Accident management has definitely proved itself to be an effective way of saving funds. According to CEI, a company that is doing well when it comes to handling accidents caused by external drivers may recover up to 25% spent on repair works, which may account for a sum with five zeros annually.
At the dawn of accident management, handling an accident claim required the whole bunch of employees overwhelmed by papers and responsibility not to miss a single tiny detail that might turn out crucial. Later, to eliminate the enormous amount of manual work, sophisticated accident management software applications emerged. They made the process of claims generating, adjusting and settling/rejecting automated, and helped neutralize the impact of human factors on the outcome.
Now, as we are steadily moving from responding to the undesirable situations on the road to preventing them, technologies that allow predicting future events are starting to rule the roost.
Predictive analytics is a nascent trend, and it has been slowly taking over high-tech industries, such as healthcare, automotive, retail, etc. Driven by the development of Big Data and a widespread usage of telematics in fleet management, predictive analytics, in its turn, gives impetus to other state-of-the-art solutions, such as real-time predictive analytics and advanced driver risk assessment.
IDENTIFYING RECKLESS DRIVERS
Proper driver risk assessment is what every fleet manager is concerned about. Taking into account almost unrestricted access to drivers’ Motor Vehicle Reports (MVR), telematics data, data provided by wearable devices, such as cameras taken pictures every 30 seconds, to timely update every driver’s risk profile, and predict a driver’s behavior on the road seem to be simple tasks. However, it turns out the traditional approach to assessing drivers is not always adequate nowadays.
It is based on gathering data related to driver performance (recent MVR, telematics) and assigning training in those fields where careless behavior is observed. Moreover, drivers are included into a certain risk category according to the number of traffic rules violations or near-accident events they engaged in. It allows managers to focus on high-risk drivers, and intervene with timely training to encourage better driver behavior.
Despite its visible positive impact on breaking bad driving habits and accident reduction, researches show that the traditional approach does not work well in terms of predicting possible future accidents as it appears to be insensitive to the so-called “hidden” high-risk drivers.
Here, we have come up to what predictive analytics can bring to the table, namely identifying risks previously unnoticed. Evidence show that drivers’ risk profiles may differ greatly depending on what approach has been applied.
Predictive analytics models rely not only on MVR and telematics data, but also demographic indicators, such as gender and age, industry a driver works in, etc. It allows creating risk profiles not only for individual drivers, but for a particular fleet as a whole, as well as to predict the number of accidents a fleet is going to experience during a certain period of time.
Preliminary findings by CEI showcase a very high correlation between the predicted and actual number of accidents. Moreover, the number of high-risk drivers has significantly increased as a result of analyzing the additional amount of data. In other words, it is now possible to identify hidden threats. Taken together, these are the significant steps towards ensuring fleet safety.
REAL-TIME PREDICTIVE ANALYTICS IN ACCIDENT MANAGEMENT
Real-time predictive analytics is not here yet, but with the rapid advancement of vehicle-to-vehicle (sometimes referred to as machine-to-machine (M2M)) and vehicle-to-infrastructure technologies, it may evolve sooner than we can imagine. Some experts say real-time predictive analytics will make it possible to eliminate the time gap between identifying a driver’s risk status and changing his behavior to the better. For example, if a fleet vehicle enters a dangerous part of the road, other drivers get notified about it through M2M technology, and can adjust their driving manner. Thus, accidents are going to be prevented in the real-time fashion.
Predictive analytics is steadily making it to accident management, which will completely change its approaches. The enormous amount of data that will become available to managers will provide a 360° picture of driver performance, and make it possible to take well-informed decisions in terms of risk assessment, accident prevention and safety programs development. Moreover, it will be possible to reduce the impact of human factors on the decision-making process due to M2M technologies, though it would hardly be totally eliminated.