Supply Chain Optimization

Stephen DeAngelis

November 30, 2011

Robert J. Bowman, managing editor of SupplyChainBrain, writes, “Before companies can apply state-of-the-art technology to that crucial exercise [of optimizing their supply chain], they need to redefine exactly what it is they are trying to optimize. It’s not just about cost and time.” [“Rethinking the Concept of Supply-Chain Optimization,” 19 October 2011] Clearly, a company needs to know “what” it is trying to optimize; but, Bowman’s point is different. He’s insisting that companies should reexamine the very meaning of the term “optimization. “Sometimes the path to innovation,” he writes, “lies in finding a new meaning for an old word.” He continues:

“Take ‘optimization.’ That’s been the goal of supply-chain managers for decades. But what exactly have they been seeking to optimize? Most have centered their efforts on cutting cost and time, says Katharine Frase, vice president of industry solutions and emerging business with IBM Research. That narrow view misses some critical, if intangible elements of a modern, best-in-class supply chain. ‘In the past,’ says Frase, ‘people tended to view supply-chain optimization a bit one-dimensionally. They wanted to do routing of trucks, or logistics in the warehouse, or the attributes of vendor management. That was all very important, but it tended to be a bit piecemeal.’ It’s not a question of tossing out traditional metrics, she says. Companies still need to keep a close watch on supply-chain costs, while minimizing the time it takes to get product to market. But they ought to be taking other considerations into account as well, as part of a much broader view of global supply chains.”

Bowman’s (and Frase’s) point is well made. There is certainly a difference between optimizing the supply chain (i.e., making it the best it can be) and simply reducing supply chain costs. The first characteristic of the supply chain that Bowman and Frase examine is one segment of its sustainability dimension (i.e., “a company’s carbon footprint or total energy bill”).

“A focus on this aspect of operations might raise the cost of moving goods, yet yield big advantages in other areas, such as brand image. For example, a food company might want to use more locally grown produce in its supply chain, to position itself as a ‘green’ organization. Implementing that policy while attempting to hold down costs involves a delicate balance, where one key metric has to give way to another. The trick, says Frase, lies in coming up with a formula that’s ultimately most beneficial to the supplier.”

I’m not sure that I totally agree with Frase on this point. In several past posts, I’ve made the point that the only sustainability programs that will endure are those for which a business case can be made. Trading higher transportation costs for better PR doesn’t sound like a good bargain to me. It sounds more like greenwashing. For more on this subject, read my post entitled Regional, Local, and Sustainable Sourcing. I’m in fuller agreement with Frase on her next comments, which involve supply chain visibility and flexibility.

“Other elements to take into account include the impact of weather, as it impacts both product flow and customer demand. Often a high-level optimization strategy must be set aside to address the needs of the moment. A sudden snowstorm, for example, might motivate a retailer to increase its supply of snow shovels. Agile providers need to make ‘on-the-fly’ decisions about where and how they can obtain more product at a moment’s notice.”

For me, that is essence of optimization — the ability of a supply chain to respond in near-real time to emerging events to enhance resilience and profitability. Frase told Bowman that “public image can trump top-line considerations in areas that go beyond environmental responsibility.” She provides an inventory example.

“A chain of bakeries might be focused on minimizing the amount of bread it has left over at the end of each day. In the process, though, it angers late shoppers who can’t find product on the shelves. A smart retailer might up its inventories to please customers, then burnish its image by giving away any leftover loaves to charity. In the end, it reaps a double advantage that wouldn’t have been possible with an approach based purely on cost.”

Again, I’m a bit skeptical of her example. Profits still drive business. Technology should be able to help the bakery chain optimize its inventory and become more profitable. It is much more important to make one sale of an item with a $10 profit than it is to make ten sales with a $0.10 profit per item. By optimizing on-shelf availability of products, both the chain and its customers benefit (and, of course, the brand’s reputation is enhanced). Bowman continues:

“Modern-day supply chains need to be viewed in the context of ‘ecosystems,’ says Frase. She uses the term to describe a web of supply-chain partners that do not always relate to one another in a linear or serial fashion. Moving up and down the supply chain from a consumer-products company, one encounters a plethora of suppliers and distributors, all of whom must be taken into account when optimizing critical business processes. Companies need to visualize this entire universe of participants in real time, Frase says. Information technology systems that help to achieve this ambitious goal are still in development.”

Frase and I are on the same page here. Her “ecosystem” conforms to my thinking about intelligent supply networks. Some consultants like to talk about value chains; but, as I’ve noted before, they are really value networks that involve much more than linearly-connected links. Like the offerings from Enterra Solutions®, IBM’s offering “pulls data from a number of sources, including point of sale, logistics operations and enterprise resource planning (ERP) systems, to help make sense of the total picture.” This data comes from both structured and unstructured sources and optimization solutions must be capable of dealing with both. Only then can you really obtain “the total picture.” Frase told Bowman, “By pooling that data outside of everybody’s silos, you can get to functions that weren’t available before. You’re allowing the system to extract approved forms of data so that other people can see it.”

 

Frase is right to emphasize the importance of breaking down silos and sharing information. For more about the reasons why, see my post entitled The Curse of Silo Thinking. The reason that you want to share data is so that corporate alignment can be more easily obtained. When different departments back their arguments using different data, confusion and conflict are more likely to result. Supply Chain analyst Lora Cecere insists that data should be written once and read many times. That is exactly what the solutions described above are aimed at doing. Frase next talks to Bowman about another of my favorite topics “what if” scenarios. Bowman writes:

“Central to this new concept of inventory optimization is the use of ‘what-if’ scenarios to review all possible options, prior to putting the best one into play. Another IBM Research initiative, this one out of its offices in Zurich, employs this technique. Called AXIO, it draws on years of historical sales data in order to determine optimum stocking levels in a variety of situations. IBM is using the tool in-house, as well as deploying it on behalf of customers such as BMW and several European ‘do-it-yourself’ retailers.”

The only quibble I have with Frase is that it is literally impossible “to review all possible options.” The reason is simple; life is too complicated to determine “all possible” scenarios that could emerge. The past is not always prelude to the future. Utilizing “what if” scenarios, however, can train decision makers to respond better and quicker even to unforeseen circumstances. The use of historical sales data is part of what I was referring to above when I discussed the fact that technology can help optimize on-shelf availability. Enterra Solutions offers its customers solutions similar to IBM’s. Like IBM’s approach, Enterra’s reaches “down to the execution level, where the realities of everyday business can play havoc with a carefully thought-out operational plan.” Ulrich Schimpel, manager of inventory analytics with IBM Research in Zurich told Bowman, “The normal theory that you see in textbooks about supply-chain optimization … is not sufficient. You have to take into account a lot of restrictions.” Another IBM analyst, Eleni Pratsini, manager of IBM Research’s Mathematics and Computational Sciences Department, told Bowman that “the gap between theory and practice is especially wide in spare parts fulfillment.” Bowman explains:

“Demand in that area can be highly erratic. … In the end, [Pratsini] says, what really matters is service level, but to maximize that metric companies might need to depart from traditional assumptions in economic order quantity (EOQ) theory. Under a contract with IBM Global Services in Germany, automaker BMW is using AXIO to support its IT platform for spare parts delivery. Known as ATLAS, for Advanced Parts Logistics in After Sales, the system operates out of a central warehouse in Dingolfing, Germany. It manages more than 270,000 parts and accessories from some 1,900 suppliers. A pilot version of AXIO reduced BMW’s U.S. parts stocks by 10 percent, according to IBM.”

Schimpel told Bowman that “‘the big idea’ behind AXIO is the notion that ‘the world is getting more and more connected.'” That “big idea” is certainly not a new revelation. There are all sorts of discussions being held about the era of “big data” and how it can be used to greatest effect. Bowman notes, “Companies don’t lack for data and the necessary sensors to track supply-chain performance.” He continues:

“They can devise any number of forecasts using complex algorithms and simulations. What they must be able to do is filter that massive amount of information into a coherent picture that conforms to reality. Outside influences that are beyond the control of supply-chain managers, such the buzz of social networks and shifts in weather patterns, must be married with hard data to create an accurate picture.”

One of the reasons that Enterra Solutions uses an ontology is because an ontology can help establish relationships that might otherwise be overlooked. When trying to make sense of massive amounts of data, a system that senses, thinks, learns, and acts is absolutely critical for staying ahead of the game. Analysts predict that even greater amounts of data are going to have to be analyzed in the future. The spreadsheet era is quickly ending. Only cutting edge technologies are going to be capable of meeting the challenges that lie ahead; that is why Bowman recommends that companies redefine what is meant by supply chain optimization.