Skip to main content

Lessons from NYC Benchmarking

By September 17, 2012November 8th, 2021Energy Efficiency, Energy Rant, Utility Stuff

New York City recently completed its report for the benchmarking of all its “large” facilities, generally with square footage of 100,000 or greater.  The results of the study are not surprising.  You may be thinking, “Who cares about NYC?”  Answer: this post includes universal challenges with benchmarking whether it’s Batswana or the Yukon.

The benchmarking was completed using ENERGY STAR Portfolio Manager, which as far as I can tell ranks buildings by source Btu[1] per square foot, otherwise known as energy intensity.  For example, it uses a factor of 3 for electricity, which is one over the efficiency of delivering electrical Btus to the site.  In other words, it takes 3 units of energy from coal, natural gas, nuclear fuel, etc, to provide one unit (Btu) of electricity.


Finding #1:  Energy intensity varies all over the place, by factors of 4:1 to 7:1 depending on building type.  Buildings at the high end of energy intensity consumed 4 to 7 times more energy per square foot than miserly buildings.  This is not surprising and can be due to a myriad of factors.

  • First, energy intensity can vary by a factor of three between misers and hogs.
  • Second, some may be “cheating” because they provide no ventilation for their occupants.  You know, clammy caves.
  • Third, some buildings may not be fully occupied, although decent benchmarking should account for this, but it’s still a crap shoot.
  • Fourth, some buildings are difficult to classify – they may legitimately have five building types – retail, healthcare, residential, office space, and lodging in a single facility.
  • Fifth, there are plenty of places to screw up and make mistakes.  More on this later.

Finding #2:  Older buildings have lower energy intensities.  This is consistent with a white paper I wrote a little over a year ago and with a related paper presented at the American Council for an Energy Efficient Economy (ACEEE) Summer Study for Buildings last month.  Stunningly, the energy intensity by building vintage in five categories decreases in each step from newer buildings to older buildings.  Older buildings use less energy.  There are no hiccups.  The reason is, HVAC designers lost the forest amongst the trees, started going down paths of centralized heating and cooling systems that are terribly complex and poorly understood.  See the papers noted above.  Older buildings may, for example, have window air conditioners with horrible efficiency and old steam radiators, but they damn well have the radiators off when it’s hot as blazes outside and the window shaker is running.  This is typically not the case with the newer Rube Goldberg colossi HVAC systems.  It would be surprising to not have cooling at the central air handler and some heating going on at the occupied space even when it’s 85F outside and the humidity is thick as a stick of butter.

Finding #3:  NYC buildings are more efficient than others in the nation.  Haha, har har, hardy har har.  I don’t buy this at all.  The results may say so, but first the data set for comparison is almost 10 years old, from 2003.  But you just said Jeff, that new buildings waste more energy.  Yes, but not many buildings adding to the “new” (post 1960) have been built in that period, and secondly, energy use in buildings tends to rise with rising plug loads and associated energy required to offset those loads (cooling).

Finding #4:  Apartment buildings have a smoother and narrower (more uniform) distribution curve than office buildings.  Again, this gets back to the single zone systems one is likely to experience in apartment buildings versus the enormous variation of systems one would see in these office buildings.  The single zone systems provide heat or cooling and not both at the same time.  This results in two things: lower energy use and more predictable energy use.


Cited Challenge #1: Building square footage.  Yes, this can be challenging, but it is easy compared to other challenges that weren’t noted.  These include the usual cases where one meter – electric, steam, chilled water, and/or natural gas serve more than one building.  E.g., district metering.  How were these allocated?  Conversely, many times buildings have multiple meters, sometimes for logical reasons such as additions. At other times, a large gas supply to a boiler plant may be interruptible and there may be a second meter serving a kitchen, and that is on firm / non-interruptible.

Cited Challenge #2: About 100 consultants provided the benchmarking services and data.  Whoa!  Thirty firms did most of the buildings.  Nevertheless, I can all but guarantee some of these firms have no business providing these services.  Benchmarking is like chiropractic or physical therapy.  You need to read the symptoms (building characteristics), assess the patient (energy data and how it matches what you see and hear), and use subtle tests to confirm your findings – does it make sense?

Cited Challenge #3: A compilation of stuff I noted above with multiple buildings served by single plants, utility meter data by service address and not by facility (difficult to match meters to specific facilities), having a building address with no matching meter address, and so on.

Learning Curve

All credit to the auditors and authors.  They have recognized the data as a bit shoddy and unreliable and thus it will not be made public on a building by building case.  My humble suggestions would be to sharply reduce the number of benchmarking companies to a handful that really know what they are doing.  While it is always far easier to recommend than do, give them access to utility customer information systems so they can investigate meter discrepancies themselves.  This is not intern work or work for statisticians or MBAs.  It requires experts with a long track record of evaluating facility fingerprints and recognizing whether the energy data provided is the right data, all the data, and only the data that applies.

[1] Btu = British thermal unit which is the energy required to raise a pound of water (think pint of beer) one degree F.

Jeff Ihnen

Author Jeff Ihnen

More posts by Jeff Ihnen

Join the discussion One Comment

Leave a Reply