How to Turn Data into a Real Asset in Global Supply Chains Using Data Analytics – A Duty Drawback Case Study

Global supply chains face many challenges due to the dependency on interconnected activities and the increased complexity of balancing supply and demand over the years. A successful supply chain process must manage costs at the lowest possible level while keeping customer satisfaction at the highest possible level. Tariffs are one example of costs imposed by governments throughout the value chain that usually serve to protect domestic manufacturers. The ability to minimize tariff/duty costs can help firms compete on price and ultimately improve their bottom line. The implementation of a Duty Drawback program is one of the best tools companies possess to minimize tariff costs.

Duty Drawback is a valuable practice adopted by exporters to recover duties, taxes and fees paid on imported goods that are subsequently exported in the same or similar condition. According to U.S. Customs and Border Protection (CBP), the main benefit is the refund of customs duties and certain internal revenue taxes.

The recovery of duty requires a robust connection between all relevant data within the global supply chain. In addition, the ability of companies to actively process and visualize their entire end-to-end processes is fundamental. Therefore, the data-driven approach used to identify constraints in business processes, thereby making the project involve improvements to those processes, combined with the expertise in data analysis is a crucial factor to optimize the duty value to be claimed.

After identifying the necessity and opportunity for a potential internal process improvement, the Supply Chain Finance and Global Operations Centre of Excellence Data Analytics Teams at Extreme Networks collaboratively designed and developed a framework capable of assisting the decision-making process. This framework is focused on optimizing and estimating future drawback claims (matching imports to exports) and ensuring compliance. This article aims to demonstrate a practical case study of how data can be turned into a real asset and how this approach can be extended to companies in similar situations that could use Duty Drawback as a cost saving measure.

The Challenge

Duty may be recovered when imported goods are the same kind and quality (SKAQ) as the exported goods. Among several methods of recovery, the substitution method described by the CBP in Section 313(b) of the Tariff Act [19 U.S.C. Section 1313(b)] provides significant opportunities for companies to recover import duties. The advantage of this method relies on the flexibility to claim back the duty paid for a product at the 8-digit Harmonized Tariff Schedule (HTS) level. Therefore, the duty paid for a product imported can be recovered partially by an export with the same 8-digit HTS regardless of the serial number or stock keeping unit (SKU) exported.

However, this method requires a more advanced analytical approach not supported by traditional descriptive analytics. The main challenge relied on incorporating essential rules in this process, including, but not limited to, the fact that (i) imports expire in a five-year timeframe; (ii) the date of the export must surpass the importation date and (iii) the value of the export might exceed the value of the import to guarantee an improved duty recovery; otherwise, the value paid is partially lost. Additionally, certain HTS codes control goods by distinct units of measurement, impacting the process of identifying the most suitable exports. Ultimately, constant monitoring of legislation and geopolitical regulations is required because they can dramatically influence the dynamic of the duty drawback process. The assistance of external customs brokers and consultants is fundamental for this task.

In this case study, a data-driven methodology was developed applying all four dimensions of data analytics (descriptive, diagnostic, predictive and prescriptive) well explored in the literature and described as follows.

In the descriptive analytics phase, the question “What happened?” is frequently addressed. This phase involved processing and exploring the import records through exploratory data analysis. In addition, these imports were linked to several sources of data and integrated into a dashboard refreshed daily, allowing all subject-matter experts (SMEs) to visualize the amount of duty available to be claimed.

The second phase of the project consisted of providing a diagnostic analysis, answering the question “Why did this happen?” by uncovering crucial information and the main factors involved in recovering the value paid. The integration of suitable exports and their current statuses were mapped, such as appropriate documentation and exports partially and totally claimed. Dependencies from other related areas were identified, and a collaborative approach removing some data silos was essential to improve the entire value stream of the duty drawback process.

In the predictive analytics phase, answers to questions such as “What might happen in the future?” are commonly expected. Machine learning and statistical models are widely applied at this phase. A many-to-many matching algorithm is the most appropriate solution for this type of problem. Let E and I be two sets of points representing Exports and Imports, respectively. Therefore, each point in I matches at least one point in E and vice versa. The main objective of the matching algorithm is to optimize the total distance between import and export value per unit to achieve highest claim value while optimizing for minimal losses. This value is constantly calculated because additional import and export records might change the dynamic of the matches.

Ultimately, in the prescriptive analytics phase, a list of imports and their respective recommended matched exports is accessible via dashboards. The value forecast to be recovered is explained by the matches processed between the relation of import-export records. In addition, all imports unable to be matched are visualized in a Pareto root causes graph analysis, answering one of the most critical questions in a data analysis project: “What should we do next?”

Results Achieved

The true achievement of this case study is the ability to stay ahead of the curve by estimating the total dollar value of duty to be claimed based on the most appropriate matches of imports and exports. The implementation of all four dimensions of data analytics unlocked the hidden insights of all data points directly or indirectly connected to the duty drawback recovery estimation.

The data visualization phase was essential to identify constraints within the value stream of the duty recovery, address the complexity posed by the substitution method and support the broker in filing for the final value of duty claimed.

The value to be claimed is measured by HTS codes; therefore, it is easy to identify the individual performance of each HTS. Imports unable to be matched due to reasons demonstrated in the Pareto root causes analysis will be still available for future claims because the automated process is refreshed daily, gathering new suitable exports. Additionally, the data-driven framework contributed to reducing repetitive and laborious tasks performed by the SMEs because the data treatment and matches performed are automated. As a result, they can concentrate their time on more value-added activities.

The Drawback Analytics Platform (DAP) was a challenging project because it required the collaboration of cross-functional areas such as finance, compliance, IT, supply chain and data analytics. However, the capacity to properly interconnect and treat data allowed our team at Extreme Networks to more accurately visualize the data flow, improve specific processes and turn data into a real asset.