What Makes Homeowners Go Solar? A Scientific, Computer-Modeled Look


A number of studies being conducted as part of the DOE SunShot Initiative aim to determine which consumers buy solar equipment and why. Researchers at the Sandia National Laboratories aim to create a model that predicts household solar energy system purchases based on variables like price, energy savings, environmental concerns and other factors.

What do homeowners want? That’s the $6 million dollar question in residential solar. From surveys on marketing to women to polls on energy choice, marketers and analysts have been trying to get at the answer.

These days, some researchers are turning to computer modeling. The DOE SunShot Initiative is supporting projects that conduct surveys and use computer modeling to examine decisions about how homeowners go solar. A primary goal is to help increase U.S. solar energy from its current share of less than .05% of energy generated to at least 14% by 2030.


One of these projects is being conducted by Sandia National Laboratories. Led by Sandia researcher Kiran Lakkaraju, the study aims to create a model that predicts household solar energy system purchases based on variables like price, energy savings, environmental concerns, and other factors.


“If we can develop effective and accurate predictive models, we can help identify policy variables that could increase purchases of residential PV systems and ultimately help advance the mission of the SunShot Initiative,” Lakkaraju said in a statement.


Such a model, he said, can be used to predict and even influence consumer purchasing decisions.


Computer models


Sandia’s approach is to collect and analyze large amounts of data, said Jerry McNeish, manager of the labs’ quantitative modeling and analysis group. The information has led to two different models: one predicts how likely an individual is to buy a PV system, and the other predicts how long that individual will take to make the investment.


Working with project partners at the National Renewable Energy Laboratory (NREL) and the California Center for Sustainable Energy (CCSE), researchers are conducting surveys of consumers in San Diego County, including 1,000 respondents who have bought PV systems and another 1,000 who have not. Data from the surveys will be studied by Sandia and Vanderbilt University quantitative modeling experts and fed into modeling tools.


Incentives, messages


Those aren’t the only organizations involved. Collaborator The Vote Solar Initiative is collecting additional data. Vote Solar will analyze how consumers respond to economic incentives, discounts, and even peer effects when friends, family or co-workers purchase PV systems. That last item is important, given that social proof is considered critical in boosting solar adoption.


Another experiment, conducted online by The Wharton School at the University of Pennsylvania, is exploring how the framing of messages can influence whether consumers will invest the time to learn more about installing PV systems.


The models include other predictive variables, including the square footage of homes, the national unemployment rate — and even whether consumers own a swimming pool.


Consumer data models


“We’re essentially creating a model that predicts household solar energy system purchases based on such variables as price, energy savings, environmental concerns and other factors,” said Lakkaraju. “But then we’re also running experiments that feed results back into the model. We have a cycle where we use the model to test and generate hypotheses about solar panel purchases, but then we test these hypotheses through experiments to improve the model.”


The Sandia-developed models, Lakkaraju said, have already predicted purchasing behavior 200-500% better than current models.


The team also is investigating financing structures that go beyond straightforward purchases, such as third-party ownership through leasing or power purchasing agreements.


The research team will test their modeling tools, recommendations, and draft guidance on using the models in additional field experiments.


“The significance of this work is that it will help identify those likely to purchase PV systems and help forecast future market trends,” said Lakkaraju. “Ultimately, it will help those in the solar industry to more effectively bring solar energy to consumers.”