Dow’s Machine Learning Journey In International Trade – Forbes

Dow’s Machine Learning Journey In International Trade – Forbes

A container vessel leaves the port of Singapore.

A ccontainer vessel leaves the port in Singapore on July 16, 2020. – Singapore’s economy shrank more … [+] than 40 percent in the second quarter as the coronavirus plunged the Southeast Asian financial hub into recession for the first time in more than a decade, official data showed July 14. (Photo by Roslan RAHMAN / AFP) (Photo by ROSLAN RAHMAN/AFP via Getty Images)

AFP via Getty Images

Dow Chemical is in the midst of a digital transformation. They have set up Centers to test out new and emerging technologies. They have had success in developing valuable intellectual property in the area of trade classifications.

Dr. John Wassick, Integrated Supply Chain Technology Fellow, Dow Inc. DD , was on a supply chain panel at the ARC Industry Forum in Orlando. Dr. Wassick leads the Digital Fulfilment Center (DFC) that focuses on “competitively advantageous technology” to develop new solutions for the supply chain market. In this context he referred to the company’s Digital Operations Center (DOC) – an idea incubator and proof-of-concept center for operations. Dow has a global manufacturing footprint and a complex international supply chain, as it produces and ships products worldwide. 

Trade Classification at Dow

For Dow, international trade offers a “huge, complex opportunity.” Dow has in excess of 600,000 cross border shipments a year. Out of these, around 150,000 of these shipments must be analyzed by trade experts to ensure the classification is correct. At the heart of international trade is the assignment of a harmonized tariff code to any material that crosses a border. Goods need to be correctly classified so that the right tariffs are paid. Underpayment of duties can lead to huge fines and bad publicity. Incorrect classifications can also lead to goods being help up at customers, impacting customer services.

Recommended For You

The Harmonized System (HS) is a commodity description and coding system, which forms the basis upon which all goods are identified for customs. The HS is used by customs authorities worldwide. Using the right HS code allows companies to pay the correct tariffs. There is an incredible gap between how products are described commercially by trade and how they are expressed in the national customs tariff schedules. This has resulted in error rates of 30%. Worse, every country or trading bloc has its own taxonomy beyond the international 6-digit level.

To say that classifying trade goods is nonintuitive would be an understatement. What a regular person would describe as “baby food” in HS speak is known as a “homogenized composite food preparation;” a “hair blower” is an “electrothermic hair dressing apparatus;” before you can classify “rayon” you have to know whether this is an “artificial” or a “synthetic” fiber; and if you were classifying an automotive part, like a car alarm, you might think you would go to the section of the HS code focused on automobiles, but no – this is an electronic signaling device. To do the tariff classification, there has to be an understanding of the components in the material, the properties of the materials, who the end users are, and what free trade agreements apply. While some materials are not all that hard to classify, chemicals tend to be difficult to classify.

Because of these difficulties, governments look for “good faith” efforts by importers and exporters. Further, the information needed for classification may not be available on the company’s ERP system or in other company systems, said Dr. Wassick.

There are also different philosophical approaches to trade classification. Some companies take the approach that if there is a gray area, choose the classification that will result in lower taxes. Other companies have a limited appetite for risk and work hard to choose the “correct” classification. Dr. Wassick said that Dow is rather risk averse and classification is done after careful verification. Also, to prove good faith, if they are challenged at any stage, there’s a paper trail to prove how the decision was arrived at. 

Dow has dealt with this problem for years and has developed subject matter expertise, which has evolved into intellectual property. The next step is to use digitalization to monetize this intellectual property to address the specific problem of classification of cross-border shipments.  

Dow has been doing trade classifications for many years and has now captured this expertise through machine learning. The process involves extracting the important features of the pre-classified materials. Next a machine is trained using a learning algorithm for classification. Once the machine learning algorithm is trained, it can classify new materials as they come in, explained Dr. Wassick. Dow’s Digital Fulfilment Center has applied for a patent for this actual algorithm. Dow has been approached to sell this technology to off-the-shelf software companies.

Dow, Logistics Digitization, and the Future

Dow is looking beyond international trade operations; they are thinking about edge computing and IoT, how to better utilize analytics, and provide new decision support tools. “Ultimately, where we’re headed is the use of multi-agent systems, where you have intelligent software agents working in concert,” said Dr. Wassick. 

In a global supply chain, the order-delivery process is not as simple as it seems, explained Dr. Wassick. Processing an order is not always smooth – the customer may have defaulted on payments, or there’s a change of delivery destination, or transport problems. However, from a customer experience perspective, Dow wants to instill confidence that their order will be delivered on time; and to be able to do that all the barriers, clutter, and exceptions have been dealt with.

This article was cowritten with Sharada Prahladrao.


Leave a Reply

%d bloggers like this: