Understanding how feedback impacts people and institutions (whether in sports, school, militaryoperations, or …) in learning and change (hopefully, improvement) is a complex and fascinatingworld. Data collection and feedback systems, as a tool for providing building/system operators the ability to make more informed decisions about energy use, are a critical pathway toward more effective (generally, more efficient) energy use. There is a reason that the U.S. Navy invested significant portions of its energy-related stimulus funding in smart meters and that the Department of Defense’s Operational Energy program has put data collection as a critical front-piece of activities: if you don’t know where you are, how can you make intelligent decisions about how to getto where you want to go?
As readers of this blog should be aware, I have a particular fascination with the impact of real-timefeedback systems on individual behavior. Considering ‘the Prius effect‘ of people becoming more energy conscious and safer/more efficient drivers with the constant reminder of the energy impact of their (for example) acceleration, there is a good possibility that one of the most cost effective paths to reduce U.S. oil demand would be to put feedback devices on all post 1996 light-duty vehicles. We “know” that people modify their behavior in the face of feedback — but do not necessarily knowhow, how much, and what the “ROI” is for the investment cost to get there.
With a belief that “feedback” offers real potential as part of a portfolio approach toward a cleanerenergy system, I was pleased to learn that the American Council for an Energy-EfficientEconomy had done a meta-study of real-world tests of real-time feedback systems’ impacts on household energy (mainly electricity) use. The ACEEE has some top-notch talent for whom I have a great deal of respect. The press release created serious concern that my belief that feedback systems were something to include in our future plans / programs might have been mistaken. Consider the title: Large-Scale Studies Find Moderate Savings from Smart GridFeedback Devices. And, the key point re “moderate”: “average savings of 3.8%”.
Okay, a savings is there — across “large-scale studies” but it doesn’t seem to be all that much. In2010, the average U.S. household used just short of 11.5 megawatt hours / year (11,496kWh/year). Projecting such a 3.8% savings would result in 437 kilowatt hours of reducedelectricity demand or about $45/year at average electricity costs. Something but, well, short of $4 /month for the average household isn’t an indication of major impact.
The report concluded that “the cost of providing real-time feedback remains high”. Looking at the table as to the most common cost, per household, to the utility of being around $500, the “moderate savings” started to seem even worse: a 10 year payback for utility costs wouldmean a 10+ year payback of costs, on a straight line basis, without even discussing utility profits nor savings to the household.
As a real ‘feedback enthusiast,’ my first glance at this report created a rather dismal mood.
A second look, however, suggests that perhaps the ‘dismal’ thoughts are misplaced. The devil, ofcourse, is typically found in the detail:
- The 3.8% savings figure excludes an outlier study which showed nearly 20 percent savings. That “large-scale study”, in fact, is an examination of a 20+ year experience in Northern Ireland which not only has more than five times the longevity of the second-longest study in the report but also had about five times as many people as all the rest of the studies combined. In a simplistic, but illuminating way to consider the situation, the ACEEE meta-study excluded from its conclusion a study with more than 25 times the people months of experience of all the other studies combined. If a more accurate reflection of end results, this would suggest annuals savings more in the $200+/year or about $18/month range, which is far more interesting to most households than $40/year.
- The $500 figure actually misrepresents actual program costs. With the exception of the Northern Ireland experience, all of these were experimental programs. The “costs” are simply dividing the total program costs (equipment, labor, administration, analysis, …) by the number of involved people. Clearly, ‘experiments’ almost always cost more per involved person than real-world business activities. The Northern Ireland program, which involved over 45,000 people (the second largest involved 5,550), had average costs of $99-$113 rather than $500 because it was an actual deployment of technologies and business processes rather than a research experiment.
Okay, if we take these figures: perhaps $400/year in savings and $100 for program costs, the payback is measured in a few months rather than ten years (and that, of course, isn’t counting the value of reduced peak demand). This gives plenty of space for sharing electricity savings value streams between the provider and consumer in a win-win space which also provides for a more efficient and less polluting household energy use.
To be clear, I have a lot of respect for the ACEEE and their research teams. The report, Results from Recent Real-Time Feedback Studies, has a lot of interesting and valuable material within it. The issues raised above about potential savings and costs to achieve those savings are illuminated simply by reading through the report. However, it is too easy for someone to walk away from the report thinking that feedback systems have very marginal impact at a high cost due to the exclusion of the Northern Ireland experience from ‘average savings’ and the use of test program costs as a surrogate for actual program costs. Thus, the title of this post: it does seem odd that the American Council for an Energy-Efficient Economy would skew report results to understate the impacts of technology to foster energy-efficient behavior.
I have to say that I strongly agree with a key ACEEE report conclusion:
More and different kinds of research are required to better understand the actions and investments that households may be making in response to real-time feedback.
More knowledge and information about energy use are not enough to solve our energy challenges but they are key to fostering an Energy Smart society.
NOTE: I will be returning to do more analysis of this and other studies re feedback devices (such as this, this, or this). As an example of another study, this 2011 RAND report (pdf) comments that
studies have shown that providing different types of electricity usage feedback can reduce usage by around 20 percent and can help shift up to 40 percent of peak demand to off-peak hours.
That is quite a different conclusion than what is seen in the ACEEE study (which doesn’t include the RAND work in its bibliography). E.g., I am intrigued by the differences between the studies and plan to learn more — and share that learning here.
Related posts: