UX Design: Peak pricing for electricity


Changing energy behavior to reduce pollution at Opower


Opower project, as part of Oracle Utilities (Oracle acquired Opower in 2016)

Opower project, as part of Oracle Utilities (Oracle acquired Opower in 2016)


The Challenge

How might we encourage people to use less electricity in their homes during peak hours?

Power companies use peak pricing to charge higher prices for electricity during popular times of the day. This is similar to Uber and Lyft who popularized surge pricing—increasing the price when demand is highest.

If people use less electricity at the same time, fewer power plants will be needed, decreasing pollution—a big impact with relatively little effort.

Project Overview

Working with a cross-functional team across R&D and regulatory, my team at Opower designed an email communications product to motivate new energy behaviors. The 6-month project started with exploratory research and progressed through an iterative design process, with two rounds of usability testing and input from power companies and regulatory agencies.


Lead UX Designer

Data visualization
Product strategy
User research
Behavioral science


Priya Kurani (UX Designer)
Kate Roberts (UX Researcher)
Vernon Laquindanum (Copywriter)
Moiz Kapadia (Product Manager)
Kostia Kugot (Eng Manager)



The Solution


Opower/Oracle is an enterprise B2B2C software company that ingests energy data from power companies. Machine learning and behavioral science help us to translate the data into personalized insights for every household. To help power companies, consumers, and the environment, we designed individualized emails for millions of people. We call these emails a customer’s “Rate Coach”.

The beta product is launching to 50,000 customers in Maryland in April 2019. Our patent application is pending.


The Rate Coach emails:

  • Educate people about their rate plan details

  • Motivate them to shift energy use

  • Compare personal data to hourly prices

  • Track people’s progress over time



Design Process



The project started with 10 in-depth remote interviews. We wanted to understand the nuances of how people think about electricity pricing, and if varying prices motivate them to change their behavior.

Coding and sorting the research data to find common themes yielded the following insights:

  • Comprehension. People understand the concept of peak pricing, but forget when electricity is most expensive

  • Learning. People don’t know which appliances use the most energy in their home

  • Personalization. Information must be personalized to avoid messaging sounding “generic” and irrelevant

  • Savings motivation. People were willing to shift their usage if they could save $5-10 a month

  • Habit-forming. “Lifestyle” changes are too extreme— focused bite-sized actions likely will lead to new habits

  • Challenges. Renters, patients using medical equipment, and larger families with kids will find it harder to change


"If I switched I'd want it to be effortless. I shouldn't have to change my lifestyle.” - Mary


Design iterations

Using our research insights and customer journey mapping, we sketched and developed wireframe flows.

We considered both repeating and rotating information. There are some topic areas we have to save for later iterations once our engineering teams build out new algorithms. Our design insights are driving the future engineering work.

Early sketches

Early sketches

Early wireframes

Early wireframes


Usability Testing

We initially set out to test comprehension of our primary graph, which maps someone’s electricity use against the hourly prices for electricity.

Our team appreciated how our line graph packed in so much information. But, when we tested it with users, we learned this graph was too complicated for the average residential customer. The double axis was confusing, and we needed to further highlight the most expensive hours. After testing we iterated on the design, separating the rate plan info from the usage info, and switching to a bar graph with only one axis.

High fidelity mocks

High fidelity mocks


Diary Study

With our new graph and additional insights from testing, we designed a series of 4 emails. We ran a diary study to replicate the experience of receiving these emails for a month. The same 11 participants were shown each week’s progressive email, first seeing an email inbox and subject line. We learned:

  • The new pricing timeline was clearly understood and participants retained the info through the study

  • The bar graph focused people’s thinking about the activities they did during each hour of the day

  • Behavior changed. Participants reported running their dishwasher or doing laundry outside of peak hours.

High fidelity mocks shown in diary study

High fidelity mocks shown in diary study


“Each week I’ve been impressed, and think that each week builds on the last one. It feels more and more personalized and more valuable.” - Allison



Using behavioral science

The exploratory research, usability test, diary study, and additional quantitative testing informed each design decision, coupled with behavioral science principles as shown here in the first email in the series.

Intro area

1. Simplification. The intro area highlights the benefits to the customer and sets expectations for the frequency and content of communications.

2. Anchoring and social norms. People rely on comparisons to make decisions. Initial research showed that customers asked, “Would I save 5 cents or 20 dollars?” Providing a range of real savings helped frame the possibilities and context. This light framing of what “other people” have done emotionally triggers competitive instincts.

Data section

3. Learning. The details of the pricing plan must be clear and easy to understand. The dollar signs remind people of Yelp and give a quick view of the differences in time periods.

4. Personalization. People want to see their own data—it makes them care. The graph subhead suggests thinking about electricity as part of a personal routine. Color coding draws connection between the customer’s use and the details of their rate plan.


5. Salience. Vividness and specificity can make a message more memorable. By including specific examples of which appliances to use during off-peak hours, the message becomes actionable.

6. Trust. Including appliances that don’t use a lot of energy builds trust that the power company is looking out for the customer.



Next Steps

The product is launching with one of our most strategic clients in April 2019, and will be sent to over 50k residents, with regulatory approval from the Maryland Public Utilities Commission. We’ll measure the success of the program using a randomized control trial and A/B testing to determine

  • Energy savings during peak hours

  • Reduced call volume

  • Customer satisfaction

This test will inform future iterations of the product. Additionally, as energy regulatory markets continue to change, the product can be adapted to meet the needs of other use cases to move toward a cleaner energy economy.