The Behavior Change Podcast by Lirio explores the various ways humans can leverage behavioral science to personalize our messaging, engage our audience, and drive better behavior at scale.

 

Guest: Beth Karlin, See Change Institute

Host: Greg Stielstra (GS), Senior Director of Behavioral Science at Lirio

Summary: Beth Karlin of the See Change Institute shares her findings with Lirio. Listen for a brief description of loss aversion and learn how behavioral science can be applied in real life to behavioral energy programs to influence energy users to build better habits.

 

Subscribe to The Behavior Change Podcast

 

Transcript

Greg Stielstra:
Hello and welcome to the Behavior Change Podcast by Lirio, the program where we explore the marvels of behavioral science and ways of applying it to make a better world. I’m your host, Greg Stielstra.

On today’s show we’ll talk with Beth Karlin. Beth is a Research Psychologist who specializes in the role of technology and new media in behavior change with a focus on residential energy, health, science, and education. She currently holds a position as a Research Fellow at USC’s Norman Lear Center. She’s the Director of Transformational Media Lab at UC Irvine and she leads a consulting organization called the See Change Institute where she helps clients, including Lirio, apply behavioral science to improving energy efficiency and healthcare engagement. She has a lot on her plate, but she made time to talk to us.

We’ll hear from Beth in a moment, but as always, we’ll begin with a Bias Brief.

Lirio Bias Brief number 24, loss aversion. Will you take this bet? I’ll flip a coin and if it’s heads, you lose $100. But if it’s tails, you win $100. Most people decline, but why? The odds are equal, 50/50. So are the stakes, $100 either way. Someone who experiences gains and losses equally would’ve said yes. Okay, how about this? Heads, you lose $100. Tails, you win $200. Ah, now we’re talking. Most people take the second bet because the reward for winning is twice the penalty of losing.

Psychologist Amos Tversky and Daniel Kahneman ran that experiment to test something they called, prospect theory. They discovered that humans experience gains and losses subjectively and the pain from a loss is more than twice as powerful as the pleasure of an equivalent gain. This difference makes us loss averse. Gains and losses are not absolute. Whether we perceive something as a gain or a loss depends on how it is presented or framed, specifically, what we use as a reference, or an anchor for comparison. So, most gains can be reframed as a loss and vice versa.

If a gas station offers a regular price for credit and a discount for cash, you’ll perceive the cash price as a gain and the credit price as normal. If it offers a regular price for cash and a higher charge for credit, you’ll view the cash price as normal and the credit price as a loss, even though the cash and credit prices remain the same in both scenarios. Almost every gain scenario can be reframed as loss avoidance, multiplying its impact by two.

You can implore someone to save money on their electric bill, a gain frame, or to stop wasting money on energy, a loss frame. You could say, “Learn to manage your chronic disease,” a gain frame. Or you could say, “Get your life back,” a loss avoidance frame. When Bose introduced its wave radio with print ads touting it as new, a gain frame, sales were slow. But when Bose revised the ad’s headline to a loss frame, “Hear what you’ve been missing,” sales took off. Loss aversion is a powerful and widespread bias, so make it work for you by rethinking how you frame your messaging. “Here’s what you’re missing,” can be twice as effective as, “Here’s how you benefit,” for motivating your audience to change their behavior.

Well okay, let’s jump into this. What do you say?

Beth Karlin:
Yeah.

GS:
Some people might think of them as strange bed fellows, behavioral science and energy conservation, but they’ve actually been doing good work together for some time. What can you tell us about the role of behavior and behavioral science in energy conservation?

BK:
It’s interesting that people don’t always automatically see the link, because really, when we use energy, we’re always engaging in some sort of behavior. If we are purchasing things that use energy in our homes or in our offices, as well as maintain or upgrade things that are using energy, as well as just turning on and off lights, televisions, coffee makers, thermostats. Those are all really behaviors. When we think about energy conservation and energy use, we’re always looking at behavior. I think the role of behavioral science is really important. We’ve learned over the past years, decades, the importance of behavioral science and the potential for behavioral science and optimizing all forms of behavior change.

In energy, I think it’s especially important, because when you think about how difficult it is for us to change our own behaviors for things that matter to us a whole bunch, like quitting smoking, and losing weight, and spending more time with our children, or working or stressing out less. Those are all things that have high personal relevance to us and energy use can be even more difficult for kind of four primary reasons.

One, it’s abstract. We don’t actually go home after a long day and use energy. We go home and turn on the TV, turn on lights. We do things that do energy, so it’s by that nature slightly removed from our intentions when we do things or behave. Second, it’s non-sensory. We can’t see, hear, taste, or touch electricity, right? It’s something that’s being used, but we don’t really see a ticker or a counter, unless we install some sort of device or app in our home.

Third, as I mentioned, it has much lower personal relevance than a lot of other things that we engage in, at least in the day-to-day of our lives. Then last, energy use is comprised of hundreds of discrete behaviors. Things like, when you think about even just lighting, installing a light bulb, setting timers, and turning it on and off when we leave a room or go to bed. Because of that, energy use presents real psychological barriers that make it even more important for us to be thinking from a psychological or a behavioral science perspective in terms of how people are using energy and how we can engage people to use less in their day-to-day lives.

GS:
It’s a challenge, but a challenge accepted. There are behavioral programs. Tell us about the kinds of behavioral programs that exist in the energy space.

BK:
Cities, governments, the Department of Energy, energy utilities throughout the country and world have been designing and implementing programs to help inform, educate, and empower people to use less energy in their homes for decades. We can think back to Carter in the late ’70s telling people to wear a sweater and bundle up instead of turning up your heater in the winter. Since then, we’ve seen a lot of different programs.

Traditionally, government and utility programs rely on incentives, which are either rebates or some sort of rewards or gifts. The second most common types of programs that we’ve seen are general informational programs. Anything from pamphlets, brochures, emails, reports being sent home. We also have energy labels. Things like Energy Star that can serve as primes and let us know, “This refrigerator might be a more energy efficient choice than that one.” And more recently we’ve seen the increase in social strategies and programs that leverage things like gamification, competition, social comparison.

Things like Opower’s Home Energy Reports and goal setting and commitment strategies. There’s a wide suite of behavioral programs and behavioral strategies that can be engaged. There have been efforts to kind of categorize them and say there are these four types or seven types of programs. But really, if we think about the kind of core components that any program is trying to target a specific group, an audience to engage in some individual or group of energy behaviors. Like I said, there could be dozens if not hundreds.

Using some sort of content, or informational strategy delivered in some way via communities, groups, schools, peers, email, organizational programs that when you think about those and then incorporating these behavioral science strategies, there’s probably an inordinate number of programs that can be designed and developed and we’re seeing more and more of those being piloted and implemented across levels from individual cities all the way up to national and international programs.

GS
If there are that many behavioral energy-saving programs, what kind of savings do we typically see from them?

BK:
Yeah, they vary. There have been various efforts going back to … If you look in the published literature, going back to about the 1970s, the early to mid-1970s, measuring piloting and measuring energy programs, and they vary hugely. Estimates of savings from these types of programs in the residential sector range from 0%, a program that just didn’t work at all, to 23% savings in some programs, especially those that are more customized, personalized, that use technology.

Most programs in the residential sector are saving between one and 5%. That’s what we’re seeing in terms of variation. I think what that says to us is that, for the most part, something works better than nothing. That doing something, some sort of intervention is probably better than not. And that how and for whom we implement these programs vary a lot. When we’re looking at variation from zero to 23, with the average even in the one to 5% range, leveraging and optimizing behavioral science could really help engage us to save more.

We did a meta-analysis a few years ago just looking at residential energy feedback programs. Those are things like devices in the home, energy reports, emails, bills, and we found average savings around 10%, but with variation from -8 all the way up to 26%. When we applied what’s called moderator analysis, what we found was that how a program is implemented, how feedback is given has a huge effect, so how frequent, how often it’s given, what medium or mechanism is used, whether it provides a goal, or peer comparison, as well as how advice on behaviors that people can use in their homes to save energy all have an impact.

While we’re seeing savings, we definitely aren’t guaranteed savings. That’s why I think it’s so important to be using behavioral science to kind of tweak and optimize to make sure that we’re getting on the higher end. If we’re spending money to reach out to people, that we’re doing it as well as we can.

GS:
What were some of those learnings in terms of reporting frequency and other factors that sort of bubbled to the top as best practices?

BK:
One of them, more frequent. More frequent seemed to work best. We saw kind of a peak around a few days a week. We also saw when we look at duration, we saw what was called a curvilinear effect. We saw that programs that either were under three months long or over 12 months long had the longest, so it looked like … There’ve been other studies that have found that people kind of grow tired of information after a while, unless … After about three months you see people kind of stop paying attention, except for when it keeps going on for a long time and then becomes habitual, and then we see people start to habituate and pay attention again.

We also saw that feedback using technology was more effective than just feedback that was sent home in the mail. We also saw individual-level differences in the types of people that respond to feedback, so not just how we’re giving it, but who we’re giving it to, so customized and personalized approaches seem to offer more potential and have higher potential for savings.

GS:
Now, when talking about the savings, you referred to it as a percentage. Tell me a little bit more about how energy programs are measured.

BK:
Behavioral energy programs need to be measured in a very different way from other forms of energy efficiency programs. Typically, and this is really important, because a lot of people don’t realize this, but these programs are funded, right? Whether they’re by cities or by utilities, they’re funded by taxpayer money, either ratepayer or taxpayer money and therefore cities and/or utilities have to prove, or justify, or deem that savings back into their state or government agencies.

It’s really important for that money to be spent, that we know that it’s leading to some sort of benefit to taxpayers or ratepayers. Typically, energy efficiency programs that have taken the form of rebates or up-stream models like Energy Star, what they can do is, they can just measure the savings. If you think about 100-watt incandescent light bulb being replaced with a 7-watt CFL light bulb, or now LEDs, you can just take kind of 100 minus seven, 93. They also last a little bit longer. You can calculate the lifetime energy savings of that replacement bulb and that way any bulb that gets distributed, you can kind of “claim” that savings.

Behavior’s very unique, in that people are different. We can’t rely on getting people to turn off lights in the room, having that same kind of verifiable consistency in savings as replacing a light bulb or a refrigerator. Therefore, behavior-based energy programs, right now regulations require that they’re measured via what’s called an RCT or a randomized control trial. This is very similar to how drugs are tested, pharmaceuticals are tested, as well as educational programs. It’s often considered the gold standard of measurement. What that means is that you have to take a group of people and randomly assign half or a percentage of them to receive an intervention. That could be any of the types of programs we talked about; a competition game, home energy report, audit, what have you.

Then another group that’s called a control group that does not receive that intervention. Those groups are then compared over a period of time. When we say, percent savings, we’re typically looking at what’s called a difference of difference score, which is the difference between the before and after of the control group and the group that received the intervention. If you just did the two groups, you would get a between subjects. If you just looked at the delta or the change, you’d get a within. But what’s called a difference of difference score really optimizes for both. That’s how we see most of these programs being measured, especially any program that is being funded by any government or ratepayer money.

GS:
Let’s talk about the most common energy behavior program, home energy reports. A little historical perspective, how’d they begin, how do they perform? Maybe remind people what they are, because they’re listening, they’re probably also getting one, chances are. And how did they come to be adopted by the industry?

BK:
Yeah. It’s a really great story. A company called Opower, it was originally called Positive Energy, was founded by Alex Laskey, was founded about a decade ago. I think it was just around a decade ago. The founders of Opower had read about some research that was conducted by social psychologists, Robert Cialdini, Wesley Schultz, Noah Goldstein, among others. What they found was that giving people what’s called a social norm message telling them what other people are doing led to changes in behavior in a number of environmental domains.

One of their most famous studies was this hotel towel study and they found that if you think about those door hangers that hang up in hotel rooms that say, “Please hang your towel,” they often would say things like, “Hang your towel, it’s good for the environment. Hang your towel, it’s the right thing to do.” What they did was they did an experiment, like I was just talking about, and they randomly assigned hotel rooms to have door hangers that either had a traditional message or one that said, “Please hang your towel, other people who stay in this hotel, or this hotel room, hang their towel.”

They found that that was more effective. They then replicated that with energy and water conservation where they had little door hangers on people’s front doors and they, for some people, said, “You should save energy or water,” and others said, “Please join your neighbors or 40% of your neighbors do this.” They found by looking at their meters that people who had received the social norm message were saving more energy, water, like the towels. Opower decided, “This is a really interesting research study and it’s just a research study. Can we actually implement that?”

They did some early pilots, mostly. They started out in California, Sacramento Municipal Water District or Energy District and Pacific Gas & Electric. Then they started expanding when they started seeing savings and they started sending out these home energy reports that, in addition to the bill, they would send out reports to a certain percentage of customers that would say, and have a graph of your energy use, and then compare it to your neighbors on the front. Then on the back, it would say, “Here are some ways that you can save energy.”

These programs were found to be highly successful. We were seeing really on average, again, one to 3%, which sounds like so little. But when you think about how much energy is being used, residential energy efficiency alone totals over five billion tons of carbon emissions annually and we’re looking at … When we’re looking at one to 3%, that actually really does add up. The potential 20% savings that’s estimated as the potential in the average home would add up if every home in America optimized and engaged in energy efficiency and got that 20% average savings, we would be saving as much energy as all three Scandinavian countries combined in a year.

So, even though it was one to 3%, it was an enormous win, especially because these programs were so efficient. They were so low cost compared to other things like in-home audits or working with retailers that might be a lot more expensive. Opower grew in popularity. It grew in terms of the number of utilities it was working with in America and across the world. I was very fortunate to spend some time in New Zealand in 2013-14 when they were launching their Opower program for the first time. The company ended up going public and was just a few years ago acquired by Oracle for … I’m going to get the number wrong, but it was something like $165 million.

They’re still very successful. Our team actually was just completing … Just completed a research project working with Oracle where we were looking, not at the social norm messages on the front of the report, but on the back of the report, how can we message and optimize the behavioral tips, so when you’re telling somebody what to do, can we seek out another quarter percent, half a percent by being more strategic in the way that we’re messaging? That program really opened the door.

The other thing that they did that was so successful, other than leveraging behavioral science, was the way they measured and tested. Until then, it was really unclear how behavior could be measured. But by applying that randomized control trial model, by applying this AB test with a control group and a treatment group, randomly assigning people to receive or not receive, as opposed to just inviting people to participate and measuring those people who did, you could … In a case like that, you could say, “Well, those are the people that were more likely to save energy.”

Because of the way they leveraged behavioral science theory and behavioral science or just scientific methodology and empirical methods, they were able to get regulatory approval. That really opened the door not only for Opower and their home energy reports, but a number of other companies that were looking to apply different behavioral science strategies. Now we’re seeing a lot more attempts and efforts to design and implement behavioral science programs at scale within the energy sector.

GS:
Paper energy reports have been working nicely, but they also represent the past. What do you see as the future of behavioral energy programs?

BK:
We’ve estimated that the average American home can save up to 20% in the home without any sacrifices to comfort or convenience. These are really optimizing efficiency in the home, plugging leaks, getting appliances and devices optimized, making sure that thermostat setting is only used when you’re in the home and need it. Things like that. With home energy reports, we’re seeing, again, on average one to 3% savings. We’ve got about, theoretically, 17% left of the table and there have been other programs that have seen savings much closer to that 20%.

We’ve seen a lot of smaller pilot programs that will get in the five, 10, 15%. Historically, a lot of those have been much more expensive. In-home audits more expensive, school programs that require a lot more work, time, effort. I think right now we have … There are few kind of convening forces that are allowing us, hopefully, and a lot of people are working on trying to figure out how we can tease out why these programs work better and make them more efficient so that we can expand on the two to 3% that we’re seeing with home energy reports to get more and more savings without spending a lot more money.

I think those three kind of trends or convening forces that are allowing us to improve our behavior programs, hopefully, are connection, customization, and collaboration. By connection, I refer to a number of things, but largely connection in our home. Our energy infrastructure has changed significantly and, most people don’t realize it, but our old energy meters, the energy meters that have been in our homes for about 100 years are analog energy meters. The patent was filed on the analog meter in the 1880s. That’s before the internet, televisions, cars, right? They’re being replaced right now.

Certain jurisdictions they’ve been almost completely replaced. In other areas, even New York City, they’re still replacing analog meters with what are called smart meters or advanced meters. Those new smart meters are digital, real-time, and wireless. They’re sampling energy information at the seven-second level. They’re pulling it in real-time at the microsecond level. And they’re sending that information. Historically, to get data to find out how much energy a home is using, one would have to look at that kind of spinning dial energy meter. You would need to send a meter reader out with a clipboard to write down numbers. Which is why historically when you got your energy bill, you’re just getting one data point per month.

But smart meters allow us to collect information at hourly, 30 minute, 15 minute, seven minute, seven second, or real-time and present that back to customers, which can give us a lot more information. When we think about the fact that energy use is comprised of hundreds of behaviors and we can’t really see it, if you just know that you spent $50 a month or $200 a month, you don’t know how to save. It would be similar to going to the grocery store and having no price tags on anything and just filling your cart with food and going home and then getting a monthly bill from Trader Joe’s.

You wouldn’t know if you want to save money, or if you want to lose weight, you wouldn’t know the calorie information. You might not know the price of those various things. Should I buy less Top Ramen or less raspberries if I want to save money on my grocery bill next month?

Similarly, this connectivity, these smart meters are giving us a large amount of data that can both be presented back to customers. We can now, if we collect and sample at a high enough granularity, we can disambiguate among the hundreds of behaviors, or at least the 30 to 50 things that are using energy in an average home. If I look, we can also split it by time, so I can see that, “Wow, my energy use is really going up in that 5 to 7 range. There must be something I’m doing when I’m cooking.” I can look at my cooking behavior.

I can also look, because we’re now seeing time of use rates, so energy costs different amounts at different times of the day. Because of our peak loads, I can go, “Wow, it would cost me a lot less money if I waited until 8:15 to run that load of laundry.” That connectivity is really helpful. We’re also seeing connectivity in terms of smart things in the home. Not only is that smart meter pulling data and sending it either to your energy utility or back to you, it can also directly talk to other things in your home. We’re seeing this increased connected home where our thermostats can talk to our doorbells, to our cars, to our speakers, to us via voice assistance.

We’re able to not just get information or feedback, but we can also control things. I can check on my phone when I’m out for the day to see if I left anything on and make sure I turn it off. Because how many of us have gone away for the day or the weekend and forgot to turn down that thermostat, or turn off lights, or we’re not even sure if we left the oven on? That connectivity is really enabling a lot of opportunity that’s just starting to be capitalized on to improve behavior programs and get even more savings and it’s giving people a lot of power and control over their energy use. What we can see and measure, we can change.

The second, and it’s related to the first, is customization. Because we’re getting a lot more information, and a lot of people might think this is scary, kind of big brother. We hear about this with the internet and other things. But also in our homes, this gives us a lot of additional information. If energy utilities and/or cities can connect that information they’re getting from the meter with other information we have about customers based on who lives in the home, who is using energy, what their occupancy patterns are, how long they’ve lived there, do they rent, do they own, we’re able to customize approaches. This presents a lot of opportunity to really target and say, “You are a family of four that has a pool in your backyard and has lived in your home for 20 years, here are some things that we think you can do.” Recommendations can be improved, information can be improved.

Then the last is collaboration. We’re seeing a lot more opportunities for utilities to work with private sector organizations, much like your own, as well as with different departments within utilities and government. The group that’s doing marketing, and the group that’s working on energy efficiency, and the group that’s working on electrification and/or solar can collaborate. Because we have these kind of connected ecosystems, both in terms of the data that we have, as well as the fact that things can talk to each other, it enables more and more groups and organizations to collaborate to solve problems together.

We can see school-based or organization-based initiatives with large consumers like grocery store chains or hotels collaborating with energy utilities and governments to see larger and larger savings. Because of those, there are huge opportunities for us to provide more connected, more customized, and more collaborative solutions.

GS:
Well, it sounds like an exciting future and it sounds like behavior change science will play a major role in it. Some of our listeners may want to learn more about you and the work you’re doing. How can they contact you?

BK:
I have a website, bethkarlin.com. I also run a small research institute and we’re called See Change Institute. See, as you can … S-E-E, you can see the change, institute.com. You’re welcome to get in touch with me through either of those channels.

GS:
You’ve been listening to the Behavior Change Podcast by Lirio. Lirio provides an email-based behavioral engagement solution that uses machine learning, persona-based messaging, and behavioral science to help organizations motivate the people they serve to achieve better outcomes. On the web at Lirio.co, L-I-R-I-O.co, or follow us on Twitter @lirio_llc.

The thoughts and opinions expressed in this podcast are solely those of the person speaking.The opinions expressed are as of the date of this podcast and may change as subsequent conditions vary. There is no guarantee that any forecasts made will come to pass. Reliance upon information in this podcast is at the sole discretion of the listener.

© Lirio, LLC

Reach Out

Want to learn more about how Lirio’s behavioral engagement solution utilizes behavioral science and machine learning to help organizations motivate the people they serve to achieve better outcomes?

3 + 2 =