Does Feedback Make You Try Less Hard? a Study of Automotive Telematics
Total Page:16
File Type:pdf, Size:1020Kb
Working Paper Series 2018/57/TOM Does Real-time Feedback Make You Try Less Hard? A Study of Automotive Telematics Vivek Choudhary INSEAD, [email protected] Masha Shunko University of Washington, [email protected] Serguei Netessine The Wharton School, [email protected] Mobile and internet-of-things (IoT) devices increasingly enable tracking of user behavior and they often provide real-time feedback to the consumers in an effort to improve their conduct. Growing adoption of such technologies leads to an important question, “Does real-time feedback provided to users improve their behavior?” To answer this question, in this paper, we study real-time feedback in the context of automotive telematics, which is the key technological disruption in the automotive insurance industry. Automotive telematics enables drivers to track their behavior using various technologies inbuilt in a mobile device and to provide real-time feedback on the driving quality. We investigate the impact of such real-time feedback on driving behavior, as measured by several parameters such as harsh braking, over speeding, and steep acceleration. Contrary to much of the existing feedback literature, we find that, on average, the driving performance of users post-detailed feedback is 13.3% worse than the performance of users who do not review their detailed feedback. This impairment in performance translates into a one-year reduction in inter-accident time. Our results suggest this deterioration is associated with increased sharp accelerations and over speeding. Drivers also demonstrate higher speed dispersion within a trip after feedback that results in 1.65% increased probability of an accident. Further, we demonstrate the critical role that insurance incentive thresholds play in the effect of real-time feedback. Key words: Real-time Feedback; Empirical Operations Management; Behavioral Operations Management; Automotive Telematics Electronic copy available at: https://ssrn.com/abstract=3260891 A Working Paper is the author’s intellectual property. It is intended as a means to promote research to interested readers. Its content should not be copied or hosted on any server without written permission from [email protected] Find more INSEAD papers at https://www.insead.edu/faculty-research/research Does Real-time Feedback Make You Try Less Hard? A Study of Automotive Telematics Vivek Choudhary, INSEAD Masha Shunko, Foster School of Business, University of Washington Serguei Netessine, The Wharton School, University of Pennsylvania Abstract Mobile and internet-of-things (IoT) devices increasingly enable tracking of user behavior and they often provide real-time feedback to the consumers in an effort to improve their conduct. Growing adoption of such technologies leads to an important question, “Does real-time feedback provided to users improve their behavior?” To answer this question, in this paper, we study real-time feedback in the context of automotive telematics, which is the key technological disruption in the automotive insurance industry. Automotive telematics enables drivers to track their behavior using various technologies in- built in a mobile device and to provide real-time feedback on the driving quality. We investigate the impact of such real-time feedback on driving behavior, as measured by several parameters such as harsh braking, over speeding, and steep acceleration. Contrary to much of the existing feedback literature, we find that, on average, the driving performance of users post-detailed feedback is 13.3% worse than the performance of users who do not review their detailed feedback. This impairment in performance translates into a one-year reduction in inter-accident time. Our results suggest this deterioration is associated with increased sharp accelerations and over speeding. Drivers also demonstrate higher speed dispersion within a trip after feedback that results in 1.65% increased probability of an accident. Further, we demonstrate the critical role that insurance incentive thresholds play in the effect of real-time feedback. Keywords: Real-time Feedback, Empirical Operations Management, Behavioral Operations Management, Automotive Telematics. 1. Introduction Mobile and IoT devices increasingly allow tracking of user behavior that was not observable before, be it a clickstream during shopping, geolocation while running, or heartbeat patterns during sleep. In addition to collecting and summarizing data, such applications and devices usually provide real-time feedback to the users, typically in an effort to change their conduct. Overall, the goal is usually to “improve” user behavior pertaining to a particular setting: for example, sleep better, run longer, reduce workload, and decrease calories intake. Providing automated feedback to the consumer in real time has potentially large implications for system performance. Numerous disruptive technologies (such as sensors in mobile devices, RFID, etc.) provide objective and insightful assessment of one’s task performance: e.g., FitBit monitors physical activity in real time and EarlySense monitors sleep quality. In industries ranging from utilities (Tiefenbeck et al. 2016) to 1 Electronic copy available at: https://ssrn.com/abstract=3260891 transportation (Toledo and Lotan 2006), technological advances already allow immediate feedback to users, who can use it to improve performance. Such feedback is important, as human decision making is often biased (Kahneman and Tversky 1973) and therefore, feedback or nudging can be used as a low- cost intervention to exploit human behavioral tendencies such as ahead-seeking (utility gain from over performing), behind-aversion (utility loss from underperforming) (Roels and Su 2014), and last place aversion (Buell and Norton 2014). One important application of IoT is in the automotive industry and it is termed vehicle telematics or telematics for short (https://en.wikipedia.org/wiki/Telematics). Automotive is a huge industry serving more than 1.3 billion vehicles on the roads worldwide and in the USA the automotive industry accounts for nearly 3.5 percent of GDP and employs more than 940,000 people directly or indirectly in manufacturing (USASelect 2018). Due to its size, transportation is the second-largest polluter in the USA and it is also one of the biggest contributors to death/injuries in the world with approximately 1.25 million deaths every year due to accidents (WHO 2015). Evidently, driving behavior can have drastic implications for safety, emissions, and vehicle longevity. Moreover, recent economic trends including urbanization, concerns about city pollution, the gig economy (e.g., Uber, Lyft and similar companies in the trucking industry like Convoy), crowdsourced deliveries (e.g. Amazon and InstaCart), and ridesharing (BlaBlaCar, Car2Go) all lead to higher vehicle utilization (Cramer and Krueger 2016). As a result, growing number of drivers operate potentially unfamiliar vehicles in unfamiliar congested locations while increasingly relying on GPS-enabled devices. Such conditions create potentially unsafe environment on the roads and pose challenges for the companies in managing their drivers and maintaining their fleets. Thus, telematics applications have become numerous: companies can monitor drivers’ behavior, including harsh braking, steep acceleration, lane changes, sharp cornering, and speeding, which all lead to higher fuel consumption, higher emissions and more accidents. Companies also use this information to manage the drivers’ pool and the fleet, and provide incentives to lower insurance premiums and operational costs. For example, with the help of telematics, left turns have been found to be associated with more accidents (Najm et al. 2001) and UPS has saved millions of dollars and gallons of fuel by implementing ‘no left turn’ policy (UPS 2016). Acknowledging these developments, telematics has been recognized as the most disruptive technology (Accenture 2018) in the automotive insurance industry, which is already a very large industry with revenues exceeding $250 billion in 2017. Telematics is utilized in usage based insurance (UBI) which is expected to have 142 million subscribers by 2023 (IHS Markit 2016). There are several other applications of telematics where insights from this study apply. For example, fleet drivers are often incentivized to improve driving behavior through threshold bonuses. Further, GPS-enabled devices, sensors, and drivers’ phones together allow UPS to capture data related to vehicle routes, idling times and safety practices (UPS 2016). Gig economy companies, such as Uber in Singapore, often lease 2 Electronic copy available at: https://ssrn.com/abstract=3260891 vehicles to their drivers and track driver performance closely (Beinstein and Sumers 2016). While telematics devices are already providing complex real-time feedback to drivers, we still poorly understand implications of such feedback on user behavior. This understanding is important given that other attempts to make driving safer have led to unintended consequences. For example, the introduction of airbags in cars in the mid-1970s led to more deaths in lower-speed collisions (Cunningham et al. 2000). In this paper, we focus on an application in which insurance companies incentivize individual drivers to drive better through insurance premium discounts and real-time feedback. At a high level, much of the existing literature suggests that “conventional” feedback (post-factum average performance, for example, utility bills