As consumers, we have become increasingly demanding when it comes to getting what we order delivered. We have a low tolerance for delays or damages, and we don’t hesitate to complain publicly so that other shoppers, and the retailer, can see what has gone wrong. Moreover, a study from 2017 showed that almost 50% of online shoppers consider ‘full visibility on delivery process’ as the most important delivery element.
This development over the past few years has seen us bring that behavior to work, meaning that high demands and low tolerance for mistakes now also apply to B2B logistics, maybe even more so. Not only are the high consumer demands trickling down the supply chain, but companies are also under pressure to minimize costs in order to compete. Precision, speed, and low costs have become parameters that can make or break businesses. Orchestrating those performance parameters requires visibility, and an essential part of visibility is knowing when shipments are going to arrive.
A Vast Number of Variables
The industry’s race to make end-to-end real-time cargo tracking and ETA (estimated time of arrival) prediction a reality has been going on for years. One challenge has been finding technology that can make it happen. Radio-frequency identification (RFID), for example, was heralded 20 years ago as the breakthrough needed, but today, bar-code scanning is still the industry’s most widely used method for tracking shipments.
Another challenge is the complexity of the shipment delivery process. If all there was to it was moving a package from A to B, then a basic GPS receiver could take care of both tracking and estimating arrival time. But road freight, for instance, is much more complex. You have organizational factors to deal with (delivering to multiple locations; delays that can happen at each location), weather factors (rain or wind can lead to congestion), traffic factors (congestion, roadwork, accidents), and human factors (route planning, driving style, experience, incentives) just to mention a few.
For freight forwarders, another big challenge is that many different suppliers can be involved in the delivery process, and setting up interconnected tracking throughout the delivery process requires everyone involved to work on the same system and provide the same output.
Road Tracking at DSV
At DSV, we are developing real-time tracking of road shipments by combining both new and old technologies. While we already offer several tracking services, the challenge remains to get closer to real-time.
In one pilot, we start the process by scanning shipments when goods are picked up from our warehouses. Before leaving, the driver receives a suggested delivery route, while the consignee receives an ETA. The ETA is based on variables including driving time and distance. If there is a delay at one of the destinations, a new ETA is estimated for the remainder of the route, and a new status message is sent to each consignee and the shipper. All the information is delivered through our customer portal, myDSV, and can include push notifications to a mobile phone.
The next development step involves using the existing GPS information to live-track shipments and add machine learning to more accurately predict ETA. Besides driving time and distance, DSV’s machine learning model crunches a wide range of historical transport data to predict how any given shipment is going to proceed. Every 15 minutes, it makes a new prediction for when the shipment will arrive based on the GPS information. If suddenly the predicted arrival time is off target, a notification is sent to the consignee. The clever part is that the model learns along the way – so if ferry times, rest times or anything else changes, the model’s predictions change with them.
DSV’s real-time tracking developments show strong results so far, and it’s an area that has our highest attention. Because with increasing expectations for precision, speed, and cost control, knowing where shipments are and when they are going to arrive can make the crucial difference.