Benchmarking a building portfolio: where the numbers come from

Benchmarking a portfolio looks like an analysis exercise and is mostly a data exercise. The formula is not the hard part. Getting twelve complete months for every meter in every building, on one calendar, with floor areas that match what is actually conditioned, is the hard part. When a benchmark gets challenged in a board meeting, the challenge is almost never about the method.

This article covers what a portfolio benchmark answers, the inputs it depends on, and how to read the result without overclaiming.

What benchmarking answers

Two questions, and they need different comparisons.

• Against itself: is this building using more or less than the same building used last year, after weather and billing calendar are accounted for. This is the question an operations team can act on.
• Against its peers: is this building's intensity reasonable for its type, size and climate. This is the question an owner, lender or board asks.

Mixing them produces the familiar bad slide: a building that improved 6 percent year over year presented as underperforming because it sits below the median for its type, or the reverse. Both numbers are true and they answer different questions.

Getting complete utility data first

Completeness is the input that decides whether the rest is worth doing. A portfolio benchmark built on eleven of twelve months, or on nine of eleven accounts, is not a conservative estimate. It is wrong in an unknown direction, because the missing pieces are rarely random: they are the account that changed hands, the site that switched to a new retailer, the water bill nobody forwards.

Two routes supply the data. Bill capture gives billed consumption and cost per account, which is what the benchmark needs. Green Button Connect My Data, where the utility offers it, delivers meter data directly once the account holder authorizes it. Getting both into one structured store on one calendar is the work described in energy data management for a portfolio.

Normalizing for weather and area

Two adjustments make year over year comparison meaningful. Weather, because a cold February moves gas consumption regardless of how the building is run, and area, because intensity is consumption per square foot.

Weather normalization works from degree days. Environment and Climate Change Canada defines heating and cooling degree days in its climate glossary as the difference between the mean daily temperature and a reference temperature of 18 degrees Celsius, summed over the period. A regression of consumption against degree days over a baseline period gives an expected consumption for any later month, and the difference between expected and actual is the part that is about the building. The mechanics, including cumulative sum tracking, are in weather normalization, CUSUM and RETScreen.

Floor area sounds simple and is the most common quiet error in a portfolio. Gross versus rentable, whether parking and unconditioned storage are included, what happens to a partially occupied floor. The number itself matters less than using the same definition for every building and every year, and recording which definition was used.

Portfolio Manager and the Canadian comparison

For peer comparison, ENERGY STAR Portfolio Manager is the common language in North America. It expresses energy use intensity as energy per square foot per year, converting whatever units you enter, and computes both site and source intensity.

The 1 to 100 score is the part people quote. It compares a building against similar buildings nationally: 50 is median performance and 75 or higher indicates a top performer that may be eligible for certification. The scoring is based on source energy and accounts for weather and property use details. For a building in Canada, the peer group comes from the Survey on Commercial and Institutional Energy Use, commissioned by Natural Resources Canada and implemented by Statistics Canada, not from the United States survey, which is why a Canadian score is not simply a United States score with a different postal code. Natural Resources Canada publishes how the score works for Canadian buildings, and our guide covers Portfolio Manager for Canadian multi-residential buildings in detail.

Not every property type has a score. Where there is none, the honest approach is peer comparison within your own portfolio plus year over year performance, stated as such, rather than borrowing a score from a type that does not fit.

Reading the result honestly

A few habits keep a benchmark defensible.

• State the period and the completeness. "Twelve months to August 2026, all 14 meters complete" is a different claim from "the last year or so".
• Separate weather-normalized change from raw change, and say which one you are quoting.
• Keep cost and energy apart. Cost moved because of price is not a performance result, which is why cost intensity needs its own treatment.
• Name the floor area definition once and reuse it.
• Show the exceptions. A benchmark with three estimated reads in it is still useful if the estimates are visible.

Where the benchmark leads is a short list of buildings worth attention, which is the output a monthly energy review report is built around. Emissions follow the same data, with published factors applied to the same validated consumption, as covered in GHG emissions from utility data.

FAQ

How many months of data do we need before benchmarking?

Twelve consecutive complete months per meter for a first benchmark, because a partial year cannot be weather adjusted sensibly and seasonal buildings will mislead. Twenty-four months lets you compare two full cycles and is what several reporting and certification programs expect.

Our buildings are mixed use. Does one intensity work?

No. Portfolio Manager scores by property use, so retail at grade with apartments above needs both uses and their areas entered. A single blended intensity hides the part that is performing badly. Keep uses and areas on file with the meters so they are entered the same way each year.

Can we benchmark without ENERGY STAR?

Yes. Internal benchmarking against your own portfolio and your own history is often more actionable, particularly for property types without a score. Portfolio Manager matters when an external party asks for a common language, which includes several municipal reporting programs.

Who keeps the numbers current after the first exercise?

That is the difference between a benchmarking project and a benchmarking process. In a monthly review, each month's bills are captured, validated and added, so the annual benchmark is an export rather than a project. Request a review to see what that looks like across your buildings, or read the use cases for how portfolios of different shapes use it.