How SKF uses a supply chain twin to enable integrated planning

At first glance, it might seem like SKF Group has nothing to worry about. With revenues of US$9 billion, we are currently the largest player in the bearings and rotating equipment market. We also have more than 100 years of history in the industry, which would make it hard, you would think, to displace us from the top of the heap. (For more information about the company, see the related sidebar, "Five quick facts about SKF.")

We knew better, however. In 2015, we realized that in spite of all our past success, our supply chain was in need of a transformation. We believed that one of the first steps that we needed to take in order to modernize our supply chain was to move from having a local planning structure that occurred at the manufacturing plant to having a global, integrated planning structure. To accomplish this, we found that we needed to create a "digital twin" of our supply chain that could enable more automation of our planning process. This is the story of our journey.

The threat

Bearings may seem like a relatively simple machine element with a very specific job: reducing friction between moving parts while also constraining relative motion to only a desired amount. But in spite of that perceived simplicity, the bearing market is set for a high growth rate. According to the market research group Stratistics MRC, the global bearings market is expected to grow at a compound annual growth rate (CAGR) of 7.2 percent from 2017 to 2023. The research firm cites multiple factors behind this growth including a rising demand for ceramic ball bearings in electric vehicles and large-scale railway, aerospace, and wind power projects. The top threats cited are cheap and fake products entering the market, along with the growing utilization of used bearings.

With this level of growth potential, the bearings market is set to attract increased competition. In fact, it's not only our traditional competitors that keep me up at night these days; it's the threat of a completely new business model disrupting our market. I look at what happened over the last years in different markets: Airbnb is the world's largest housing provider, yet owns no hotels. Uber is the world's largest taxi company, yet owns no vehicles. And while these companies and many other recent disrupters are now acquiring assets, who's to say such a competitor with new ideas and a different business model won't emerge in our space, significantly disturbing our growth or even threatening our existence?

It can be tough to convince people of the need to change before a crisis actually hits. But the larger and more set in its ways a company becomes, the harder it is and the longer it takes to change. The sudden demise of once great companies, like the photography company Kodak, provides a valuable lesson: Large-scale transformation must begin well before a major threat appears on the horizon.

Fortunately for us, a new CEO, Alrik Danielson, joined SKF in late 2015. He came in to shake up the status quo and asked: "What is our demand chain vision? What is our guiding star for the next 10 to 15 years?" He encouraged us to take inspiration from any source and not be constrained in our thinking by any technical, organizational, or other limitation. "Start with a blank sheet of paper!" So we did.

The vision

In setting our vision, we not only tapped into some of the most experienced brains in our company but also reached out to leaders in manufacturing and other industries like fast-moving consumer goods. In addition, we contracted with the supply chain consulting and applications company Optilon to support us in this journey. Early in 2015, a diverse team of supply chain experts, visionaries, planners, and information technology and data experts worked up a rough sketch of our vision, and it literally fit on a single sheet of paper. That vision was to consolidate and integrate planning on a global basis.

We felt that moving to a global planning structure was our highest priority because we had observed that doing planning at the local level led to operational inefficiencies and friction between internal organizations. An integrated planning process would optimize how we made daily, operational decisionson such matters as production, stocking, and distribution. For example, it could help us better decide how many of a lot of 1,000 items we should ship to Shanghai, and how many we should ship to Mexico. Or, we could use it to decide what product should be produced next week based on the demand we are seeing now from customers around the world and the inventory we currently have in different warehouses. These decisions would no longer be made based on what would be best for a particular location or function but rather based on what would be optimal for the supply chainas a whole. 

It was just as well that our CEO had asked us to brainstorm without limitations, because we knew this would be a mammoth transition. First off, there was a lot of complexity involved. SKF has grown both organically and through acquisition. Consequently, over the years, we have accumulated many manufacturing sites, distribution centers, and warehouses. These new sites introduced more complexity into the planning process, especially as we were running different enterprise resource planning (ERP) systems in different regions. The transition would also affect many people and their incentives, and it would require many of our employees to develop new skills.

For example, we decided that to do integrated planning effectively we needed to create a new role, the global planner, who would see, manage, and supervise all of the material and information flows associated with our products, including finished goods, components, and raw material. The global planner would have end-to-end responsibility for these flows from analyzing customer demand up to submitting purchase orders to suppliers. Orchestrating all the associated activities, however, is a lot of work. The only way the global planner could do it efficiently was if we achieved a much higher degree of automation throughout our supply chain, which would be enabled by what's known as a "digital twin" of the supply chain.

The digital twin

Our supply chain's large size and complexity means that it's difficult to consistently make optimal trade-offs at the planning level between "cash" (such as safety-stock levels and goods in transit), "cost" (such as staffing and output of production and transportation methods) and "customer" (service levels). To gain the needed visibility to make these trade-offs and feed accurate, timely data into our new planning system, we immediately realized that we needed a "digital twin" of our supply chain, or a digital model that's an exact copy of the entire supply chain. Essentially, it is a cluster of structured tables of master data and operational data.

We needed to have this structured data so that we could feed it into the automation and visibility tools that would improve our operational planning processes. The digital twin would enable us to create visibility for truly fact-based decision making; to link low-level, operational decisions to high-level strategic goals; and perform simulation and "what-if" analysis.

Fortunately, we already had an "embryo" for this new approach because we had previously worked with our partner Optilon to establish a tactical multi-echelon inventory optimization (MEIO) process. MEIO optimizes complex distribution networks like ours by using simulation and what-if scenarios to determine the right level of inventory to keep in each location, by time period, in order to best serve the customer. Having a simulation-based process in place meant our extended team was already in the right mindset for the cross-functional groundwork needed to build and maintain the digital twin. For example, team members already understood how powerful it was to have structured data and how important it was to optimize end-to-end.

Here are the steps we took to build the digital twin and then apply it to our planning process:

Step 1: Create a supply chain network map. The first step to representing our global supply chain in the digital twin was building what we call a supply chain network map. This is essentially a view of how our items are manufactured, stored, moved, and sold around the world. A key part of this step is preparing data (including cost, service-level targets, lead times, and bill-of-material information) for our roughly 500,000 stock-keeping units (SKUs). This involves pulling in data from about 40 instances of five different ERP systems, then "normalizing" the data (cleaning and making it more consistent), before moving it into a central repository. Taking the time to normalize the data is important. An item might be produced in one facility, stocked in 20 warehouses, and sold through 40 sales operations. Normalizing SKU data ensures that it will be interpreted in a consistent way across the entire network.

This map is not a fancy graphical representation of the supply chain but simply a relationship database model. However, it really made our hearts sing to now have supply chain information at the lowest level of detail readily available at our fingertips.

Step 2: Bring the digital twin to life. Once the map was designed and the SKU data normalized, the next step was bringing the supply chain network map to life by adding operational data: things like open customer orders, goods-in-transit, and inventory levels. This alive, digital twin provides visibility to answer questions like: What's the total number of external customer orders for one item across all our global sales operations right now? How much inventory do we hold of that item in all our locations? How many of those items are currently in transit in all our transportation lanes? These are just a few examples of simple questions that the digital twin can answer about the present; its value really increases when we apply it to plan for the future.

Step 3: Use the digital twin in planning. The next step was feeding the data from our central relational database repository into our supply chain planning software, ToolsGroup's SO99+. We had already been using part of this software for the aforementioned MEIO process to set safety-stock levels. Now we started using it for daily operational planning. SO99+ enables us to optimize demand, inventory, and planning for all the SKUs in all locations in our supply chain. We can now more easily sense demand deviations and address them more proactively. Crucially the software lets us optimize inventory levels against desired service levels so that we are able to compete effectively in our industry without incurring excessive storage, transportation costs, and obsolescence. For example, when we face supply shortages, we are now able to make informed decisions about which warehouses to prioritize sending inventory to, minimizing the impact on customer service levels. The planning tool was in place, but the journey to transform the whole of SKF had just begun.

One production line at a time

To ease SKF through this ambitious transformation, we are introducing our new planning process—and all the associated changes for adjacent departments like customer service, manufacturing, controlling, and human resources—one production line at a time. Today we have two factories—one in Steyr, Austria, and one in St. Cyr, France—where global planners now supervise the worldwide distribution of their product's assortment and balance the supply and demand for it. These planners use SO99+ to focus on handling exceptions rather than planning manually. They are globally responsible for all inventories and customer service levels for their share of SKF's products, which are often stocked in more than 20 warehouses stretched around the world from Latin America and the United States to Europe; Dubai, United Arab Emirates; Russia; Shanghai, China; and Singapore as well as smaller warehouses in places like Thailand and New Zealand.

Using the data from the digital twin, SO99+ automatically calculates each item's external demand forecast, factors in lead times and actual stock levels throughout the supply network, calculates safety-stock levels and net forecasts for every warehouse, and creates replenishment plans to satisfy future customer demand to set service levels.

SO99+ also alerts our planners before any potential problems reach a critical point. For example, if we are in danger of not being able to supply an item to one of our warehouses, the planner is alerted and can then act on this exception before it becomes a crisis. Or, if we observe unusual customer demand, the planner is alerted and able to proactively investigate and manage the suspected deviation.

SCOR-ing well

Soon each product will have one global forecast, one planning method, one person responsible, and true end-to-end accountability. This means we will align much better to the Supply Chain Council's Supply Chain Operations Reference (SCOR) model, a diagnostic tool for benchmarking performance. Before we started, our organizations and people were taking on multiple roles and struggling to perform them all effectively. Now our planners in factories have the global planning mandate. In essence, we have consolidated the Source, Plan, and Make functions of the SCOR model into one organization. This makes end-to-end optimization possible, empowers individuals, and reduces internal competition.

With integrated planning in place, we are looking forward to the next stage in our transformation. It is a big change we are driving, but in the grand scheme of things, we have just laid the foundation. We are already exploring multiple options to extend the horizon of planning upstream and downstream. This might include doing planning for and on the behalf of customers. This planning would be based not only on customer orders but also on signals from SKF sensors in customers' machineries and other demand signals. All of these signals would be connected to our core planning procedures. We expect these changes will allow us to reach greater service levels, operational efficiency, and waste reduction.

After devoting my entire career to SKF, I will probably always find some reason to stay up at night, thinking about how to improve our supply chain. At least now that we have embarked on a journey to transform ourselves, the threat of a major industry disruption taking us by surprise is considerably less likely.