Dawn Mathew Varghese, a Senior Consultant at Infosys, writes, “Supply Chain Analytics is a very hot topic today and has gained considerable mindshare among our customers too.” [“Supply Chain Analytics Fact, Fiction or Fantasy,” Supply Chain Management, 14 June 2011] Varghese goes on to assert that the term “supply chain analytics” means different things to different people. She compares the current situation to the old Indian fable of the blind men who were asked to describe an elephant. She continues:
“Too many areas/subareas have been attributed to analytics and [have] claimed to be a part of analytics. Broadly what analytics leads to is superior business performance through data driven intelligence. In order to achieve the different levels of intelligence (simple to advanced predictive analytics), it requires an organizational dimension based on inputs in terms of Processes, Policies, Procedures and Practices. … Also it requires a computational dimension fired by data. Both of these dimensions form the basis of the analytical intelligence an organization can leverage on. It is this intelligence which leads to insights for a supply chain planner or a warehouse manager for him/her to act upon in such a way that it leads to superior performance.”
Varghese touches on several important points. First, analytics is not about computational power (although that matters) it’s about turning data into knowledge. Data cannot be obtained only from structured sources, but it must also be drawn from her 4Ps (i.e., processes, policies, procedures, and practices). Some (or most) of that data is unstructured, which makes it much more difficult to include in the supply analytics process. Excluding unstructured data, however, provides an incomplete picture and could lead to erroneous results. It certainly doesn’t lead to better knowledge. Second, Varghese is right that the best analytics have an organizational dimension. I would go further and state that all business intelligence (BI) needs to be drawn from the same pool of data, whether it be used for supply chain purposes or any other business activity. As supply chain analyst Lora Cecere has commented, data should collected once and read many times. You will never achieve corporate alignment if everyone is using different data sets. Finally, good analytics requires good data. From the beginning of the computer age, people have understood the adage “garbage in, garbage out.” Varghese continues:
“Organizations worldwide are keen to capitalize [on] this new way of doing business. Organizations that started with merely reporting … are today looking at exploiting analytical prowess to have end to end visibility into the extended supply chain [in order to] enable management by exception.”
She points out that a good automated system will allow decision makers not only to track shipments but determine if they are going to be late and why. Such a system will also be able to provide insights into future consequences of disruptive events, like earthquakes or hurricanes. She continues:
“Today’s Supply Chain innovators are trying to be pro-active and are clearly heading towards a notch higher than management by exception. Classical management by exception is the fundamental premise for all Supply Chain Visibility programs. This is enabled by defining events based on the business process, understanding and establishing norms which fit into the definition of an exception and generating action points (based on breaching/meeting of norms) for supply chain managers. This framework [is] encapsulated in most IT enabled supply chain solutions today [and it] is helping organizations to drive an end to end visibility of their supply chain. The information derived out of historical data can further be organized, sliced, diced and served on a dashboard for insights into key performance indicators.”
The importance of business intelligence visualization should not be underestimated. With mountains of data being generated and analyzed, analytical insights are only useful if they get into the right hands in the right format at the right time. Visualization expert Stephen Few writes, “Big, old, traditional BI companies are good at producing technologies that enhance the infrastructure of business intelligence—more and faster—but not the actual use of data in ways that lead to greater intelligence.” [“Old BI and the Challenge of Analytics,” Visual Business Intelligence, 7 March 2011] He says that to be useful, BI systems must “support decision making: data exploration, sensemaking, and communication.” In order to achieve this objective, systems must focus as much on “the humans who use it” as on the technology that drives it. Varghese claims, “Few Supply Chain and collaboration heavy businesses, like 3PL companies, are exploiting visibility and analytics as a strategic weapon by conceptualizing a control room/tower (akin to an air traffic control room or a turbine control room of a nuclear station) which will give them round the clock visibility to the extended supply chain, help monitoring supply chain events and KPIs and help control and mitigate risks them to a large extent.” To learn more about the concept of “control towers,” read my post entitled Have You Heard about Supply Chain Control Towers?
The editorial staff at Supply Chain Digest believes that analytics can help manage inventory in a more sensible and cost effective way; but, even with good analytics, managing inventory is daunting challenge. [“Can Smarter Analytics and Optimization Finally Reduce the Out-of-Stock Challenge in the Consumer Goods to Retail Supply Chain?” 25 May 2011]. The authors write:
“For at least two decades, consumer goods manufacturers and retailers alike have been focused on reducing inventories and out-of-stocks on store shelves through such initiatives as Efficient Consumer Response (ECR), Collaborative Planning, Forecasting & Replenishment (CPFR), RFID and other programs – with somewhat mixed results. While each of the programs delivered some of the expected results, that was usually accompanied by the sense that some of the potential benefits from these initiatives were left on the table. In fact, inventory levels in consumer goods manufacturers stayed flat throughout most of the 2000s, and out-of-stock levels at the shelf have also been resistant to improvement.”
Through interviews I’ve had with corporate executives, I can attest to the fact that these challenges remain. Reducing inventories and out-of-stocks on store shelves are areas with which executives have asked my company to help. The SCD article continues:
“A well-publicized 2007 study by Dr. Thomas Gruen of the University of Colorado and Dr. Daniel Corsten of the IE Business School Madrid, based on funding from Procter & Gamble, estimated that manufacturers lose something close to $100 billion in sales annually due to out-of-stocks at the shelf. Solving the out-of-stock challenge can therefore pay big financial dividends – especially for the companies that can reach new levels of in-stock performance first, before competitors do.”
A hundred billion dollars in lost annual sales due to out-of-stock situations is an amazingly large number. Neither manufacturers nor retailers are happy losing that kind of business. The SCD staff notes that this doesn’t have to be the case. It reports that “one beverage company IBM recently worked with was able to increase its sales by 12.3 million cases by significantly reducing its out of stocks, in this case throughout its distribution network that served local stores.” It was able to accomplish this through better analytics — specifically using inventory optimization software. The article continues:
“What is inventory optimization software? In short, it is a tool that looks at how inventory should best be positioned at different nodes and levels of the supply chain holistically, rather than just optimizing each node/level individually, as is the case in most traditional supply chain planning environments. The result can be significant reductions in inventory with constant or even improved customer service levels.”
In other words, inventory optimization software helps determine where the inventory is and where it should be. By getting the analytics correct, inventory velocity can be increased and out-of-stock situations can be reduced — a win-win-win for suppliers, retailers, and customers. The article continues:
“One place to start with inventory optimization is an analysis that shows not only what SKUs have too much inventory … versus demand, but also what SKUs have too little. … [Better inventory management] is accomplished by changing inventory policies and safety stock rules for these SKUs at different levels of the supply chain. [IBM’ Dr. Michael] Watson … noted that the SKUs that have too little inventory usually result in costly mitigation strategies, such as last minute production schedule changes or expedited transportation. Of course, no company plans to have too much inventory of some SKUs and too little for others. Then why is it so hard to get network inventory levels right?”
The simple answer to that question is: because too many supply chains don’t have enough visibility. The article goes on to discuss a “number of variables that must be considered,” including:
• Forecast error and demand variability
• Lead time and lead time variability
• Order and production cycles
• Committed times from manufacturing or vendors
• Transit times and transit variability
• Order minimums and increments
The article notes that this is not a simple matter because “those factors must further be considered across increasingly complex supply chains that have multiple levels or ‘echelons’ and often involve multiple production steps (e.g., intermediate good production, then co-packing operations).” It continues:
“It is this total complexity that almost requires strong technology support to manage, as optimizing across all these variables for hundreds or thousands of SKUs is beyond what humans and spreadsheets can accomplish.”
I’m surprised the SCD staff wrote that this complexity “almost” requires strong technology support. It definitely requires strong technology support. Numerous supply chain analysts are pushing the notion that supply chain complexity must be decreased. In many cases, believing that can be done is irrational. However, the proper use of supply chain analytics can provide the perception that the complexity has been reduced because the technology deals with the complexity and displays only pertinent knowledge.
Dr. Watson told the SCD staff something that I have been stating for some time, “Lean manufacturing strategies can also sometimes conflict with overall inventory optimization.” When a system becomes too lean it becomes brittle. This is not a good trait when most supply chain analysts believe that the best supply chains need to be agile and flexible. The article provides an interesting example of how a lean practice can actually cost money. The article concludes:
“Efforts to decrease raw material levels at the plant along Lean concepts can actually result in more finished goods inventory in the network, which is more costly than raw materials or WIP inventories further back up stream. Inventory optimization will help identify that optimal balance and where inventory buffers are most effectively maintained. Watson also noted that the right approach to inventory optimization means companies will check plans and inventory policies at different ‘cadences’ throughout the year. For example, more strategic looks at the role of inventory in a business unit might be performed annually, while more parameter ‘tuning’ activities might be performed quarterly, monthly or even weekly depending on the type of task.”
A good BI system can help determine what cadences are required as well as help with inventory allocation itself. The bottom line is that big data is becoming more available and, when used correctly, this data can help companies more effectively manage their inventories to save money and increase sales.
