Alerts from telematics systems for risky driving behaviors like speeding, harsh braking or logbook violations have to be acted on quickly. And so must risk observations captured by video event recorders like distracted or fatigued driving.
Besides acting on actual events, fleet managers may have to respond to predictive information that shows which drivers are most likely to quit or have a severe accident.
Failing to act on any risk information can have serious consequences, whether or not the information seems relevant at the time. In the event of an accident, a plaintiff attorney could seek punitive damages by finding evidence that management knew of the risk but failed to act.
For these and other reasons, fleets are starting to use new driver management systems that synchronize their information with immediate corrective actions.
Some systems do this by giving drivers real-time coaching and feedback. Others do it by assigning and documenting training to drivers that violate company policies for safety, compliance or performance.
Overall, the technology is lightening the workload for fleet managers to focus more time and resources on the more serious offenders.
In 2016 the company developed an ELD Driver Retention model that predicts what drivers in a fleet will be next to quit based on more than 1,000 variables derived from driver hours-of-service (HOS) data.
The company says that blind validation studies repeatedly show the model predicts turnover with 80 percent accuracy. It also developed a model that uses HOS variables to predict which drivers will have severe accidents.
The ELD Driver Retention model predicts turnover based on factors that cause job dissatisfaction like fatigue and variability in drivers’ work schedules, says Lauren Domnick, director of analytics and modeling at Omnitracs Analytics.
Information from the models is reported through an online portal called the Driving Center. It gives visibility to the drivers with the highest turnover risk, the reasons why, and suggestions for how to remediate.
Managers use this information to proactively reach out to drivers and have a conversation, she says.
“The idea is to have a positive, non-transactional conversation,” she explains. A manager would begin the conversation by asking how the driver is doing. After listening, the manager could offer suggestions and help resolve problems caused by stress, fatigue, or inconsistent work schedules.
Omnitracs has a database of current hours-of-service data on more than 500,000 drivers. The database is updated constantly by fleets that use its electronic logging software. The ELD Driver Retention model leverages this data to increase the accuracy of predictions for individual customers, such as those with smaller fleet sizes, she says.
The model can also obtain HOS data from fleet customers that use competing ELD products, she says, and from third-party compliance firms such as RAIR Compliance Services.
The cloud-based IoT system will factor in behaviors of drivers over time to detect an increase in risk. For example, a driver might be predicted to have 10 “hard brake” events a week when routed through Chicago, but only two events when routed through Nebraska, she says.
When the system detects a critical safety event – or an exception from what would be considered normal behavior given the circumstances – it can instantly determine if the behavior is fatigue related based on HOS and other variables.
The system could provide real-time coaching by sending a text message to the driver with a recommendation to pull over for a nap or get to sleep earlier tonight, she says.
“Once (drivers) realize that data is being transformed to keep them safer and happier, that is a wonderful thing,” she says.
Meanwhile, fleet managers will be able to see recommendations in the Driver Center in order to have an effective conversation with drivers about the event and to document that interaction, says Drew Schimelpfenig, manager of partnership integration for Omnitracs.
“What if we could stop bad things from happening by showing you risk that you didn’t know existed based on information that you didn’t think applied?” he says. “The pre-emptive interaction could just be a simple conversation with that driver, realizing that he needed someone to ask him if he was ok and that he needed to get back into the game.”