How much energy does a hotel use? Benchmarks by hotel type and climate

What question is your hotel benchmark supposed to answer?

Start by choosing the decision. An owner budgeting utilities needs quantities and prices. An engineering team looking for waste needs equipment and schedule information. A portfolio manager comparing buildings needs consistent boundaries and peers. A single national average cannot answer all three questions, even if the number is presented with several decimal places.

Use the bills and costs hub to separate physical energy from spending. Higher dollars do not necessarily mean higher consumption, and higher consumption does not automatically imply avoidable waste. A hotel may expand its kitchen service, add EV charging, bring laundry in-house, or experience a more demanding heating season.

Write a short benchmark brief: the property boundary, included fuels, measurement period, metric, peer population, and intended action. This prevents the comparison from changing mid-analysis. If the eventual finding is simply that more detail is required, that is still a useful result when the next investigation is clear.

Which energy metrics belong in the comparison?

Energy use intensity, or EUI, expresses annual energy use relative to building floor area. It helps compare different building sizes, provided the energy and area boundaries are consistent. EPA’s explanation of EUI distinguishes site and source metrics, which should be identified explicitly in every comparison.

If a supplier quotes electricity alone, use electricity quantities for the purchase analysis. If ownership wants whole-building energy performance, include gas and other material fuels. Preserve the raw quantities so unit conversions remain auditable and a later equipment decision can return to the relevant fuel.

What do public lodging datasets actually represent?

EIA’s 2018 CBECS lodging page distinguishes hotels, motels or inns, dormitories, nursing or assisted-living properties, and other lodging. A figure labeled lodging therefore does not necessarily describe hotels alone. Identify the reference year and subcategory before carrying any published result into a management presentation.

A national building-stock dataset answers a different question from a collection of properties voluntarily entered into a benchmarking platform. The former can describe the survey’s building population; the latter can help explore characteristics of measured benchmarked properties. Both can be useful, but neither makes an unqualified hotel cost promise.

Do not recast a published energy-use statistic as a current retail energy price. Energy demand reflects buildings and operations, while a supplier quote reflects an account, delivery period, product, and market. Keep benchmark research in the operating analysis and actual supplier offers in the procurement analysis.

How should hotel type and amenities change peer selection?

Separate properties by meaningful service differences rather than by a brand label alone. Limited-service hotels usually have a different equipment boundary from full-service hotels with substantial food preparation, banquet operations, laundry, spas, and pools. Two properties with the same room count can have very different common-area and back-of-house energy needs.

Record which services are on-site, separately metered, outsourced, or seasonally closed. A hotel that outsources laundry shifts part of its energy use outside the building; it has not necessarily achieved the same improvement as efficient on-site laundry. A restaurant tenant can create a similar boundary issue if its meter is included at one property and excluded at another.

Meeting activity also deserves context. A large banquet event can increase kitchen, ventilation, lighting, and hot-water use while contributing few occupied room-nights. See the cost per occupied room guide for why an occupancy denominator should be accompanied by operational indicators.

How do climate and weather enter without hiding poor operation?

Climate describes the broader environment; weather describes the measured period. Use both. A cold-climate hotel and a humid coastal property face different heating and moisture-control demands. Within one hotel, a mild year and an extreme year can create different quantities even when schedules and equipment remain unchanged.

Weather normalization can improve comparisons, but retain the unadjusted figures beside the modeled result. State the model, included variables, and limitations. A model that ignores a renovation or a newly operated pool can mistakenly assign that change to weather. Engineering observations help explain when the building itself changed.

Humidity management can matter even when outdoor temperature looks comfortable. Do not assume that less cooling runtime is always better performance. The hotel’s indoor conditions, ventilation requirements, and equipment capabilities must guide any operating change. Benchmarking identifies questions; it does not authorize aggressive setbacks in occupied spaces.

What does a transparent illustrative calculation look like?

Assume a hotel uses 1,000,000 kWh of electricity and 30,000 therms of natural gas during a complete year, across 80,000 square feet. EIA’s Btu conversion reference gives 3,412 Btu per kWh and 100,000 Btu per therm. These physical conversion constants are distinct from any energy price.

The assumed electricity contributes 3,412,000 kBtu and gas contributes 3,000,000 kBtu. Total site energy is 6,412,000 kBtu. Dividing by 80,000 square feet gives 80.15 kBtu per square foot for this illustrative boundary. The result is a worked example, not a recommended target or a measured property’s performance.

If the reported area omitted a restaurant while its meter remained included, the ratio would be overstated. If the gas meter was missing, the ratio would be understated. Boundary errors can be larger than the operational improvement being evaluated, which is why data validation precedes ranking.

How can you use an explorer without overstating the result?

EPA’s Portfolio Manager Data Explorer offers filters for property characteristics. Use those filters to investigate a relevant comparison group and retain the choices with the downloaded result. Check whether the remaining sample actually represents your hotel’s size, location, and use pattern.

An extremely narrow group can leave few observations. A very broad group can mask major operating differences. Present that tradeoff openly and avoid claiming a precise percentile when the data or methodology do not support it. Public comparison tools are most helpful when they make the next question better, not when they provide a flattering rank.

Retain the search date and exported dataset with the benchmark file. If the platform later changes its available population or filters, a saved extract allows the team to explain why a comparison changed even though the hotel’s measured energy did not.

What should happen after the first benchmark?

1. Verify all meters, fuels, floor area, and twelve-month coverage.
2. Compare the property with its own consistent prior periods.
3. Select peers based on services, climate, and operating boundary.
4. Investigate the largest unexplained end-use or schedule differences.
5. Establish measured project baselines before assigning savings.
6. Revisit the benchmark when operations or the boundary change.

Move consumption assumptions into next year’s energy budget, while keeping price scenarios separate. Review tariff selection when load shape changes, because operational changes can affect cost beyond total energy. A benchmark becomes valuable when it guides a specific, testable decision with the hotel’s actual data.