64 votes



This idea is active.
App/Services "Killer Ideas" »

Energy Waste ('Crash') Report: Generating Key Behavioral Data

While human behavior undoubtedly drives building energy waste, this problem is rarely quantified -- and often ignored.

Similar to 'crash reports' that help fix software bugs, an app must be developed to generate vital behavioral and operational info during periods of anomalous energy use. This will create a compelling and urgent rationale for occupants and facility managers to disclose details about behavior, ease the transfer of this info, develop a wealth of 'behind the meter' data, and craft new energy efficiency solutions.


Operational energy waste is all around us: buildings are heated and cooled with no one in them; computers stay 'idle' instead of 'sleeping;' we forget to turn off the lights. Estimated behavioral savings range from 5-30%+, but, critically, the data is sparse. Little is known about the scale of the problem -- or the relationships between the countless variables.

Most efficiency efforts revolve around the easily identifiable, from sustainable design to efficient HVAC and lighting. That's not enough. Regardless of system design, efficiency, or ability to automate, human behavior remains highly impactful. The basic thermostat is illustrative: no matter how advanced, a human being decides the temperature set point; even if capable of intelligent scheduling, there’s always a 'manual hold' function.

Meter data is essential, but we're missing a key piece of the puzzle: What's going on behind the meter?


It's a familiar IT experience: Software crashes, restarts, and you're asked to 'send report' to help fix the bug. This data (about operating system, application specifics, etc.) is then evaluated, previously unknown issues are identified, and the problem is fixed.

The energy space needs a similar tool. With simple, voluntary, and potentially anonymous reports about behavior when energy consumption spikes -- and with basic regression analysis against Green Button and other available data -- patterns of energy waste will surface.


AppA: Aggregation of anonymous behavioral data. Potentially led by, or in partnership with, DOE.

AppB: Customer-focused approach with individual solutions. Led by private sector.


1. Flag unusual and out-of-bounds energy use.

2. Alert user, via mobile device and email, to send report. Examples of possible info: occupancy and space use (self-generated or automated through geolocation, calendar integration, etc.); temperature set point; occupant comfort (e.g. too cold/hot?); plug-load (e.g. computers, mobile chargers, etc.)

3. AppA: Combine anonymous behavioral info with Green Button, weather, and other data; aggregate and analyze; identify relationships and patterns. AppB: Uncover user-specific inefficiencies.

4. Pick up fruit on the ground.

Submitted by Sam Brooks 9 months ago

Vote Activity Show

(latest 20 votes)

Comments (9)

  1. winner. something actually doable, and addresses major issue

    9 months ago
  2. An app I can actually understand and look forward to using. #1 for me.

    9 months ago
  3. Great idea Sam! Maybe we can persuade our local Metro Washington utilities to support and participate.

    9 months ago
  4. I like this idea a lot. It would also be interesting if you could include some kind of energy demand/use threshold alert for operations/facilities managers to utilize on this app. If a threshold for occupied and unnoccupied hours was determined for each building, you could set an alarm to send a notification via text or email to the operations manager notifying them that something is out of the ordinary. For example, if the HVAC system was not programmed for a holiday and the building systems came on anyways, or if your cleaners left all of the lights on in the facility after hours, the operations manager would be alerted and would be able to log in remotely (if your BMS has that capability) or come in to turn the system off and save potential wasted energy. Then that threshold information could be sent in anonymously disclosing the peak energy amount used. the DOE could cross reference that info with Energy Star as well. Though this would likely only be doable in your AppB scenario..just a thought.

    9 months ago
  5. Great idea. Data usually collected from building Automation System is entirely focused on system / technological failures…incorporating human and operational info would be a huge plus to identify and predict energy waste.

    9 months ago
  6. Big ROI on this one.

    9 months ago
  7. Love the idea and the proposed use case. But what data are you exactly using to do it?

    9 months ago
  8. Great idea! Let's rally more folks to vote this one up!!

    8 months ago
  9. Sam Brooks Idea Submitter

    Hey kuenley: The data that'd be used to identify anomalous / spiked energy use would be Green Button -- that is, electricity usage data, at intervals (usually 15 minutes), that's streamed in real- or near-real-time.

    Through APIs already in development (both by utilities and third-party developers), this information can be analyzed, relatively easily, to identify when energy consumption is anomalous. Then, when analyzed against key behavioral and operational information (as opposed to system (e.g. HVAC system failures) info traditionally tracked by building automation systems) that's created through the user-generated "crash reports,” patterns of behavioral/operational waste can be identified -- both in aggregate and for individual buildings/homes -- and addressed.

    8 months ago