Energy disaggregation technologies are unique in their specific methods of parsing out the energy usage of individual loads like air conditioners, appliances, lights and electronics, simply by analyzing the entire home or building’s electricity data. But in broad terms, they tend to fall into two categories: low cost and lower accuracy, or high cost and higher accuracy.  

Cambridge, Mass.-based startup Sense has definitely chosen the latter route. Its home energy monitor costs $299 and requires skilled installation of induction clamps on a home’s electrical mains to pull the sub-second data that it needs. That hasn’t stopped it from selling tens of thousands of units to date, mostly direct to DIY homeowners — but it is a stiff barrier to mass-market adoption. 

On the other hand, its in-home, internet-connected hardware allows Sense’s speech-recognition-based disaggregation technology to parse out one to two dozen different loads in real time, down to garage doors opening, refrigerator compressors failing, or appliances being used. That’s a level of accuracy unmatched by lower-cost disaggregation based on smart meter interval or other sampled data. And unlike historical analyses, Sense knows the second it happens, and can tell its customers via web and mobile alerts, if they’ve asked for them. 

Michael Phillips, Sense co-founder and CEO, believes this kind of real-time, highly accurate monitoring capability can do a lot more than just inform homeowner energy decisions. In his view, it could be one of the foundations of the smart home.

Phillips was quick to highlight that the lead investor in its $18 million Series B round that closed early this month is Schneider Electric, the French electricity and grid technology giant. “We’re not talking about specific product plans yet,” he said in a Wednesday interview. But “you can see how Sense plus Schneider could have potential. I feel we have a very aligned vision around how energy systems should become smart.” 

Schneider makes a wide range of low-voltage electrical equipment, from circuit breakers and switches to energy storage and industrial control systems, and also sells smart thermostats under its Wiser home automation brand. These are the kinds of embedded building electricity infrastructure that could become the points of contact for Sense’s disaggregation data-collection and analysis — and that could significantly cut the upfront cost of getting the technology into homes and businesses. 

“We are doing this with a retrofit that goes inside your electrical panel today, because that’s the only way to get the signals we need,” he said. “But the processing we do ought to be built into the core infrastructure. It could live either in the panel or the utility meter — these are the two places that power comes into your house.” 

"Once it’s built into the infrastructure, that cost goes away,” he said. “The incremental cost of silicon, of computation and sampling, we’re riding projections that these things are going down to zero, essentially. If we can trade off a bit of computation for a bit of energy, that’s a good tradeoff in the long run.”

Sense is not pitching itself to homeowners as an energy efficiency tool, he stressed. Instead, “the real-time aspect is the key thing for engagement. You can open this app on your phone and see what’s happening right now in your house.” This has helped drive “really good engagement numbers” among the tens of thousands of Sense owners out there, he said. About half are still using the service at least once a week, and of those, the average user is checking it 12 times per week, or “almost twice a day,” he said — most of them to check if they’ve left things on when they’re at work, or to respond to alerts of unexpected energy usage like an opening garage door. 

“We want to further engage those engaged users” in energy efficiency, he said. “Once they’ve used the app to check if their garage door is open or their oven is on, we’re hoping they’ll trust us when we say, 'Your air conditioner needs maintenance'” — an insight that Sense can glean from changes in the electrical signal of a failing compressor, one big source of inefficient cooling in homes. 

To date, Sense’s most extensive test of energy-efficiency-driving capabilities has been with Efficiency Vermont, which reported in July that a six-month trial in select homes had yielded improvements of about 8 percent, or $100 per year.

But Phillips believes that Sense’s detailed data can drive much deeper cuts in wasted energy, starting with encouraging homeowners and equipment makers to curb "vampire loads" drawn by set-top boxes, televisions and other consumer electronics, and extending into more use of its fault-detection capabilities. "That’s an important value proposition and an opportunity for the utility," said Phillips.

Sense’s new round brings its total funding to $38.6 million, including its $12 million Series A round in 2016 and about $5.6 million raised before that. Previous investors that joined the new round include Prelude Ventures, Capricorn Investment Group, Shell Ventures, iRobot, and Energy Impact Partners, the utility-backed investment fund that’s put more than $200 million into 20 startups to date. 

Sense’s relationship with Energy Impact Partners means the company is "spending time with utilities to see how the consumer-facing implementations we have, and the data behind the screen, can be useful as utilities increasingly try to manage when energy is used,” he added. 

Phillips stressed that Sense’s real-time analytics technology isn’t actively controlling in-home devices. But it is integrating with key home automation players, including Amazon Alexa since June 2017 and Philips Hue starting this July. More integrations are coming, he said. 

We’ve been tracking a number of efforts in high-fidelity energy disaggregation, some stretching back more than a decade. These include Intel Labs’ experiments with in-home sensoring technology, and Belkin’s integration of Zensi, a startup it acquired in 2010, into its WeMo line of smart home products. Neither effort seems to have yielded publicly announced advances in putting disaggregation to use in a commercial context. 

In the past five years, other startups, such as San Francisco-based Verdigris, Belgium-based Smappee and Vancouver, Canada-based Neurio, have promised technology that can yield accurate disaggregation of household loads as well. More recent entries include U.K.-based startups Verv and Ecoisme and Swedish startup Watty.