Supply Chain Modeling: What Does the Future Hold?
Sophisticated supply chain modeling and simulation tools help companies predict the outcome of business scenarios and foretell the future.
It was 2008, and JBS S.A., a Brazil-based meat-processing company, was still managing its supply chain and analyzing data to make future decisions using basic spreadsheet tools. Key leaders understood the company had outgrown manual spreadsheets, and it was time to embrace a more sophisticated model.
JBS manages more than 40 industrial facilities in Brazil, each with a range of available resources, capabilities, and certifications. The company moves 1,500 different SKUs per month. “The complexity of our operation is considerable,” says Celso Batista, coordinator of planning and control at JBS in Brazil.
Eight years ago, it became clear to JBS that its low-tech approach to supply chain management caused it to miss out on crucial opportunities because of hazy forecasts and an inadequate ability to study the full spectrum of possible scenarios, Batista says. The company had grown too big and intricate for its old-school modeling and forecasting methods.
JBS elected to radically alter its approach and wade into supply chain modeling and simulation. Wanting a flexible software option that could adapt to its unique circumstances, JBS decided to work with AIMMS, a global decision support technology company that offers an analytics, modeling, and optimization platform. Supply chain modeling helps organizations visualize the future of their supply chains in new ways to identify opportunities, inefficiencies, and risks.
“We sometimes lost track of supply chain opportunities, for example the type of return we can get by changing a location,” Batista says. “We were seeking new ways to identify all the opportunities our business affords.”
Today, the information that companies can integrate into modeling software applications to develop forecasts and investigate hypothetical scenarios is impressively thorough. For instance, LLamasoft’s Supply Chain Guru modeling software incorporates attributes such as production facilities, individual lines, capacities, distribution facilities, suppliers, retailers, costs, demand volumes and frequency, sourcing and inventory policies, and transportation modes, among other components, says Toby Brzoznowski, co-founder and executive vice president of Michigan-based LLamasoft, a supply chain optimization software provider.
When the model is right, and properly integrated with accurate and comprehensive data—such as through an existing enterprise resource planning (ERP) software—organizations can begin to mimic the real world in a way that allows them to study the future with detailed foresight. They can make decisions conceptually, and peruse projected results in graphs, dashboards, maps, charts, and videos, without any risk.
A Sunny Forecast
Advances in the field mean that forecasting is no longer limited to occasional work that requires many hours of staff time to investigate a single scenario. Forecasting now can be performed routinely and swiftly, producing answers to complicated questions in minutes.
For example, organizations can calculate new safety stock levels daily or weekly, instead of automatically maintaining a four-week safety stock level. This results in more precise, less wasteful supply chain management, and frees up cash for other investments, such as developing new products or exploring new markets.
“Companies can start playing with reality without actually affecting reality,” explains Marcel Mourits, supply chain optimization lead at AIMMS, a global company offering integrated multidimensional modeling software. “They can start asking the important question: What if? What if we close this warehouse? What if revenue grows by 20 percent? What if we start using multimodal shipping instead of relying only on trucking?”
After making a decision, organizations can use modeling to anticipate possible events resulting from that decision and to develop plans to prepare for them.
For instance, if an organization changes its supply capacity, it can play out a random series of demand scenarios for an upcoming period of time to see what likely will happen. The organization can then identify where both opportunities and problems will develop. Do certain scenarios create inventory or capacity issues? Is a particular product vulnerable to some developments? Does production in a certain country meet special challenges? The supply chain model reveals all.
The model also gives supply chain planners and designers a crucial head start.
“Simulation allows companies to determine how best to respond to a problem if it occurs, and avoid surprises,” Mourits says. “Simulation software lets companies spend the time now to either mitigate the risk or, if it happens, to respond as quickly as possible, so the situation doesn’t deteriorate.”
“Modeling and simulation software is all about making decisions, or evaluating the repercussions of a decision, in the digital world before acting on it,” Brzoznowski adds.
Modeling and Simulation at Work
Glass is among the materials Nampak uses in its large packaging operation based in South Africa. Glass production represents a nettlesome planning challenge because the process is so delicate. The company uses large, extremely hot furnaces to transform sand and other ingredients into glass bottles for clients around the world. The conditions need to be precise depending on the type of bottle—its size, shape, thickness, and color. As a result, different bottle types cannot be created in the same furnace at the same time, or quality suffers. And if Nampak doesn’t properly build its schedule, delays and gaps in production time can result.
Because of this complexity, Nampak faces tough adjustments when circumstances change. For example, when the company added capacity by installing a new furnace, it created numerous new planning questions that continually arose.
To address those challenges, Nampak decided to work with AIMMS. Now, if a client calls to change an order from its usual 100,000 bottles per week to 150,000 bottles one week, Nampak can use the software to explore that request. It can study its decision based not solely on cost, but also taking into account revenue, scheduling complexity, margins, and other factors.
The software could show, for example, that fulfilling the request creates a domino effect that would disrupt the schedule for several other clients and prove less profitable than declining the request or only partially fulfilling it. Or, the software could reveal that accepting the request and postponing orders for other clients would produce a windfall.
“What’s transforming for Nampak is that it now has the option to choose the most profitable production schedule for each customer request,” Mourits says.
Like manufacturers, retailers also employ sophisticated modeling tools to develop supply chain forecasts. One such retailer is Michael Kors, a designer of luxury accessories and ready-to-wear apparel. In recent years, the company has enjoyed rapid growth, resulting in a steady addition of new stores. As it looks to the future, Michael Kors uses LLamasoft’s Supply Chain Guru to anticipate how continuing growth affects the supply chain, and how it needs to be adjusted to accommodate that growth.
In particular, Michael Kors can layer its forecasted demand growth on top of demographics, and then use the supply chain software to identify where it will likely face capacity constraints over a particular time, when and where it will need to add capacity, and how to optimize for that over time. This information helps the company prepare to properly support projected growth down the road.
Similarly, when the retailer wanted to study the implications of a more robust e-commerce strategy, it used supply chain modeling to forecast which stores would service the online orders and then need to be replenished.
Changing with the Times
Organizations vary widely in the emphasis they choose for their supply chain models. They also diverge through cycles when times change and company goals shift with them. For example, Brzoznowski says automakers, including LLamasoft clients Ford and General Motors, used to optimize their supply chains to cost during the recession when the industry encountered tenaciously poor demand. Today, though, as the economic picture has improved, the industry is more apt to optimize to service.
Through supply chain modeling, many organizations learn that they have been squandering opportunities that have gone unnoticed. For instance, JBS used its modeling software to consider a spectrum of scenarios for one of its products, and was baffled when the AIMMS application suggested the company should be producing a much higher volume of that product than historical data indicated was appropriate. The finding seemed far out of step with what JBS understood—and had understood for years. Stumped, JBS leaders investigated their historical data more closely, and discovered an error within its ERP software. The error caused the company to view the decision improperly and miss the opportunity that the product’s underproduction represented.
JBS consequently revised its approach and increased volumes to the levels the software proposed. The change led to a project payback within just two months. This example was one of many that convinced JBS “we could be missing simple data that represents a lot of value for the company,” Batista says.
gaining buy-in is a key challenge
Learning to use the new tools, and understanding how to integrate them into the company’s operations, proved less challenging than securing buy-in within JBS, including among leaders from a number of different departments. In the early going, when the AIMMS platform suggested alternatives that conflicted with solutions the more traditional methods produced, there inevitably was pushback. Some decision-makers were wary of setting production and sales plans based on data they still did not understand. “It’s difficult to change,” Batista admits.
The value and potential of the forecasting tools, however, became evident. “Heated discussions always revolve around information conflicts, but the modeling system survived,” Batista says. “We improved the quality of the data we were moving, so everyone could see the solution’s benefits. The barriers came down.”
“The single biggest challenge is the quality of the data,” agrees Brzoznowski. In the 1990s, it would have been impossible to implement these kinds of projects because of the lack of data standardization within companies. One benefit of Y2K (when 1999 turned to 2000, it was feared that computers would shut down completely), is that it drove many companies to bring their data under a centralized system.
Still, Fortune 1000 companies frequently maintain a number of different ERP software applications, often because of acquisitions made over time. Organizations and service providers must collaborate to access these different data streams and blend them to optimize the supply chain.
Coming Up Short
Data shortcomings can slow a move to modeling software. Just ask A.W. Chesterton, a Massachusetts-based manufacturer of sealing solutions, which has undergone a recent major shift, employing business intelligence to improve data management and analytics.
The next step the company wants to take is to add modeling and forecasting software, says Tom Meier, vice president of information technology at A.W. Chesterton. However, the company still has work to do in that area because it only recently adopted data integration capabilities.
“We need more years of historical data to support it,” Meier adds.
Discrete event simulation is another forecasting tool that would have been unthinkable not that long ago. Through discrete event simulation, users can observe where problems will emerge for a specifically defined event before the event occurs, and then adjust accordingly. This differs from supply chain optimization modeling based on averages and flows.
Companies can use this form of simulation, for instance, when designing a new manufacturing plant and wanting to identify where bottlenecks could occur in the internal transportation system, such as among moving products, people, or vehicles. The simulation allows companies to run the events through a model almost as though they were happening.
Sometimes the simulation displays a 2D or 3D image that enables users to observe waiting lines, moving equipment, and congestion in real time—all before the facility has even been built. It is like watching a movie before it has been released.
Users can choose to run the event for a specific time to see how circumstances evolve and how the operation behaves. If key problems arise in the model, then users can change their plans. It is a way of ensuring supply chain design changes perform as expected. In the case of the hypothetical manufacturing plant, the model can demonstrate if the interior design leads to poor flow of resources, such as a crossing that forces vehicles to wait too frequently or even leads to potential collisions. The user can then alter the design to prevent the problem, and run the new design through the model for fresh testing.
Asking Better Questions
After switching to supply chain modeling and optimization software eight years ago, JBS today can respond rapidly to market prices that change daily for its products, and make comparisons with the newest demand scenarios, production capabilities, transport possibilities, and different types of costs.
In the future, Batista hopes JBS will be able to better align its use of the software with its processes and people. “The main factor for continued business success is to tighten the relationship among the software, its processes, and our people,” Batista says.
One key quality of the modeling software is the way it forces businesses to challenge the kind of assumptions and conventional wisdom that can lead to lazy thinking. No matter how JBS uses modeling to peer into the future, the process and tools will challenge the company and force it to scrutinize itself in a way spreadsheets never did, Batista says.
“Modeling software makes companies question their thinking, and get creative,” he says. “It’s not the kind of system that simply provides production plans, and all the user has to do is push ‘print.’
“Solutions that make you think, such as modeling and forecasting software, are the best ones for your business,” he says.
Off-the-shelf vs. Tailor-made
Supply chain modeling and optimization software can be organized into two primary categories: off-the-shelf and tailor-made. The solution that’s right for you depends on your organization.
Off-the-shelf software uses a standardized application. One example is LLamasoft’s Supply Chain Guru, which is designed for use by companies across a range of industries while still meeting their individual needs.
Tailor-made software offers companies the opportunity to get creative based on their unique characteristics and objectives. Marcel Mourits, supply chain optimization lead at AIMMS, says he is constantly surprised by the applications companies develop on their own for AIMMS’ tailor-made software solutions.