Inside Advanced Scale Challenges|Saturday, March 24, 2018
  • Subscribe to EnterpriseTech Weekly Updates: Subscribe by email

AI Fights Online Merchant’s Nightmare: Late Shipments 

Source: LateShipment

Valentine’s Day is right around the corner and you’ve got nothing for your significant other. In a panic, you go to Amazon, pick a gift and check the delivery date. Good news! There’s two-day delivery, the gift will arrive by 8 p.m. on February 13. You hit “Complete Order” and draw a sigh of relief.

Amazon, you are the best!

Two days later, you track the order. Eight o’clock comes, then goes. A few hours later you get an email from Amazon. The gift will arrive on February 15 (or later).

Damn you Amazon, you ruined Valentine’s Day!

Sound familiar? (It does to this writer, it happened to him last month.) But  in blaming Amazon (or other online merchant in a similar B2C or B2B scenario) we’re probably being unjust. Late shipments are usually the fault not of the merchant but the shipper.  In addition to getting blamed unjustly and damaging customer relationships, merchants typically aren’t even aware of shipping problems until it’s too late to warn their customers (and blame the shipper). They only find out when they get an irate customer phone call.

This week, LateShipment, an Orlando, FL, carrier package tracking software company for online and mail order merchants, announced an AI-based tool, called Pulse, that utilizes predictive analytics to notify merchants that late shipment are about to happen.

Founded in 2012, LateShipment is the story of a company sitting on a gold mine of data that, with the emergence of machine learning, decided to develop algorithms to exploit its value. The company was originally formed to enable merchants to collect refunds from shippers for late or damaged deliveries. Customers include ToysRUs, clothing and apparel retailers Roots and Michael Kors, and Carnivore Club, a home-delivery food club for meat eaters.

CEO Sriram Sridhar told EnterpriseTech that the LateShipment platform collects more than 130 real-time data points on 15 million packages per year.

“Looking at our data, we realized…the data we were holding onto gave us a lot of opportunity in terms of what we could tell customers about their shipments in transit – way beyond what a carrier was sharing with them,” he said. “The carriers share only a portion of the information with the merchant.”

Sridhar said Pulse came out of project work done by a data analytics team parsing through the reams of data collected by the LateShipment platform “trying to find relationships where the data could be put to use. One of the projects was trying to see how far ahead of time we were able to predict a service failure. When we ran data tests across  several customers, we realized when we cross-referenced that data with data we already had that we could build a pretty efficient model where about 70 to 80 percent of service failures could be predicted.”

Today, most merchants don’t have the tools or resources to track every shipment. Pulse reports anomalies and issues failure warnings. The company said that over the past three months, Pulse identified more than 138,000 delivery exceptions before they were reported by carrier s and predicted more than 67,000 delays before they occurred.  The system also reported more  than 1,500 instances of packages likely to be returned to a sender, informed merchants of more than 7,000 lost shipments before the carriers’ reports and spotted more than 10,000 instances where internal labels were generated for packages that weren’t shipped, according to LateShipment.

Pulse can be integrated with order management systems to sync customer data to send automatic notifications to customers, and enable support teams to view detailed shipment tracking information.

Sridhar gave the hypothetical of a New York-based merchant that receives an order from a California customer. The merchant engages with a shipper, which says the delivery will take four or five days. Tracking the package, the merchant, or  the merchant’s customer, might see that the delivery is making its way across the country on schedule. But let’s say there are delivery problems in the recipient’s area of California. “Everything seems to be going well and then suddenly the day of or day before the delivery date an exception comes up, and they tell you it will be delayed,” he said.

According to Sridhar , the carrier may have known there were delivery problems in California, “but they don’t share that with the merchant because it’s not in their best interests to tell them there will be problems.”

Pulse collects real-time data on a delivery, then cross-references it with data from millions of other shipments the company has tracked. “That’s where our data and the intelligent algorithms come into play,” Sridhar said. “When that shipment goes out of New York, and we know that particular region of California that there will be a disturbance there, we can have that information much earlier than carriers will report it. We track thousands of similar shipments around the country heading in the same direction. So we’re able to get that data out and apply it to the (Pulse) shipment predictive service.”

Pulse issues dashboard-based updates, and when it notifies a merchant of late delivery, the merchant can then intervene, either warning the customer of package tardiness, or it may even re-ship the order overnight.

On-time shipping is an important factor in customer relations and retention. Industry research indicates that one in three customers who have bad delivery experiences don’t re-order from that merchant. And more than 65 percent say shipment timeliness is a critical assessment criterion.

Add a Comment

Share This