While calculating ETA is one of the most fundamental requirements for any supply chain, it can also be exceedingly difficult to get right. The time it takes a commercial truck driver to get from Point A to Point B can vary widely, based on weather conditions, traffic, hours driven, time of day, operational factors, and many others.
There are three prevailing methodologies when it comes to calculating ETA in the supply chain industry, ranging in complexity from basic math to advanced data science.
Method 1: Distance Remaining / Average Speed
Many shippers still rely on this simple formula to calculate ETA. This rules-based approach is easy to implement, but doesn’t account for dynamic conditions such as weather or traffic, which can completely derail a shipment. As a result of historically low accuracy with this method, companies buffer their time commitments, safety stock levels and labor planning to mitigate risk.
Method 2: Consumer-facing Algorithms
Consumer-facing ETA tools, such as Google Maps, use cutting-edge data science as the basis for their predictions, and typically provide more accurate ETAs than the method outlined above. They incorporate real-time data from dynamic sources, but they are still limited when it comes to specific use cases for supply chains. Consumer-facing ETA tools are built on basic assumptions of consumer transportation habits, which tend to be very different from the behavior of commercial truck drivers. Average consumers are looking for an ETA that assumes they will take the quickest possible route from Point A to Point B, without making any stops during transit. Shippers, on the other hand, have a very different set of concerns, and therefore need an ETA that accounts for regulated hours of service (HOS) breaks, additional stops along the route and business hours of facilities.
Method 3: Purpose-built Algorithms for Commercial Shipping
Purpose-built algorithms for commercial shipping have become the gold standard for calculating ETA. These take into account hundreds of industry-specific data points for the most accurate ETA possible. Data streams include facility business hours, driver breaks and route restrictions based on load type, as well as on-road events like travel bans and seasonal slowdowns. This approach allows shippers to have confidence in scheduling labor and inventory for a leaner, more efficient supply chain.
Shippers who implement purpose-built machine learning algorithms to calculate ETA can save millions of dollars by avoiding OTIF and late rescheduling fees; optimizing inventory and labor planning; and improving customer satisfaction. However, not every shipper is ready to take advantage of this technology right out of the gate. These algorithms depend on good data to deliver accurate results, which can be challenging for organizations that are new to machine learning.
Here are a few things shippers can do to ensure success when calculating ETA through a machine learning algorithm:
Enforce Operational Consistency
It is important to commit to a consistent supply chain to fuel your ETA engine. To gauge the consistency of your supply chain, ask yourself the following questions:
Ensure Data Quality
To generate an accurate ETA, it is important to understand the current state of a shipment by analyzing high-quality data. The most advanced algorithms use data from hundreds of internal and external sources. If foundational data inputs are missing or outdated, the ETA accuracy will suffer. It’s best practice to review the ETA with every geolocation update of a shipment to monitor for current and upcoming factors.
To effectively compete in the on-demand economy, companies must continuously innovate to provide a differentiated customer service experience. Those who take steps to improve ETA accuracy will realize efficiency gains across their supply chain, and build a sustainable competitive advantage in the marketplace by better serving their customers.