- By Steve Banker
- Factory Automation
- Supply chain control towers using digital twins and other sophisticated technology can support agility.
- The COVID-19 pandemic created supply chain disruptions, illustrating the requirement for agility in manufacturing supply chains to be successful.
- Being able to support better visibility is based on using application programming interfaces rather than relying on electronic data interchange, emails, portals, and phone calls.
Achieve agility using supply chain control towers and digital twins.
Supply chain control towers have come of age. And, as the supply chain disruptions related to the current COVID-19 pandemic have further highlighted, agility across manufacturing supply chains is essential. The concepts that allow modern control towers to support agility are easy to describe. However, the technology that empowers these concepts is quite complex. Supply chain control towers using digital twins and other sophisticated technology can support this agility, but to do so effectively, the technology must do the following:
- Provide near-real-time visibility to inventory, shipments, and risks.
- Provide support for concurrent planning. Organizations have different plans for supply and demand planning. They have different types of plans for long-range, intermediate, and near-term planning. With concurrent planning, these plans are connected. A change in manufacturing scheduling, for example, ripples up to the integrated business plan. Changes in short-term demand forecasts get highlighted in the supply plans.
- The system needs to be intuitive. Let’s take the example of a “hot load.” The trick is to identify what is “news” and what is “noise.” Perhaps a truck is late, but it contains product for which a company has plenty of stock. On the other hand, an ocean container might come in early but contain products that are in a low stock position. So, the system needs to tell users which to receive first.
- Supply chain applications need to make more accurate forecasts. Downstream data increases forecasting accuracy, particularly for near-term forecasts. Forecasting accuracy and estimated times of arrivals are improved when machine learning technologies are applied.
- The systems need to be easy to use so that planners can easily run scenarios to understand how best to mitigate exception situations.
- The solutions need to support internal and external collaboration, playbooks, and war rooms. Not all problems can be solved using math. Collaboration with diverse stakeholders is often necessary. If certain types of exceptions occur on a regular basis, a playbook can be built to help automate these recurring problems.
Technology needs to be complex so solutions can be simple
Rather than delving into each of the above bullet items to describe the overall complexity, let’s focus on just one of those areas here: getting near-real-time visibility to shipments, inventory, and risks. The ability to support better visibility is based on using application programming interfaces (APIs) rather than relying on electronic data interchange, emails, portals, and phone calls. Because many large companies use many different business and supply chain applications across different business units and regions, the integration problem is significant. Companies may need to use data lakes and technologies that help to harmonize the data from the different applications.
A data lake consumes all the critical data from the different applications. Significant work is required to generate a harmonized data layer. The critical master data for supply chain management includes objects like sales orders, shipments, inventory, and lanes. Different business systems can define all the fields associated with these forms of master data differently. When implementing this, a company needs to ask its business users what data points they need for each of these master data objects. It does not matter what various business systems, like SAP, think a “purchase order” is. A harmonized definition of “purchase order” is created based on what is needed to support the control tower.
Then these objects need to work together to create a time-phased view into what inventory will be available at what locations. To understand the current inventory position, a company takes its initial stock position and then calculates how that inventory position is changed by things like quality inspections, purchase orders, production orders, and intercompany demand.
The improved visibility to shipments was based on federal regulations requiring truckers to electronically track their hours of service. Suppliers like FourKites or Descartes MacroPoint provide near-real-time visibility based on API integration to truck carriers’ GPS-enabled electronic logging devices or by getting small carriers to agree to use smartphone applications that support tracking.
The solutions that provide real-time visibility to supply chain risks are fascinating. These solutions work by continuously monitoring a wide variety of risks, identified based on monitoring hundreds of online and social media sources, and then linking those risks to a map they have created of their customer’s end-to-end multi-tier supply chain.
The other things supporting agility—concurrency, intelligent alerts, the use of downstream data, and so forth—all have their own complexities. But the basic point is this, we have dreamed about having robust supply chain control towers that can provide control that is analogous to what airport control towers or war rooms provide.
Medtronic’s digital twin supports its ability to respond in the pandemic
As ARC learned in a recent webinar, Medtronic, a large manufacturer of medical devices, has been engaged in a digital transformation. This included building a digital twin of the company’s end-to-end supply chain to make it easier for people to make better and faster supply chain decisions. The company built a data lake to feed LLamasoft’s Supply Chain Analytics Platform, llama.ai, to create its supply chain digital twin. The system has already been rolled out to support distribution flow path analyses. The solution allows Medtronic to make better decisions.
Medtronic is a big company with a complex supply chain: hundreds of manufacturing and contract manufacturing sites, dozens of sterilization sites, over a thousand suppliers, more than 250,000 distinct products. What is manufactured is distributed to hospitals, clinics, third-party health care providers, distributors, government health care programs, and group purchasing organizations. Approximately 60,000 daily shipments flow through a network of cross docks, forward stocking locations, and distribution centers, across multiple modes of transportation, in shipments ranging from less than a pound to full-size crates, out to over 150 countries. This work is done by more than 3,000 employees working in Medtronic’s supply chain operations.
The digital twin journey
For Medtronic, a supply digital twin is a continually updated digital representation of the company’s end-to-end supply chain. This digital twin allows the company to run more scenarios and make better decisions in a timelier way. Before it started on the digital twin journey, this was a struggle. For example, if the company wanted to analyze how certain products were flowing to customers, and whether node skips would make sense, a project might take seven months, by which time the supply chain would have already changed.
Medtronic is still in the early stages of its supply chain digital transformation. The initial focus of this implementation centered on a distribution flow path analysis—how products flow through the global network to customers. LLamasoft helps it run trade-offs. A slower mode of transportation requires holding more inventory to attain targeted service levels. Faster delivery paths lead to higher transportation costs but require less inventory. The company can also look at whether adding or subtracting a node—like a cross dock—would lower the total landed costs. And this analysis can be “pan-Medtronic.” For example, a cross dock currently used by one business unit might provide additional benefits if other business units’ goods flow through it.
“Journey” is the right term to describe Medtronic’s digital initiative, and the difficulties associated with this kind of project should not be minimized. One key challenge was the need to pull data from many systems. Often the data was not clean or complete enough. In some cases, the data just does not exist. While the company understands its transportation costs at a granular stock keeping unit (SKU)/mode/origin-to-destination level for most shipments, that was not universally true. In some regions, for some products, it does not have visibility into which specific SKUs are in a shipping container. Understanding the landed costs at a granular level is necessary to doing the trade-off analysis. Getting this data became a project involving a good deal of labor.
In other cases, getting data was easier. For example, the company did not understand the fixed and variable costs drivers at some warehouse and sterilizer facilities. Here it conducted a survey with facility managers to get this data.
Impact of COVID-19
Faced with the challenge of pandemic-related supply chain disruptions, Medtronic’s supply chain team worked to provide visibility to what was happening around the world. One part of this was stand-up meetings twice a day and once on the weekend—to talk about the latest developments and issues.
The supply chain impacts have been huge. The supply chain has been whipped around by the need to protect the company’s people, unexpected demand surges and crashes, and countries shutting down and closing customs. Shipping by air became problematic. Much freight moves in the belly of passenger planes. As these planes stopped flying, air capacity vanished.
As COVID moved from a China problem to a global problem, the team responded by pulling inventory forward into countries before inbound shipments became impossible. They had discussions with their leading carriers where they explained what they were doing and pressed for commitments to move the company’s goods.
The Llamasoft product was one of the tools used to respond to the pandemic. Ninety-five percent of the data Medtronic needed to run COVID scenarios was already cleansed and validated in its digital twin model. The COVID analysis mainly consisted of tracking down the last 5 percent of the data points the company did not have and running “what-if” scenarios.
While demand has dropped off for many products, at some point it will bounce back. Medtronic needs to be ready for these demand spikes in preparing for transportation and distribution capacity. So, as one example, Medtronic ran a total landed cost scenario for how to meet potential surges in demand for its cardiovascular products. If it did nothing, the warehouses would be running near 95 percent capacity.
There were various levers to solve the warehouse capacity bottleneck: the company could put inventory in overflow warehouses, it could slow down manufacturing, change where certain products were manufactured, or it could slow down shipments by using ocean carriers turning the boats—in effect—into floating warehouses. The total landed costs of all these options were examined.
Catching up to the vision
Until the last few years, the technology needed to support manufacturing supply chain agility, such as demonstrated at Medtronic and other forward-thinking manufacturing companies, was not available. But finally, the technology has caught up to the vision. Companies are using these solutions to great effect. The early adopters have been rewarded with a greatly enhanced ability to respond intelligently to the COVID-19 pandemic.
Medtronic’s digital twin will be an always on, constantly updated model, that will improve not just the company’s day-to-day operations, but its business continuity capabilities.
We want to hear from you! Please send us your comments and questions about this topic to InTechmagazine@isa.org.