This research provides energy and environment policy makers with an up-to-date summary of eco-driving research. Our review of an extensive database of ecodriving studies reveals the fuel and emissions reduction outcomes achieved to date and the conditions under which those reductions were measured or estimated. We summarize the interventions used to achieve those outcomes and consider promising strategies. The research underscores a clear imperative: we need to understand the behaviors that constitute eco-driving in order to design effective interventions. In short, we seek to inform policy makers of (1) the presently understood savings potential of eco-driving, (2) promising pathways for effective policies, and (3) the need for a behavioral perspective to accurately identify the true savings potential of eco-driving and the most effective policies. Starting with a broad definition of “eco-driving,” this paper addresses four questions: Why eco-driving? What are ecodriving behaviors? How much do they save? How are they promoted? Before answering these questions, it is helpful to note that eco-driving is one example of a long- and widely held distinction between technical potential (often assumed to be the best possible performance) and realized performance in the hands of end-users. One common ecodriving reference is the statutory measure of automotive fuel economy for compliance with federal Corporate Average Fuel Economy (CAFE) standards. CAFE is not (nor was it ever intended to be) the best fuel economy any given driver might expect from any given vehicle. Given the central organizing idea of a difference between technical potential and realized performance in the hands of end users, there is far too little eco-driving research that explicitly adopts a theory of behavior to guide research design and data analysis. The section on “Why eco-driving?” introduces a central theme of the overall discussion: Answers vary. The two most common answers are increased fuel economy and reduced emissions. There has been little study of interactions between these and related functions such as safety and reduced pollutant emissions other than carbon dioxide. Functions are considered at the level of the individual driver; little consideration has been given to traffic-network level impacts of eco-driving. This variability extends to the second question, “What are eco-driving behaviors?” Definitions and classifications of the behaviors that constitute eco-driving are inconsistent, sometimes contradictory, and never (to date) comprehensive. To support more precise and systematic definition and classification we propose a framework of function, form, and context; the framework emphasizes the importance of understanding eco-driving in terms of what behavior is to be enacted, by whom, in what context, to accomplish what function. Regardless of the difficulty in comparing or summing results across studies with disparate definitions of eco-driving, it is possible to confidently make two general claims in answering the last two questions addressed in this paper (“How much does it save?” and “How is it promoted?”). First, in addition to the effects of behaviors at the point of vehicle purchase and maintenance, end users of automobiles can reduce energy and emissions intensities while driving their automobiles. Second, there is much more work to be done to support those end users in making and sustaining the maximum practicable and safe improvements. Most studies assess the effectiveness of just one strategy to promote eco-driving. These studies have produced ample evidence that some kind of intervention is typically better than none, but beyond that is difficult to assess since few studies compare interventions. Few studies test hypotheses based on behavioral theory, so the empirical record is difficult to generalize. Rigorous statistical meta-analysis of existing research and additional systematic empirical work, both based on behavioral theory, will garner a greater understanding of which interventions work, for which behaviors, and in which contexts. There is clear potential to engage people—as buyers, owners, and drivers of automobiles—in energy and emissions reduction policies; there also is much to be done to identify the most promising paths for policy development. In sum, while eco-driving research has developed and deployed sophisticated models of technical systems, such as the vehicles and feedback devices, achieving eco-driving’s full potential requires an equivalent sophistication in our understanding of human behavior. Understanding the behaviors that constitute eco-driving is the crux of the challenge to develop most effective policies that achieve maximum savings potential.