 ÷ 220,000 = 79.9 kBtu/sq ft/yr

Against a US office median in the region of 50–70 kBtu/sq ft, this building is above median and the filing will say so. What the filing does not say is why — and that is the limit of benchmarking as a discipline.

The argument benchmarking sets up

Whole-facility measurement and verification — IPMVP Option C — cannot reliably detect savings below roughly 10%, because the signal is buried in the building's own variability. A project that genuinely saves 6% is unprovable at the whole-building meter, which is exactly the situation a facilities director is in when defending last year's capital request.

Submetering the affected system converts that into an Option B problem: isolate the system, measure before and after, and the signal-to-noise ratio improves by an order of magnitude. That is the standards-grounded reason submetering exists, and it directly supports 179D's measurement pathway and the M&V obligation in every energy savings performance contract.

For M&V specifically, the binding constraint is completeness, not accuracy: at least twelve months of data at 90% or better coverage, a weather station within a defensible distance, and a documented log of non-routine events. Non-routine events destroy baselines far more often than weather does.

What energy benchmarking is not

It is not an energy audit. Benchmarking produces a number and a percentile. An audit produces a list of measures with costs and savings. A poor score is a reason to investigate, not a diagnosis.

A score is not a saving. Improving an ENERGY STAR score from 42 to 68 is a reporting outcome. It becomes a financial outcome only if the underlying consumption fell, and consumption falls only if something operational changed.

It does not require submetering or interval data. Twelve months of whole-building bills across all fuels satisfies the filing.

It is not emissions reporting. Building performance standards convert consumption to emissions using fixed per-fuel coefficients with no time-of-use grid factor. eGRID is annual with roughly a 13-month lag, and hourly Scope 2 matching is a proposed GHG Protocol revision with final publication expected in 2027. A platform's value against these laws is in reducing the number, not in reporting it.

Reporting is the easy half. Every jurisdiction in the table above prices the gap between the reported number and the required one. Filing on time avoids the administrative fine; it does nothing about the emissions penalty.