top of page


  • Writer's pictureMirka Karra

Debunking myths about accuracy in energy disaggregation

What we often come across as a common focal point in discussions around energy disaggregation technology with utility professionals is accuracy. Utility professionals often wonder about the specifics: "Can you tell when someone uses a kettle or a hairdryer?" Even though the answers to these questions might satisfy curiosity, more often than not, they are not what leads consumers to reduce their energy consumption. Based on our experience, the important thing is not whether the kettle is detected or not, but whether the energy insights provided are useful and personalized enough to enable consumers to take action to reduce energy consumption and costs.

Debunking myths about accuracy in energy disaggregation

Reality check

Even with widespread smart meter implementation across Europe, the U.S., and Australia, we're not yet at a point where this technology can offer disaggregation based on high resolution data which would enable energy disaggregation on appliance level. This would require a smart meter reader or advanced metering infrastructure equipped with Distributed Intelligence technology. Typically, the lower resolution data coming out of smart meters every 15, 30, or 60 minutes can offer energy disaggregation on activity level (heating, cooking etc.) and not on appliance level (air conditioner, stove etc.). Is that a problem for consumers who would like to understand how to reduce their energy consumption and costs? Based on our experience, the answer is no.  

The mere fact that consumers who previously were kept in the dark with regards to what they would pay for in their energy bill can now have an understanding of how they consume energy at home is already a significant step towards helping them save energy and lower costs. It is the actionable accuracy that moves the needle and not the absolute accuracy in low resolution energy disaggregation.  One example of actionable accuracy is the AlwaysON category consumption which refers to electricity consumed by appliances which either appear to be off or are on standby.

The AlwaysON category

Surprisingly enough, a lot of appliances, no matter the status they want you to think they’re in, ultimately, they’re on and add to the electric bill without any practical reason.

The convenient thing is that the greatest part of the AlwaysON consumption, which  in most households concerns computers and peripherals, as well as home entertainment appliances, can be easily reduced or even eliminated by using a power strip with switches. Plugging these appliances into power strips that can be switched ON/OFF allow for reducing AlwaysON consumption simply with one action.

Receiving information on your AlwaysON category, coupled with such a recommendation on how to reduce consumption, is already a concrete step that enables the consumer to reduce electricity consumption in this category.

What about high resolution data disaggregation?

Τhere are instances when appliance level disaggregation, based on high resolution data, is feasible, for example, when there is extra hardware in place or advanced metering infrastructure equipped with Distributed Intelligence technology (see Itron's next-generation smart meters, Riva). This capability allows for immediate detection such as identifying when electric vehicles begin charging. Utilities employing such technologies can offer innovative services, like Automated Demand Response programs or real-time energy consumption advice, similar to NeN Energia's project with II Robo, which provides tailored, timely recommendations to customers the moment they are about to exceed their power threshold. 

The strategic value of actionable disaggregation

Although real time data services are the destination, most utilities have not yet laid the groundwork for getting there. In  pursuit of such a goal, a utility that wants to stay competitive and cater to the evolving customer needs through personalization and digitization might as well offer services based on low resolution data leveraging the widely available smart meter.


bottom of page