In the past year, you’ve probably heard something about machine learning. This branch of science, which involves crunching massive datasets to find hidden patterns, is helping companies solve problems that used to be unsolvable. Machine learning algorithms keep spam out of your inbox, and sound an early warning when someone else might be using your credit card. Down the road, they might save your life
. At their core, machine learning tools capture lots of complex information, learn from it, then apply what they learn to better estimate unknowns and predict future events. As the keepers of enormous datasets that defy conventional analysis, utilities could benefit from machine learning in a big way. Here are seven fundamental business challenges it could help them solve.
1. Uncovering hidden energy use patterns
Utilities are under pressure to personalize the customer experience. Generic mail won’t cut it in 2015: consumers worldwide expect companies to deliver carefully tailored insights and offers, and utilities are no exception. Machine learning can help. At Opower, we’ve used it to identify hidden trends in customers’ energy habits, segment customers by their individual energy behavior, then target those customers with information that’s actually relevant to them. For example, people who aren’t home during the day can benefit from programmable thermostats. Those who are home are primed for demand response programs. Unsupervised machine learning
makes it possible. We start with average energy usage at each hour of the day for hundreds of thousands of customers. Plotting this data all at once, you get an undecipherable hairball:
But when you apply a clustering
technique to the data, you discover that usage patterns generally correspond to one of a few archetypes, or “energy personalities
Utilities can and should treat their customers differently based on their energy personalities. When they do, marketing dollars go to better use, demand-side management results improve, and homes and businesses get a deeply personalized experience.
2. Getting more people to enroll in utility programs
Participation rates in energy efficiency and demand response programs are notoriously low. Traditionally, utilities have tried to boost them by breaking customers into a handful of hand-picked demographic segments, then marketing different programs to different segments. High-income customers might be more likely to take advantage of an expensive appliance rebate program, for example. Supervised machine learning
offers a better approach. If you know which customers participated in which programs in the past, you can train a model to make accurate predictions about who will participate in the future — and automatically optimize program targeting. Initial research suggests that by combining static information (like income level and household square footage) with behavioral information (energy usage, previous utility interactions, and so on), machine learning models could ultimately raise program participation rates as much as 20 percent.
3. Identifying untapped energy efficiency opportunities
Behavioral messaging can make a powerful impact
on energy efficiency. But if you really want to motivate people to save energy, you need to tie it to personalized insights and analysis. Breaking down a home’s energy consumption by appliance — air conditioning, water heating, and so on — can help customers understand their energy behavior and pinpoint their biggest savings opportunities. That’s why we wrote a usage disaggregation algorithm, which takes in meter reads, weather data, household characteristics, and other variables to create a personalized profile of a customer’s home energy use — and corresponding efficiency advice. The end result is a chart which can be embedded everywhere from a customer web portal to the utility bill. Best of all, the algorithm even works for customers who don’t have smart meters.
4. Determining how customers heat their homes
Accurate, personalized energy insights boost customer engagement. The reverse may also be true: homes and businesses that receive the wrong advice could be more likely to disengage, and tune out their utilities. One place that comes into play is home heating. Utilities don’t always know how customers heat their homes — and while natural gas efficiency tips are helpful for people with gas heaters, they might be a turn-off for those with electric ones. Machine learning techniques can ensure that customers get the right advice every time. By analyzing load curves from homes with known heating types, an algorithm can consistently predict customers’ heating hardware based on usage data alone.
5. Optimizing thermostat setpoints
In the same way, utilities can also feed smart meter data and weather patterns into a machine learning algorithm to estimate how homes and businesses are setting their thermostats. The algorithm’s output might look something like this:
Utilities can apply thermostat setpoint estimates, which don’t require any hardware in the home, toward a variety of ends. They might segment homes with inefficient setpoints, and offer personalized savings advice — including how much money customers could save by choosing a more efficient setpoint. Alternatively, they could use setpoint estimates and weather data to deliver bill forecasts halfway through the billing cycle. Or they could target homes with inefficient afternoon setpoints for demand response programs, like this:
6. Integrating electric vehicles into the grid
Utilities have a huge incentive to know when electric vehicles — the largest home appliances ever — are plugged into the grid. In the immediate term, it’s a customer engagement opportunity: EV detection would allow utilities to suggest time-of-use rates and encourage overnight charging. Down the road, electric cars could also prove to be an important demand response resource
. Last year, we analyzed the EV owners’ signature energy usage patterns
, which are represented below. In the months since, we’ve developed machine learning algorithms that help utilities detect when their customers plug a new EV into the grid.
7. Keeping customers for the long haul
In Europe and around the world, utilities in competitive markets are working hard to cut churn and win their customers’ loyalty. Helpful, personalized services like high bill alerts and insightful call centers are giving them the tools to succeed. Machine learning can help utilities take customer care a step further — helping them identify homes at greatest risk of switching providers, listen to their concerns, and offer solutions. A variety of industries — banking, insurance, and telecommunications among them — are already using churn modeling to deliver better customer experiences. By feeding customer characteristics, behavior, and billing information into a machine learning tool, utilities can start doing the same.
This article first appeared in Intelligent Utility on March 23, 2015.