Paper presented at the DA/DSM DistribuTECH '98 Conference, Jan. 13-15, 1998 in Tampa, Florida
David Trumble and Sandy Smith, Electrotek Concepts
In today's climate of deregulation, customer service takes on a new, more important role in an arena in which customers are expected to have more choices of energy providers and services. Proven DSM tools such as bill enhancements, residential audit services, and annual reports of weather-adjusted energy usage have an enhanced value in providing customer service via the Internet. The Internet, particularly the World Wide Web, makes possible the low-cost delivery of information and analytical services to residential energy customers. Analytical tools developed around the user interface and conventions of Web browsers minimize the learning curve associated with adopting new tools and technologies. Easy user access is possible with "Wizard"-type options for implementing these analytical tools based on a combination of decision rules and statistical modeling procedures. The result is that a customer's billing history may be analyzed in a detailed and automated fashion with a minimal number of user inputs to produce a useful and thorough instantaneous report. As additional customer information is provided, a more detailed analysis and accurate report is possible.
The first section of this paper makes the simple point that customers need better information and tools to make informed and effective energy decisions. Good energy decisions should save money and result in a more satisfied customer. The second section discusses the type of information that customers need and want to know to make effective energy decisions. The tools and methods used to produce this type of information are also discussed in this section. Good information and analytical tools are critically necessary for consumers to make informed decisions. The third section discusses the Internet technologies that enable a low-cost interactive distributed-computing approach instead of an expensive, capital-intensive one. A distributed-computing approach means that the analytical tools can be developed to provide the best answer possible without incurring the cost of providing additional server resources. Thus, investing in providing proven DSM tools using the new Internet technologies is likely to provide a low-cost means of helping consumers make effective energy decisions.
Numerous energy-related studies have shown that customers need better information to make sound, informed energy decisions. Utility bill information does not give timely energy usage information nor does it break usage down by individual appliances or behavior. Evidence that a consumer's energy decisions are typically sub-optimal is provided by a variety of studies, which include aggregate economic studies, case studies of consumers, and evaluations of information-based DSM programs.
Aggregate studies have shown that customers demonstrate a discount rate of over 100 percent when engaging in some types of energy-efficiency investments. One explanation for this finding indicates that "market barriers" due to a lack of information, feedback, and decision support (Sanstad and Howarth, 1994) are responsible. A special issue of Energy Policy (Fall 1994, Volume 22, No.10) is devoted to this debate.
Case studies of individual consumer decisions involving residential energy efficiency find that a number of common mistakes and over-simplifications are frequently made. Dollar values from customer monthly bills are used instead of actual energy units (Kempton and Montgomery, 1982). For example, the monthly bill for a non-heating and non-cooling month such as April or May could be used to generate an estimate of the non-cooling portion of a cooling month like July. This approach can greatly overestimate actual cooling cost and underestimate base load costs, especially in a summer-peaking service area which is likely to have a higher average kWh cost in the summer. In addition, consumers have been found to be highly inaccurate in estimating actual energy savings from previous energy-savings measures (Kempton and Layne, 1994). Even consumers who made an extra effort to measure energy use and make well-informed decisions were just as inaccurate.
A review of evaluations of information-based DSM programs has generally demonstrated a 5 to 15 percent range of energy savings (Kempton and Ramakrishna 1995). The authors conclude that these results are evidence of market failure and a need for better "feedback," based on the empirical finding that better information generally provides greater motivation for consumer behavior, and a subsequent reduction in energy use.
These various piece of evidence from aggregate studies and individual case studies, indicate that there are significant problems in making energy information readily available and understandable. The next section discusses what types of energy information consumers want and the best means for them to obtain it.
A variety of studies based on focus groups, case studies, and evaluations of DSM programs have found that residential customers want to know
In using billing data to compare annual energy usage, it is important to take into account variations in weather, pricing changes, and billing periods to obtain an accurate measure of whether energy usage is increasing or decreasing over time. A number of weather normalization procedures have been suggested in the literature, of which the Princeton Scorekeeping Method (PRISM) is probably the most widely used. These methods generally break total consumption down into a sum of thermal loads for base, heating, and cooling loads. Heating and cooling degree-day variables are computed using daily temperatures to derive measures based on the billing periods used to calculate energy consumption. These variables are then used to estimate a statistical model of energy consumption (which typically has a non-linear form resulting from the way thermal loads are modeled as a function of daily temperature data). The resulting model parameters can then be used to estimate individual thermal load values as well as weather-normalized values (i.e. what energy consumption would have been with historical long-run average temperature). The question of whether energy usage has changed over time may be addressed by a series of statistical tests of whether these estimated model parameters are constant over time. The customer may then be informed that his expected heating costs have decreased (or increased) $X a year or that it is essentially the same. If desired the confidence level of the test may also be included, such as 90% versus 99%.
These model results can even be used to provide automatic notification of usually high bills. The latest bill received can be evaluated to determine whether it is unusually high for the temperature and duration of that billing period. Users may even select a probability level at which they are notified of unusually high energy consumption.
The comparison of one household's energy usage with another requires a cross-sectional model. The thermal load components from the weather normalization process provide a convenient starting point. A cross-sectional model of the thermal loads as a function of key factors such as number of household members, home type, and size provides a useful means for comparing households with minimal input from the customer. For example, with just a few key questions the customer can learn that while his base and heating loads are about average for a household of his size, his cooling load is unusually high. He may then want to focus on this particular load type and provide more information with regard to his cooling system and cooling practices in order to evaluate a range of cooling-related energy savings measures.
End use energy cost breakdowns of annual and seasonal energy costs are possible with an energy survey approach. The thermal load model results provide an initial, statistically oriented, starting point for layering of additional information. This information can be used to compute engineering-type estimates as well as refine the statistical model results. The estimated benefit for a number of efficiency improvements may be compared with costs and degree of uncertainty in ranking the resulting recommendations for the customer. For example, in focusing on a high cost cooling load, the thermal load estimate of annual cooling cost may be combined with a description of the installed system to compute the savings from a more efficient system or from other measures such as shading, window film, or a reflective roof.
The fast-growing popularity and acceptance of the Internet and the World Wide Web provides a new avenue for applying proven DSM tools. Several new features of Web browsers are ideally suited for implementing new Web versions of proven DSM tools. One important new development in Internet technologies is the introduction of a new computer language, Java, which is designed to allow programs to be downloaded and run on the user's computer in an automatic and seamless fashion. This development allows customer-service oriented DSM tools to become more user friendly with "Wizard"- type features, as well as become more comprehensive without consuming a greater number of server-based resources. Indeed, the distributed-computing approach that this development makes possible should lower server requirements while making the programs more interactive in nature by running on the user's own computer.
The next generation Web browsers recently released by both Netscape Communications and Microsoft Corporation have introduced the option of subscribing to an Internet Channel. Channels allow the user to describe the type of information to be delivered and the schedule for delivering and updating it. Dynamic content can be designed for channels capable of operating in a non-network mode. The channel mechanism allows the downloading and installation process to be automated and scheduled at convenient time for the user.
A utility customer may for example subscribe to a Home Energy Management Channel to have an interactive Energy Analyzer placed on his computer. The Energy Analyzer, which has a copy of the customer's billing data, can then be run as a local program, providing quick interactive results at a time convenient to the user. In subscribing to this channel the customer provides his account number and meter identifier and schedules a download time, such as overnight. At the download time, all necessary content and billing information is saved and installed on the customer's computer. Each time the Analyzer is used, it loads previously saved inputs and determines whether more current billing data is available. If more current data and a network connection are available, the Analyzer downloads this data from the utility server and uploads a copy of the customer's answers. This data transaction is fairly small in size so that there is virtually no waiting, even for customers with slow network connections. An express report option allows an initial report to be generated with a minimum of user inputs. This report may have a number of different charts that include:
A report with more specific estimates and recommendations can be created as customers provide additional information using the energy survey modules targeted at specific loads.
To enhance these reports a number of additional services may be added, such as an On-line Store and Forum targeted for specific energy end uses. These additional server applications are easily integrated with a number of readily available server-side scripting facilities. An on-line energy store provides a useful benchmark cost for a recommended energy measure, as well as a means for taking immediate action on it. Similarly, an on-line energy forum may provide additional motivation for users taking action by being involved in an energy conscious group and contributing to it.
The Internet and the World Wide Web are highly desirable vehicles for making DSM tools available to residential energy customers. The inherently dynamic nature of this medium makes constant updates to software and information not only possible, but also expected. Most utilities maintain a presence on the Internet, providing information on the organization and its services to both customers and external audiences. Some utilities have taken advantage of the interactivity of the Internet coupled with its relatively low entrance requirements (a personal computer and means of accessing the Internet) to make it possible for customers to check on billing status, receive safety and other information, and even pay bills and order services. Extensive use of password-protected access technologies not only ensures the privacy of user transactions with the host utility, but also provides a means for the utility to monitor usage of the Web site and its services.
This ability to compile information on user energy usage and other demographic data (especially from on-line energy usage audits or questionnaires) holds tremendous value for utilities as they devise new strategies to win and maintain customers, and as they evaluate the success of their marketing and customer service programs. For example, if more and more customers indicate that they are using gas water heaters, it may indicate that an electric utility should do more to encourage the usage of high-efficiency electric water heaters.
Popular literature and common assumptions indicate that individuals using the Internet have a higher degree of interest in technology and applications of technology, and are more likely to have a higher propensity to participate in on-line DSM programs and undertake investing in energy-saving measures.
Incorporation of DSM tools into the utility Web site experience yields two significant benefits for the sponsoring utility. First, the nature of the medium and the nature of the user employing the medium will ensure that the DSM tools are used and that the user will employ some energy savings measures - generating demand reduction benefits. Second, this ability to analyze and understand energy usage coupled with the opportunity for the customer to take direct responsibility for managing his or her energy usage will result in the customer realizing a sense of empowerment, and a stronger sense of loyalty to the utility. In a time when customers will have the chance to choose energy providers, this increased opportunity for customer attraction and retention will become as valuable to utilities as the DSM benefits themselves.
Usage of the newly available Internet technologies will ensure that these benefits can be made available using readily available personnel and resources, and can be maintained and kept dynamic with a steady effort by the host utility. The opportunities posed by adoption of these technologies and implementation of on-line DSM tools provides utilities with the means to compete more successfully in a changing business environment, and enjoy the usual benefits that result from successful DSM programs.
Fels, Margaret F. (ed.) "Measuring Energy Savings: The Scorekeeping Approach." Special Issue Energy and Building, 9(1&2), 1986.
Hadely, S. and E. Hirst. "Utility Demand Side Management Programs, 1989 to 1998." Oak Ridge National Laboratory, 1995. ORNL CON 405.
Harrigan, M, Kempton, W., and V. Ramakrishna. Empowering Customer Energy Choices: A Review of Personal Interaction and Feedback in Energy Efficiency Programs, published by The Alliance to Save Energy, June 1995.
Kempton, W. and L. Layne. "The Consumer's Energy Analysis Environment." Energy Policy 22(10) (1994): 657-665.
Kempton, W. and L. Montgomery. "Folk Quantification of Energy," Energy - The International Journal of Energy and Buildings 10 (1982): 817-827.
Parti, M. and C. Parti. "The Total and Appliance-Specific Conditional Demand for Electricity in the Household Sector." Bell Journal of Economics (Spring 1980).
Sanstad, A. H. and R. B. Howarth. "Consumer Rationality and Energy Efficiency." In Proceedings of the ACEEE 1994 Summer Study on Energy Efficiency in Buildings, 1.175 - 1.183. Washington D.C. and Berkeley California: American Council for an Energy Efficient Economy, 1994.
Train, Kenneth. "Discount Rates in Consumers' Energy-related Decisions: A review of the Literature." Energy 10 (1985): 1243-1253.