By Margaret Lindquist | October 2021
Many utilities are setting aggressive decarbonization or net-zero goals—more than 30 United States utilities covering 78% of the population have set these targets—but meeting such goals requires more than finding the right new technology and energy sources. Decarbonization also must address affordability, to ensure that limited-income residents don’t bear too big a burden as utilities make expensive investments in the electrical grid.
At the same time, however, COVID-19 has made the affordability risks even greater. In a recent survey by Untold Insights for Oracle Utilities, 23% of respondents said that last year they were—for the first time—unable to pay a utility bill. Fifty-six percent were not aware that programs were available to help.
“Limited-income households spend nearly five times more of their budget on energy utility bills than the average household,” says Jameela Belyeu, product strategist for Oracle Utilities Opower team. As utilities make expensive infrastructure upgrades, which are funded by the price of energy, related to decarbonization goals, they know they also need to improve affordability for limited-income customers. “We just can't let limited-income customers foot the bill for a clean energy system,” Belyeu says.
Utility companies have three challenges in aiding their limited-income customers: identifying them, making them aware of programs and support they qualify for, and reducing the hassle of enrolling in those programs.
Solving these problems isn’t a “nice to have” for utilities. Many states have laws or regulations that mandate support for limited-income customers. For example, Pennsylvania has specific goals that utilities have to meet by a certain date, and in Michigan energy-efficiency programs for limited-income customers must be supported by all ratepayers.
To reach these people, utilities need to layer datasets that identify limited-income customers and use AI to predict eligibility and deliver personalized program offers at the right time. That might mean sending an email or SMS offer of support, or providing a one-stop shop on the web that makes it easy for customers to take advantage of the array of available programs. Here are some of the ways utilities are using AI and behavioral science to help support limited-income customers.
Finding the customers. Limited-income customers aren’t using more energy than the average consumer or paying higher rates. But they’re more likely to face late fees, or reconnection charges if their power is turned off due to nonpayment. There are numerous bill assistance and energy efficiency programs available, run by utilities, state and federal government, and community organizations, but the sheer number of those programs makes it difficult for consumers to find the specific program that can help them.
“Utilities face real challenges in just getting their income-qualified customers the assistance that's available to them, and that starts with finding them,” says Paul McDonald, senior director of product strategy for Oracle Utilities Opower. Utilities often look at basic third-party demographic data, but those using Opower can segment their customers using data from the U.S. Census Bureau and other reputable government sources such as the Energy Information Administration and Department of Energy. That’s all in the service of identifying and reaching out to customers who might be at risk and might qualify for different types of aid. Right now, the Opower team is working on new applications of deep learning—the most sophisticated machine learning technique in the world—to use all available data to predict which customers need and qualify for assistance.
“Utilities face real challenges in just getting their income-qualified customers the assistance that’s available to them, and that starts with finding them.”
Connecting customers with the right programs. The next challenge is making customers aware of the programs that can help. “Not a lot of us think about energy,” says McDonald. “Especially for limited-income households, there are a lot more urgent things to do.” Utilities are using Opower to provide customers with weekly, personalized outreach that includes bill updates, program offers, and AI-powered recommendations for customers to save energy and money.
Making it easy for customers to sign up. Research from the American Council for an Energy-Efficient Economy shows that limited-income households in the US spend an average of 8.1% of their household income on utilities, compared to 2.3% for non-low-income households. A quarter of limited-income households spend as much as 15% of their income on energy. But it’s not easy for many utilities to enroll customers in programs to help them, even though it’s critical. Because programs originate with multiple entities—utilities, government, and nonprofits—finding the right program can be challenging. “These programs take a lot of paperwork,” says McDonald. “The result is this fractured, time-consuming, discouraging experience.” What’s needed is a single place for customers to see what’s available and start the enrollment process. “We’ve created a one-stop shop where a customer can view programs they are eligible for,” says Belyeu. “It generates a personalized, curated list for each customer right within their utility’s web portal.” Opower is working on functionality to create a seamless experience for customers to enroll in those programs.
Even without government mandates, utilities are committed to helping limited-income customers. Ratepayers will end up supporting the decarbonization programs that are part of most utilities’ roadmaps. Since many of the programs—around household solar installations, for example, or electric vehicles—benefit more affluent customers most immediately, the burden can’t be allowed to fall disproportionately on limited-income customers, who comprise around 30% of all residential customers. Utilities know they must address this risk. “Utilities aren’t going to be able to meet their decarbonization goals unless that 30% is part of the strategy,” says Belyeu.
Illustration: Wes Rowell