Tendril Networks has gone through a lot of changes in its decade in the home energy management business. But throughout its faltering as a provider of energy dashboards and displays, as well as its later reinvention as a software and data analytics provider, it's almost always had access to a key source of data: utility energy bills.
That made sense, since utilities make up most of Tendril’s customer base. But plenty of tasks on Tendril's menu rely on data sources beyond the meter. Demographic research and market segmentation, physics-based home energy profiling based on property records and consumer databases, and cost-benefit analysis of which customers make the best targets based on past behavior are a few examples of what it's been doing.
Tendril says it’s gotten pretty good at figuring out the energy essentials from these outside-the-meter sources -- so good, in fact, that it can do it for pretty much every address in the country. On Thursday, the company announced it's making that capability available to customers, through what it calls “zero”-data calibration.
The “zero” is meant to underscore the fact that it’s using no utility meter or billing data to deliver high-resolution understanding of each home’s energy profile. Most utility-connected “virtual audits” of this kind rely on at least thirteen months' worth of data to run the linear regression analyses that allow patterns and predictions to emerge for individual addresses or accounts, Chris Black, Tendril’s chief technology officer, said in an interview this week.
“We do it differently," he said. "We have a physics-based home simulation model that allows us to provide the same exact experience to homes that don’t have thirteen months of past data." Tendril uses public data, as well as consumer data from providers such as Experian and Axiom Data Corp., to flesh out its analysis, and runs more than 200 variables for each home in a typical model, he said.
“With this zero-data model, where all we know about the home was obtained through third parties and simulation, we’re achieving 70 percent accuracy,” in terms of predicting “what the home will consume, and how they will consume it from a disaggregated respect,” he said. “We see closer to 90 percent when we have data that’s supplied by the users,” such as web portal interactions or surveys, he said.
Tendril’s data can be useful even to utilities already crunching meter data for much finer-grained predictions of energy usage, he noted. That’s because it incorporates all that consumer behavior data, which allows it to avoid the mistakes that befall mass-marketing approaches standard for utilities today, like sending appliance rebates to people who have just installed a new kitchen set, distributing electric water-heater offers to people using natural gas, or asking renters to install insulation, he said.
“Customer satisfaction is the primary motivator, but it has the ancillary benefit of increasing the uptake of offers from our utility partners,” he said. “When the recommendation that’s made is very accurate, then they’re much more likely to take actions that do require some investment on their part…and our utility partners are able to sell more things to their customers.”
“I would say that in easily a third of all RFPs we get now, the main thrust is how can you help us in upselling and cross-selling across multiple programs,” he said. “It’s now, 'Tell us what you can do to segment our customer base, come up with new products and services, and help us solve these demand-side management problems we’re facing.'”
But the zero-data model also offers non-utility players a view of energy usage that’s hard to obtain without having a utility as a partner. Specifically, “If you’re asolarcompany, you won’t have that kind of access to utility data,” Black said. “Solar is a great example of that, and you’ll see an announcement along those lines coming later,” he said, though he wouldn’t go into more details.
It will be interesting to see how Tendril’s new capabilities might complement the work that’s already underway on that front. Third-party solar providers like SolarCity, Sunrun, Vivint and NRG Energy have been spending lots of money to acquire marketing and customer acquisition startups. At the same time, one of Tendril’s biggest competitors, Opower, plans to incorporate solar PV, electric vehicle chargers and other grid-edge devices in its new Opower 5 platform.
According to industry buzz, Tendril has won some utility contracts lately against competitors like Opower on the strength of its new capabilities. Black wouldn’t name any names, but he did say that Tendril recently won a contract against a competitor, based on indications that it’s able to deliver a 30 percent greater return in terms of efficiency per dollar spent. Duke Energy, another customer, said it got a fourfold increase in customers responding to mailers after Tendril took over, he said.
Even so, there’s plenty of uncertainty out there about the value that big data capabilities like these can bring to the broader push to make energy efficiency more efficient, so to speak. There’s also plenty of competition, with companies like Opower, C3 Energy, Oracle, Silver Spring Networks and others offering home energy analytics, and startups like FirstFuel, Retroficiency, Lucid, Gridium and Pulse Energy doing similar work on the commercial and industrial front.