Next-Generation Supply Chain Visibility

If supply chain and logistics executives were granted one wish to make their work lives easier and more productive, most of them would probably say, “I wish I had end-to-end supply chain visibility.” It’s been an unfulfilled goal for many companies for many years, but while significant challenges still remain, advancements in technology are helping companies move closer to achieving this goal.

Why has end-to-end supply chain visibility been so difficult to achieve? How are emerging technologies like AI, machine learning, and predictive analytics helping in this effort?

That was the focus of my conversation with Adam Compain, CEO at ClearMetal, in a recent episode of Talking Logistics.

Supply Chain Visibility: The Elusive Goal

If end-to-end supply chain visibility has been a goal for so long, why has achieving it been so elusive? “It’s because the underlying data problem hasn’t been solved,” said Adam. “Regardless if you’re a carrier, a 3PL or 4PL, or a technology provider, the ‘garbage in /garbage out’ problem is difficult to deal with. Much of the transportation data coming from carriers is inherently flawed. As that gets passed through to other systems it becomes impossible to rely on the data to understand what’s happening to shipments and when they will arrive. The EDI transactions typically used for this today are often conflicting, have too much latency, or have basic errors. You can’t rely on them.”

Technology Enablers

Part of the problem is that often nobody is responsible for data quality management, yet people have to use the information to run their business. A decade or more ago supply chain visibility applications emerged to try to aggregate the data and display it on dashboards, but they seldom fulfilled their promises because the underlying data quality was lacking. Now with the explosion of data from mobile devices and IoT, the problem is even more complex. So I asked Adam if newer technologies such as artificial intelligence (AI), machine learning, and predictive analytics can help.

“While the older systems did a good job of gathering and aggregating data, the challenge today is to make sense of the mountain of data available,” Adam said. “That is where AI and predictive analytics can help. For example, a retailer may be relying on six-month old shipment charts with average transit times to decide when to reorder stock. This data may not be accurate or current enough to provide precise shipment times, so inventory may be ordered early as a buffer. By using AI and predictive analytics to resolve conflicting, confusing and missing data issues, more accurate and timely predictions of actual transit times can allow the retailer to match orders to needs and reduce costly buffer inventory. Other examples include resolving issues around transshipment problems and accurately scheduling labor.”

From Static Rules to Dynamic Learning

Adam explains that a drawback of older systems is that they are generally rules-based, using If/Then/Else logic to parse the data. He gives the example of a container that is supposed to arrive at Long Beach but actually arrives at Los Angeles. The old rules logic says the shipment hasn’t arrived as planned. Through more dynamic machine learning, however, an AI system may recognize that the container arrived at a port in the same area and arrange for drayage there. The new technology takes a more dynamic and holistic approach to make sense of potentially conflicting or missing data in order to provide a more complete, actionable view of what is really happening.

The Value Proposition

Adam defines the value proposition for this new technology in four areas. Besides the many obvious opportunities for cost reductions, he notes the potential for profit improvement. “A supplier may be able to upcharge for products based on the reliability of delivery or a retailer may be able to increase profitability through reduction in lost sales.”

A third area of value Adam notes is service. Customer service reps, for example, could spend less time scrambling to dig up data to answer customer questions and more time providing value-added services. And there can also be financial cash flow benefits from more accurately knowing delivery times and when to order, among other examples.

Of course, as with any new technology, companies are making mistakes in trying to deploy it successfully and have many questions about how to get started. Adam provides lots of insights on this, as well as more examples of some of the topics discussed above, so I encourage you to watch the full episode for all the details. Then post a question or comment and keep the conversation going!