Can Geocoding AI Solve the Last Mile Delivery Problem?

February 16, 2018

Reading 4 minutes
Detrack Dashboard

One of the biggest challenge for all types of delivery companies is the different ways humans and computers process addresses. Humans are used to looking at a long string of information that includes everything from house numbers to localized postal codes. Humans easily understand different abbreviations, even when others use non-standard ones. But, computers understand addresses as a pair of geographical coordinates.

There are a variety of ways for computers to translate addresses into coordinates through geocoding, but unlike humans, computers often can’t deal with non-standard terminology, format, or abbreviations. This creates a major problem with the last mile of any delivery. Many have come up with solutions to solve this problem but none have achieved significant success.

Now, imagine there exists a geocoding artificial intelligence that can learn how to understand human written addresses and transform this information into accurate geocodes that all computer systems can understand.

Systems have difficulty locating written addresses on the map

Before we delve deeper, let us first look at why delivery addresses not being able to be geocoded into their accurate latitude and longitude constitutes a big problem for the last mile.

How Big is the Last Mile Problem for Ecommerce?

The last mile from the warehouse to the final destination comprises 28% of the total costs to move goods. Anything that can improve the efficiency of last mile delivery will result in huge cost savings and improved competitiveness for firms.

Current geocoding solutions are not adequate for efficient last mile deliveries.  And areas without structured postal systems like Southeast Asia region and the Middle East, have even greater last mile costs and delays.

The problem is compounded by the rise in customers’ expectations. Customers all over the world are becoming accustomed to being able to track the progress of their goods and services in real time and they are now demanding for it to be part of the deal when they make a purchase online. People are used to being able to see on their mobile phones everything from what stage of production their pizza is at, to how quickly their Uber will arrive. Any company that could offer this type of real-time tracking during the last mile would reap the benefits of greater customer satisfaction and loyalty.

Customers expect deliveries to be fast, accurate cheap and trackable

In order for deliveries to be fast, cheap, trackable and accurate, businesses will need to make use of technology. This is where the crux of the problem lies. For many technologies to work, accurate geocodes is an essential ingredient. For instance, technology like routing software systems that can automatically optimize hundreds and thousands of deliveries stops into the most fuel, time and resource efficient routes for drivers, navigational apps that can enable drivers’ accurate navigation to the customers’, real-time tracking technology that can provide full last mile visibility and allow customers to track their delivery live on a map with the estimated time of arrival, and many others; are impossible without accurate geocodes.

Obstacles Faced by Current Geocoding Services

Almost all of the current geocoding automation systems require standardization to work perfectly. If everyone always used a standardized address format, automated geocoding software would quickly convert thousands of addresses into the latitude and longitude coordinates routing software needs to help delivery drivers deliver packages in the most efficient way possible.
But, even in places where there is a structured postal system different people will write or enter the address differently. Consider these possible variations of a single address:

  1. 61 Kaki Bukit Avenue 1, Singapore 417943 (proper street level format)
  2. 61 Kaki Bukit Avenue 1, #04-34 Shun Li Industrial Park, Singapore 417943 (proper postal address but includes unit number and building name)
  3. 61 Kaki Bukit Ave 1, #04-34 Shun Li Ind. Park, S (417943) (postal address using common abbreviations)
  4. Singapore 417943 (the postal digits 417943 points to a particular building in Singapore so this is a valid postal address)
  5. 61 Kaki Bukit Ave 1, #04-34. The 4th level in the main building. Turn left after coming out of the lift and find the last unit along the corridor.

All of these addresses point to the same location, but automated geocoding software would not be able to process most of these formats. It would take a human delivery driver who knows the area to make sense of the addresses.

The problem is even more complicated in areas such as Vietnam, Malaysia, and Indonesia where e-commerce is booming but there is no structured postal system. The lack of standardization has made geocoding automation all but impossible.

Will Geocoding A.I Change Everything About the Last Mile?

There is a solution to this problem that works in countries with and even those without structured postal systems. George, the world’s first geocoding A.I being rolled out by Detrack, could very well be the key to the geocoding problems faced.

We all understand it may not be possible to change the way people write their addresses, but George could bridge the gap between humans and systems. He possesses the capabilities to geocode all the delivery addresses in the world to finally bring an end to the problems of the lack of accurate geocodes that stands in the way to efficient and optimized last mile deliveries.

Join Detrack’s mailing list (sign up on the bottom of the page) and be updated on the release and capabilities of George.


Fanny See

You might also enjoy these related posts

Stay ahead of your competition with a handy email straight to your inbox with the latest posts, updates and industry insights.

Detrack is smart delivery management software that helps improve your business and reduce costs.