Delays are inevitable in international freight and logistics. Estimated time of arrivals (ETA) are often an educated guess, but over the last few years there have been efforts to improve the array of estimates offered, how those estimates are generated, and improvements in making estimates actionable.
Knowing estimated time of arrival is crucial in supply chain management and logistics. Reliable estimates allow for better communication and planning. Shippers rely on ETAs to calculate order lead times, ensure that deliveries are on time, and aren’t racking up costly detention and demurrage fees. Manufacturers depend on ETAs to avoid stock-outs, which brings production lines to a screeching halt and hammers the bottomline.
How do companies use these ETAs?
Active visibility into supply chain delays provides Shippers advance notice of a disruption to their business.. Proactive measures can reduce the overall negative impact and instead increase customer satisfaction. It shouldn't be a surprise that intermodal shippers want more visibility. Today a consumer can order a pizza and see it tracked every step of the way, but they can't find out where their $30,000 container is located - just that it’s between an origin and destination. The demand for enhanced visibility has evolved over the years. It's no longer good enough to know a container is planned to arrive. Today’s ETAs leave much to be desired.
Types of Estimates
In the ocean freight industry, the term ETA is generally used for the estimated arrival of a vessel at the destination port of the shipment. Carriers and different parties, however, can provide estimates for different event types.
The estimates available to shippers depend on the carriers, forwarders, brokers, and technology partners. Some ocean carriers report many estimates, but others report only a handful.
Here is a table of the types of estimates Vizion has measured from the ocean carriers.
Where do ocean freight ETAs come from?
Most ocean ETAs begin with carrier marketing and operating schedules. Marketing schedules are what is published to carrier websites and provided by their agents. Operational schedules are typically internal only and managed by a separate operations team. These operational schedules are being human managed on a daily basis.
Another form of estimate is broadcasted directly by the vessel’s captain in their AIS transmission. This broadcast contains the captain’s estimated arrival at the vessel's next port of call. The estimate may not be the specific event the Shipper is interested in because the next port of call may not be the destination port for the shipment.
Many companies for years have relied on statistical modeling to develop their own ETAs fashioned after lead time designs. The mean duration and the standard deviation between ports is taken into account with ocean schedules to sum together the typical journey time. This was quite an improvement over carrier ETAs, but not due to accuracy improvement. The value of the statistical model approach is that it provides the window of probable delay. Meaning if there's a delay how great could the delay be.
Utilizing machine learning to constant recalculate estimates as events unfold is the major distinction between predictive and statistical approaches. The statistical ETA may be published as a table and updated infrequently. The predictive approach is constant reevaluating the actuals against the subset of the statistical models that best match the current conditions. This approach works well the more impact factors the system can intake and calculate, such as port congestion and terminal processing times. This challenge in this approach is that it is only taking in data from past events, hence it can only predict futures that look similar to the past. Although, the predictive engine approach has a major improvement over other options because of its increased frequency of updating an estimate. It is continuously recalculating taking into account factors changing in real-time.
How Reliable Are ETAs?
In the business of ocean freight visibility there are two constantly asked questions. Where is my shipment and when will it arrive? Over 2020 this question was being asked at an ever increasing rate due to the unprecedented supply chain interruptions. Frequently the answer to this question is provided by container and/or vessel tracking combined with ETA at Port of Destination. But Vizion frequently hears Shippers ask if they can rely on ETAs because of their experiences in inconsistent service levels from the ocean carriers. Few initial ETAs are accurate and they are updated infrequently. On average Vizion sees one to three ETA for Destination Port updates per shipment journey, with the majority of those updates happening at the end of the journey.
How often an ETA is recalculated is its most valuable feature. Most of the pain associated with incorrect estimated events is not due to how correct or incorrect the estimate is, but how often the estimate is recalculated and updated. Ocean carriers today provide very few update estimates and those often are towards the end of a trip when not much can be done to mitigate the impact of a delay. Estimates that are recalculated every few hours and take into account the latest real-time statistical measurements of network velocity will identify potential delays far sooner and become more accurate as the shipment progresses. This means that update frequency is directly connected with exception identification. More updates means seeing wild swings in estimates sooner and now there is more time to act.
Estimate or prediction accuracy is measured with the difference between the observed values and predicted values at a point in time. Measuring accuracy isn’t a challenge, but comparing the accuracy of two different approaches to generating estimates and predictions can start to dive into the weeds. Here is a great article that breaks down the four most common ways of comparing predictive model accuracy. Each approach has pros and cons, but for the sake of comparing ocean carrier estimates to a predictive engine approach we will use MAPE (Mean Absolute Percentage Error). Because this approach takes into account the percentage of accuracy it makes it much easier to compare models. We will be diving into the details of accuracy differences between the sources of estimate and predictions in a future blog post.
Value of Estimates & Predictions
With higher quality estimates and predictions organizations can save money on administration by automating specific procedures that allow their teams to concentrate on exceptions. This means shippers can better allocate key staff to higher-value tasks rather than mundane and time-consuming tasks like calling carriers for freight location updates, sending delivery notifications, and scheduling docks.
Exceptions and delays have an impact on the entire supply chain. The end-customer who is waiting for finished products is the most impacted by a delay. On the other hand, a delay will stifle production, warehouse teams, and distribution centers. Accurate estimates and predictions reduce the risk of overall disruption and allow these teams to make necessary changes to keep things moving.
Real-time predictive engines also enable Shippers to work more closely with their carriers and gain a better understanding of their operations. This openness ensures that freight costs are fair and reasonable, particularly when unexpected costs arise. In the event that a shipment is delayed, carriers may benefit from advance notice as well. It helps them to decrease their dwell time and change their operations to make better use of their fleet and available resources.
VIZION CAN HELP
When it comes to running a profitable company, having a reliable estimate can make all the difference. We at VIZION believe that estimates and predictions data are critical, so we've worked hard to make it as accurate and up-to-date as possible. Our predictive system enables our customers to shift from a reactive to a proactive mode of operation.
VIZION also believes that having the ability to compare two data sets is the best way to hold your carrier responsible, so we always provide carrier estimates for whichever events they support - not just vessel arrival. This is an excellent reference point when deciding which carrier to use when booking your freight.
To learn more about our VIZION visibility data and what it can do for your business, be sure to download our free brochure today.