Engineer your supply chain – using supply chain network design to determine the right structure for your supply chain
“Do you let your supply chain grow ‘ad hoc’ over time, or do you take control?”
Companies know that the performance of their supply chain is of vital importance to the overall competitiveness of the organization. Yet most supply chains have grown unplanned, a tactical decision here and a merger there, but never looked at as a whole. More and more companies start to realize that the structure of their supply chain is too important to be left over to chance or ad hoc decision making. They recognize that the time has come for a disciplined, engineering approach to the design of their supply chain network.
While supply chain network design has been around for some time, a lot of companies are adopting it now because a few trends have come together. Due to growing uncertainty and volatility in the business environment, combined with supply chains that have grown more global and complex, the need to relook at supply chains that are more resilient is there. At the same time many supply chain organizations have matured to the level where they are ready to use more sophisticated technology, sufficient data is available in their ERP systems, and the software tools available in the market have become more user-friendly.
So with the need there, and the tools available, now is the time for your company to start to architect your supply chain for the future.
Basics of Network design
A supply chain is defined by the suppliers, plants, warehouses and flows of products from each origin to the final customer. The number and locations of the facilities is a critical factor in the success of any supply chain. Approximately 80% of the cost of running the supply chain is locked once the location of facilities and product flows between them is decided. An engineering approach is therefore required to strategically plan and design such network.
“Supply chain network design is the disciplined process of determining the optimal location and size of facilities and the flow of products using a mathematical modeling approach.”
As supply chain networks have increasingly become global and more complex, companies need to be able to analyze the impact of factors such as:
- External factors
- Changes in government regulations (such as tax, incentives) that impact the supply chain
- Changes in demand patterns, business environment etc.
- Internal factors
- Growth plans that require a need to check the suitability of the network to support the plans
- Introduction of new revenue streams (new products, new markets etc.) and the need to design networks to support them
- Merger & Acquisitions and need to rationalize and redesign network for new entity
- Need of companies to periodically revalidate and rationalize the existing supply chain network
The figure below shows the typical questions you can answer with supply chain network design.
Decisions about the location of your facilities impact many aspects of your business and require you to make trade-offs. This includes:
- Transportation cost
- Location drives distance to transport goods, and available transport infrastructure
- Service level
- Distance to customers is a major factor determining how fast you can get your products to your customers
- The number of locations for a critical activity (sourcing, production etc.) determines the level of risk in the supply chain to be impacted by major disruptions (fire, natural disaster, strike etc.)
- Availability of labor, skills, materials
- Location determines the labor cost, availability of required skills, percentage of how many raw materials you can buy locally etc.
- Taxes depend on where location is located and what type of operations are done.
- Shipping to and from locations also impacts import/export taxes
- Carbon emissions
- Locating facilities to minimize distance travelled has a side benefit of reducing carbon emissions
The objective of network design is to balance service level against total supply chain cost, which consists of:
- Production/ purchasing costs
- Inventory carrying costs
- Facility costs (fixed and variable)
- Transportation costs
Some of these factors cannot be quantified (easily) and are thus not considered directly. This does not mean that optimization is meaningless, nor does it mean that you should ignore these factors. The right approach is:
- Run the optimization with as much quantifiable information as possible
- Run a variety of scenarios to give a range of potential solutions
- From this range, then apply consideration to the non-quantifiable aspects
For example if closing an existing facility and opening a new one in a new location increases your political risk, then the result from a network study indicating that the savings are 500 K or 50 Million enables you to judge whether the extra risk is worth it.
Type of software available in the market
There are a number of commercial software packages in the market for network design. This kind of software models the real world supply chain as a series of nodes for sourcing, production and distribution locations with associated product flows and adds the relevant constraints and parameters such as capacities, lead times, costs etc. It allows you to specify what is the objective of the network study (such as highest service level, lowest cost etc.) and then use mathematical optimization techniques (powerful algorithms) that evaluate all possible alternatives in order to find the best result.
Network design software requires various data from the supply chain to be collected. This typically includes:
- A listing of all products
- Location of customers, stocking points, plants and sources
- Demand for each product by customer location
- Transportation rates
- Warehousing costs
- Shipment sizes by product
- Order patterns by frequency, size, season, content
- Customer service goals
Because strategic decisions based on the network design study have to be valid for the next 3-5 years, all data needs to be collected both for the current and future operations. This means that future projections will have to be made for:
- Customer demand
- Transport cost, incl. new sources and destinations
- Other key parameters, such as % shipped by truck/rail, % customers that orders FTL vs LTL, etc.
Typically a scenario approach (for example optimistic/realistic/pessimistic) is used to analyze a range of expected future demand (and other parameters) over the planning horizon.
There are two types of software for network design: off-the-shelf and custom-configured. While off-the-shelf software can be immediately used, the drawback is that it may not be able to take into account the unique constraints and business rules of your company. Custom-configurable software allows you to take into account the specific business objectives and constraints of your company, but requires (at least initially) more support from external consultants. Note that the price of custom software is not necessarily higher. Modular and customizable software allows you to start small and customize towards your unique requirements, and then gradually grow in complexity (and cost) over time.
However, keep in mind that optimizing your company’s supply chain requires more than just the software tool. It is the “soft or knowledge” part that determines the overall success:
- Quality of translating real business operations into a model
- Quality of the “what if” improvement scenarios defined
- Ability to assist with following up on the recommendations of the model: the implementation & change management efforts
Thus the choice of implementation consultant is as important (if not more) as the selection of the software.
Implementation of supply chain network design software requires a project approach with the involvement of the various stakeholders such as senior management, representatives from the business functions and IT and/or Analytics staff. The figure below shows the typical steps that are done.
While a fair amount of the effort is on data collection and the ‘technical’ part of the project, it is of greater importance that the ‘business’ side of things is at all times the focus. This starts with a clear definition of what business questions need to be answered with the study. It requires the buy-in of senior management to not only support the smooth execution of the project but also afterwards take ownership of the results, and pay attention to the change management aspects of following up recommendations from the study with real changes to the supply chain during the actual implementation.
Art of modeling and optimization: how to avoid common pitfalls & mistakes
While the algorithms inside the network design software are definitely advanced science, doing a supply chain network study is still both an art and science.
Knowing how to model a supply chain and in what level of detail to get the right accuracy of results requires experience. The following best practices can be followed to avoid common mistakes:
Best practice 1: Experiment, run lots of scenarios
By running as many different scenarios as possible, you can better prepare for the future. Because you need to use forecasts for demand and projected figures for costs in the future, it’s best to run lots of what-if scenarios to test the impact of changes in:
- Supply / distribution patterns by region
- Gaining / losing new customers
- Cost, for example what if petrol price goes up?
- A study resulted in identified transportation and warehousing savings of 4M. SGD with no change to service level
- However, when analyzed in more detail these savings were driven by one very large customer. If this customer were to be lost or its demand lower, the new network would be 0.5 M. more expensive than the existing structure
- By coming up with alternatives, a solution was found that still saved 3.2 M. SGD if the customer stayed, and 2.8 M. if the customer was lost
- In this case, 800 K. savings was given up, but resulted in a more robust solution
Best practice 2: consider carefully the constraints
When setting up the constraints you need to carefully strike the right balance between enough constraints to capture reality of the business versus restricting too much so that model is limited in the optimized results it can return. Sometimes a lack of constraints can lead to discover solutions that initially you thought were not possible. This can lead to significant savings, if the change management aspects are accepted and well managed. An overrestricted model can lead to results close to the current state. To avoid this, make sure that constraints are ‘real’ constraints (such as legal/physical limits), not just current business practices.
Best practice 3: Don’t forget the softer elements of the study
In some cases a network study is done too much in isolation of the rest of the organization. In such cases the study can produce great results, but they are not adopted by others, leading in a worst case to no implementation at all. By including also the ‘softer’ aspects related to executive sponsorship, impact of the results on the organization (in terms of people, process & technology), change management issues etc. this can be avoided. For example, using a structured approach for Supply Chain improvement such as SCOR is recommended.
“Does your global supply chain network still make sense if China’s labor cost increases by 20 percent, oil costs $100 a barrel, and shipping lanes have 25 percent excess capacity?”
Latest trends in supply chain network design
There have been two key trends in recent years:
- Use your supply chain network as a hedge: shift away from ‘lowest total cost’ focus, to include other key performance attributes such as resilience and risk management
- Growing awareness about importance of network design and growing adoption by companies
Ad. 1. Use your network as a hedge
The traditional reason to do a network design study was to minimize cost while maintaining or improving customer service. However, the ability of supply chains to withstand a variety of different scenarios could influence the profitability and even the viability of organizations in the not-too-distant future.
It’s critical for organizations to determine which of the many questions like these are right to ask and to invest energy in understanding the global trends underpinning them. Companies should design their portfolios of manufacturing and supplier networks to minimize the total landed-cost risk under different scenarios. The goal should be identifying a resilient manufacturing and sourcing footprint—even when it’s not necessarily the lowest-cost one today. This approach calls for a significant mindset shift not just from operations leaders but also from CEOs and executives across the C-suite.
Ad. 2. Growing adoption of network design
It used to be (and it might still be the case in many companies) that supply network optimization would be run yearly by a few statistics geeks. But the number and significance of changes and volatility nowadays means that being more alert and adaptable when it comes to one’s supply network can often differentiate the winners from losers. Smart companies are making supply chain design a core business function, and use this competency as a competitive weapon for their business. Often they will establish a centralized supply chain design function to continuously optimize their end-to-end supply chains.
This results in:
- Expanded scope of what is being analyzed
- In the past, companies would do a network optimization to analyze the structure of the supply chain. Now they are doing that hand-in-hand with inventory optimization, with vehicle routing for multi-stop truck runs and with simulations that show how their supply chain would be impacted by low-probability, high-impact events
- Supply chain design is becoming much more granular, towards the SKU level rather than the product family level
- Companies typically did a supply chain design project every one to three years. The trend is to increase this frequency and to look at tactical as well as strategic decisions every year and sometimes even half yearly
- Model at lower level of detail
- Model more frequently
Building an in-house capability for network design
As more and more companies realize the importance of network design, they decide to make supply chain design a core in-house competency to be used as a competitive weapon. They are using modeling technology to continuously optimize their end-to-end supply chain to improve service levels, identify major cost savings, reduce risk and stay ahead of their competitors. As a result they are outsourcing fewer of these supply chain design activities and doing more in-house by building up their internal teams of supply chain design experts.
It requires the classic components of people with the right analytical skill set, running well structured and repeatable processes, supported by the latest technology. Another vital ingredient is continuous education, as network design (but also wider supply chain analytics & optimization capability) requires well-educated staff.
Companies also need to consider the right organization structure, which can change over time and evolve with the growing maturity level of the organization in using Analytics and Optimization. A typical path is for the company to start with separate analytics teams supporting the various business functions (such as supply chain, sales & marketing). In order to standardize processes, technology and also recruiting, a Center of Excellence can be started. Such Center would support the various analytics teams in the other business functions to increase learning and adoption, reduce duplication of efforts, standardize IT used, and increase the visibility in the organization. As the company’s analytical capability matures, a final step can be to create a separate function for Analytics in the company that takes care of all the analytics requirements through centralized data structures, business processes, recruiting policies etc.
Building an in-house capability for supply chain network design does not have to be done overnight. It is pragmatic to follow a step-by-step approach that begins with a small group finding tactical quick wins to build momentum, to creating a Center of Excellence, to eventually driving key strategic decisions in the supply chain.
- Even though supply chain optimization can identify major cost savings, the recommendations can often be disruptive and time-consuming to implement (close factories, open new DC, rationalize products etc.)
- In order to establish credibility quickly, companies can identify quick-win projects that are easier to implement and still deliver significant cost savings (DC-customer assignments, inventory right sizing etc.).
- Quick multi-million wins can gain executive attention and establish early credibility to justify further investment in staff and technology
Establish a Center of Excellence (COE)
- The staff involved with Supply Chain Optimization need to be able to address the entire business to optimize the end-to-end supply chain and not just a specific business unit or function
- Establishing a Center of Excellence (COE) can pool talent and technology to provide optimization capabilities to the entire organization
- After some tactical wins, the Supply Chain Optimization team should explore what is truly possible through modeling, optimization and analysis of the end-to-end supply chain
- By removing pre-conceived business constraints and designing new green-field supply chain operations, this could lead to game-changing business practices and create competitive advantage
Companies that do a network design study typically report savings in the range of 5-10% on their overall supply chain cost. In most cases it also results in higher service levels and a supply chain that is more resilient to volatility and unforeseen events. If your supply chain in Asia has reached a certain size and complexity (at least 2-3 plants and 5-10 DC), now is the time to start engineering your supply chain!
This article has previously been published by Logistics Insight Asia