How HERE is Mapping the Most Accurate Arrival Times
Synopsis: In our increasingly on-demand economy providing accurate and fast estimated times of arrival (ETAs) is more important than ever. But creating and analyzing data to make those estimates is anything but simple.
Working alongside HERE Technologies, AppGeo is using our expertise in GIS to analyze and improve how we generate ETAs. This webinar (recorded on December 7th, 2021) brings together Aaron Doucett, Sales Engineer and GIS analyst from AppGeo, Dr. Amol Naik of HERE Technologies, and Alexander Osaki of HERE Technologies to discuss the technologies behind the most accurate ETAs.
Meet the Presenters:
The Importance of ETA
Dr. Naik: “If you put that ETA in the context of logistics, and especially in today’s time you realize the value of it, it’s tremendous… ETA is not just a time, it has become more of a prediction. How can I rely on that time? Not only how can you solve that problem, but what does it take to solve it. That’s what we want to share with you today. Is it only location data? Is it more than that?”
Our modern supply chains are incredibly complex and forming predictions that take into account the many moving pieces they involve is a serious challenge. When dealing with such complicated interactions, businesses need to be able to account for delays and other unexpected situations. Part of the struggle to provide accurate ETAs is the high demand that has continued since the COVID-19 pandemic, as well as higher consumer expectations for 1 or 2 day delivery. These pressures make the already difficult task of providing ETAs to consumers even more challenging. Consumers want more information about when their packages are going to be delivered, as well as control over aspects of the process. To adapt to these circumstances, businesses need to make increasing use of location intelligence and technology.
What goes into an ETA?
Osaki: “Big data becomes kind of interesting here, that’s when you see companies deploying machine learning to analyze things like driver behavior, things like vehicle profiles. Understanding that when you’re looking at an ETA how a light commercial vehicle behaves is going to be different from how a scooter behaves or how a heavy vehicle behaves. Not just because the routes are different, but because the driving times themselves are different.“
There is a huge amount of data required to form accurate ETA estimates. Factors that affect timing can range from simpler components like weather or speed limits, to more nuanced elements like vehicle characteristics, time of day, or traffic patterns. As a result, determining an ETA for a given delivery can involve a huge variety of considerations that won’t always be applicable for other deliveries. For ETAs to be accurate enough to meet ever-higher consumer expectations a one-size-fits-all approach won’t work.
How does HERE technology get a more accurate ETA?
Doucett: ”HERE technologies has a real strength in getting down to understanding what every lane of that highway is dedicated for. It’s no longer point A to point B, but all the little details of what’s in between. One of the interesting examples you guys were sharing with me ahead of time was you could take a given curve in the road and know that if we have an 18-wheeler it’s not whipping around that at 70 miles per hour, it’s gonna have to slow down. That’s the kind of deeper insight that we’re starting to use in our calculations that are making them much more accurate, versus the assumption that a truck will just be going 65 miles per hour the whole way and it’s going to be about an hour. We have a lot more complexity going into the equation.“
HERE approaches the problem of determining an accurate ETA by combining information from two sets of data. One set is made up of static data, which is consistent over time, such as roadway information or speed limits. This provides a consistent picture of certain aspects of a delivery route, such as distance from distribution center to a given address. The other set is dynamic data, which accounts for circumstances that can change moment to moment, such as traffic, weather, or the type of delivery vehicle available. By working with HERE, businesses can bring their own static and dynamic data and combine it with the location platform of static and dynamic data that HERE provides. HERE uses the combined data to find patterns and analyze routes days in advance. Even during the delivery however, HERE is still tracking and absorbing live data to improve estimates and update routes in real time. By linking these predictions with real time information, HERE is constantly improving its ETA accuracy to provide the most efficient routes to its partners.
How can consumers deal with ETA updates?
Dr. Naik: “A lot of discussions stop at ‘how do we calculate the ETA?’ but I think the real thing is how do we consume that information? We also have to look at how many times the ETA updates are required. Let’s say you show a picture of a long route where 80% of it is an express highway where nothing changes. What we have seen is that customers require fewer updates on those, maybe once or twice. But as soon as you enter an urban area of a busy city, you require more frequent updates because at the same time the warehouse operations have to start allocating resources for loading and unloading. A lot of things also become more dynamic depending on the traffic, depending on the checking clearances all of that. One thing we have to look at is how many times the updates are required? The other thing is how do I respond to those?”
For ETAs to remain accurate they may have to update over the course of a delivery. This poses a problem for businesses on how to address these updates in ETA. Missing ETAs or providing unreliable ETAs can cost a business their customers and erase trust. When an ETA update occurs business can try to reroute, finding a faster or more efficient route to make their deadlines. However, sometimes rerouting is not an option, and businesses must be able to alert their drivers, the customer, and any other actors involved of changing conditions. Only with real time data and adaptive predictions can such interactions take place to keep ETAs accurate and customers informed. ETAs can no longer be purely predictive, but must also be reactive to changing conditions.