One of the biggest problems motor carriers face is finding and retaining enough drivers to meet their needs. Numerous industry studies have shown that driver retention is top of mind for fleet managers, and this concern is warranted — especially in the current, highly competitive labor market.
How do fleet managers identify and intervene with drivers who they fear may want to leave? Many of them — especially those who have been in the industry a while — rely on a gut feeling informed by personal interactions, the types of loads a driver has taken, the number of miles he or she has driven, and other factors.
This type of intuition is valuable and, typically, fairly accurate. However, tracking this information manually can be difficult and time consuming. There’s also the time invested in working with the driver; first listening to his or her concerns, and then following up to see if (a) those concerns were addressed or (b) the driver walked.
Advancements in technology offer new ways to automate this process and eliminate much of the time required to manually track this information. With thousands of data points being tracked and collected from just one truck, key variables can be analyzed using machine learning to predict a driver’s risk of leaving the company, and insights from those data points can provide — and prescribe — recommendations on what type of intervention is needed.
Predictive analytics can use a multitude of data points about a driver, such as tenure, type of load, equipment and pay to predict their risk of leaving. Prescriptive analytics, however, shows fleet managers possible outcomes for each driver based on hypothetical next actions, helping them determine who should take which load and when. Used together, these two types of data analysis can give fleets a recipe to manage and improve the satisfaction of drivers who are at risk of leaving.
For example, for hours of service, predictive data includes the amount of time a driver is behind the wheel, what time of day they drive, when they are in the sleeper berth and if they are on duty between midnight and 4 a.m. These data points are used to prescribe actions that help to retain drivers. Examples of these prescriptive analytics include minimizing the amount of time a driver is on duty between midnight and 4 a.m., scheduling a driver’s breaks during times of low alertness as determined by circadian rhythms, and avoiding overworking and underworking a driver as determined by amount of total time on duty and on duty driving.
Most often, there is no single event that causes a driver to leave a company; it is a series of events that could occur over a period of time spanning weeks, months or years. How do you know if the next load assigned to the driver will be the one that pushes him or her out the door? By the time a manager is able to act on his or her intuition about an unhappy employee, it may be too late. Data and machine learning can automate that intuition, helping managers intervene sooner and more effectively.
To be clear, using data to predict when drivers are at risk of leaving does not fully automate the process; a good driver-manager relationship is vital for building and ensuring employee satisfaction. However, this type of data is a great starting point for retention conversations.
There is no field in any database for a driver who misses their child’s soccer game three weekends in a row. But think of a driver who turns down a load on a Thursday, and then has two hard-braking events on a Friday morning that are uncharacteristic of his or her normal behavior. This data might indicate that the driver is just trying to rush home to see the kid’s game.
If a fleet manager had infinite time, he or she would check in with every driver every day. But, we all know this isn’t practical. By relying on data, fleet managers could focus on those drivers who need the most help and attention.
However, data is only the beginning when it comes to addressing the concerns of an unhappy driver. Fleet managers and dispatchers who work with drivers every day must check in with drivers and make sure they are satisfied with their work. Perhaps that means sending a quick check-in message, a call to discuss how things are going or an in-person meeting to talk through issues, concerns or opportunities.
Often, the easiest answer to driver problems seems to be, “They need a bonus. They need to be paid more.” While pay packages are important for driver retention, it’s undeniable that pay is just the start of a strategy for driver retention; money is not the only determination of whether a driver stays with your fleet. It’s how they feel as an employee. Are they feeling valued? Because ultimately, the most important part of retention is treating drivers as people who are doing a very difficult job.
Orban is vice president of data science at the transportation division of Trimble.