Home » Big Data » Supply Chain Analytics

Supply Chain Analytics

February 16, 2011

supplu-chain

In a post entitled Trends in Supply Chain Software, I cited a blog post by Lora Cecere in which she identified a number of trends she is going to watch this year [“Ring in the New Year …” Supply Chain Shaman, 22 January 2011]. One of those trends was “New Forms of Supply Chain Intelligence.” I was pleased Lora that mentioned Enterra Solutions® in relation to that trend. This is what she wrote:

“One of my favorite trends that I am following is the changing world of supply chain intelligence. New capabilities are evolving that can deepen our abilities to design, sense and respond value networks. Historically, supply chain intelligence has been limited to simple work rules, optimization and simulation. The use of Software as a Service (SaaS) changes the game for optimization. The use of SaaS enables parallel processing, expert tuning, data cleansing, and benchmarking opening up new horizons. I am also following the use of artificial intelligence to map ‘multiple ifs’ to ‘multiple then’ conditions, advanced pattern recognition to drive listening and sensing platforms and the use of optimization in combination with simulation to model the feasibility of solutions to drive the best answer. Technologies to watch include Enterra Solutions, Predictix, Revionics, SAS, Sockeye Solutions, Solvoyo, and Terra Technology. This is a great trend to watch in 2011.”

Lora isn’t the only analyst watching the business intelligence sector. Jeremy Friedler from the Maxiom Group suspects that supply chain analytics is going to be “the next big thing” in business intelligence [“Is Supply Chain Analytics The Next Big Thing In Business Intelligence?,” Maxiom Group Blog, 18 January 2011]. Friedler writes:

“I want to focus on [an] emerging supply chain trend that seems to be gaining significant traction by allowing companies to increase operational responsiveness and flexibility. This area, known as supply chain analytics, hopes to allow for supply chains to quantitatively segment multiple data points to make fact based predictive decisions as a means [to]become ‘data driven’ and improve operational performance. Supply chain analytics, a somewhat nascent yet growing business intelligence platform, led by companies such as Teradata, is proving to be a powerful force. While most companies across industry verticals measure supply chain performance based on what has occurred solely in the past through metrics, this approach will no longer be sustainable in the future. By contrast, analytics ties future and predictive performance together by analyzing trends based on millions of data points gathered from operational transactions in real time throughout the globe.”

As business analytic systems get “smarter,” they approach something like the sense-think-act paradigm that prompts human behavior. In fact, “sense-think-act” supply chains and “demand-driven” supply chains are probably the same thing identified by different names. In another post, Lora describes some characteristics shared by companies that approach having a demand-driven supply chain [“What happened to the concept of demand driven?” Supply Chain Shaman, 12 January 2011]. She writes that “demand-driven companies have by definition, five characteristics”:

  • Demand sensing: The early definition of demand-driven value networks was a network of trading partners that can sense and respond to demand fluctuations with near zero latency. The concept of demand sensing with near zero demand latency is the first step in becoming demand driven. A good example of this is how supply chains can be transformed by moving from a push to a pull-based signal using Wal-Mart Retail Link.
  • Active demand shaping. However, as companies mature and learn demand-driven concepts, they quickly learn that it is not sufficient to just sense demand, but that it is critical to also actively shape demand through demand orchestration. The process of demand orchestration is horizontal. It spans functional silos of sell, deliver, make and source to and enable market-to-market corrections (from the outside in) to stimulate, or deaccelerate demand. There are multiple demand shaping levers – new product launch, marketing programs, sales/channel incentives, price management, trade promotion management, service programs/after-market support, and the sale of Slow and Obsolete products (commonly referred to as SLOB) at the highest margin – that can be leveraged to speed-up or slow down demand. The key of being demand-driven is to actively shape demand to maximize profitability ensuring alignment of all functions to deliver against the opportunity. All too often, in supply-driven organizations, demand is shaped in isolation. (E.g. a promoted product that is not in stock or a price decrease for a product when a commodity market is tightening.) Samsung does this well.
  • Design of value networks for demand. A characteristic of demand-driven value networks is that they are designed, not inherited. The company also recognizes that they have multiple value networks with each having unique characteristics for response time, cycle times, and flows. The design of these value networks for demand-driven leaders recognizes the characteristics of demand flows and incorporates demand variability into the overall design. The goal of the supply chain shifts with demand variability. Networks with low demand variability (e.g. low MAPE) can be push-based supply chains based on efficiency goals of lowest cost per unit. Networks with high variability and high volumes (seasonal products) need to be designed to be responsive. These value networks sacrifice the lowest cost to have the most responsive value network. Networks with high variability and low volumes need to be designed for agility. In these networks, companies design for the same quality, cost and customer service given the level of demand fluctuations. They know what is possible and become active modelers to understand the trade-offs. Intel’s supply chain mastery program allowed the company to redesign the value network quickly to launch the Atom chip.
  • Agility through demand translation. In the management of demand networks, demand is orchestrated—sensed and actively shaped to maximize margin—through the use of an agile supply network. In this agile network, manufacturing and distribution flexs to maximize value. (E.g. Manufacturing load is shifted from plant to plant, and shipping modes are changed to maximize value in the value network.) This orchestration of demand—sensed, shaped and translated to maximize profitability – is usually coordinated through a series of horizontal processes designed from the outside-in (from market facing processes to supplier processes) of revenue management, Sales and Operations Planning (S&OP), and New Product Launch Commercialization.
  • Focus on outcomes. As companies implement demand-driven value networks, the focus shifts from selling into the channel to selling through the channel. It can also shift from a product-based focus to a value-based outcome focus that combines product and service (E.g. An example of value-based outcomes is the shift that is happening in healthcare. It is no longer sufficient to just sell insulin to diabetes’ patients, they want devices for monitoring and direct linkage to their physicians.) This may seem like a subtle shift; but the change management issues loom large for an organization implementing the strategy.”

Lora’s first two characteristics (i.e., demand sensing and active demand shaping) obviously focus on the “sense” portion of “sense, think, and act.” They also serve as part of the feedback loop. Since companies actively try to encourage customers to buy their products (through active demand shaping), the returning demand signal helps them measure how well their efforts work. The final characteristic (i.e., focus on outcomes) deals more with the “act” portion. Sensing demand is obviously meaningless if you can’t do anything about it. The middle and next to last characteristics (i.e., design of value networks for demand and agility through demand translation) provide the “thinking” connection between sensing and acting. Those two areas are where business analytics can fulfill their most vital roles. Friedler continues:

“The idea of analytics utilized within supply chain management becomes even more important as companies look towards cost containment and freeing up working capital, risk management, sustainability, and increased visibility within their supply chains. Therefore, automating the entire supply chain through advanced modeling and predictive performance is not only here for today, but very well may be the way of the future. The promise of supply chain analytics allows for ‘intelligent systems’ to learn and adapt to how a particular supply chain is run and then make adjustments accordingly through advanced simulation and complex algorithms. For example, this avant-garde system allows for predictive buying and selling patterns (i.e. what is the supply chain impact of a particular promotional event or seasonal trends) to coincide with the proper level of production, inventory, and distribution to more effectively align costs with revenues. In addition, some companies are even using supply chain analytics software as a competitive advantage, to adjust their supply chains based on market conditions, weather patterns, or customer segments. This ultimately provides the correct level of ‘end to end visibility’ of the performance of each specific node within the supply chain in order to improve responsiveness, contain costs, and improve customer service. Sophisticated modeling and analytics software may also allow companies to uncover data that may be overlooked within traditional ERP and CRM systems.”

One often reads about sense and respond systems — but little mention is made of the analysis or “thinking” that must take place between the sensing and the responding. This missing piece of the puzzle is what good analytic programs provide. Because analysts like Cecere and Friedler believe that business analytics will play an important role in the future, it should come as no surprise that workers skilled in such analytics are predicted to be in demand [“Analytics, Storage and Cloud Skills Will Be in Demand,” by Jennifer LeClaire, NewsFactor.com, 17 January 2011]. LeClaire writes:

“Experts say there is plenty of action across the board for technologists with specialized skill sets. IBM is betting on business analytics. Tim Powers, a spokesperson for Big Blue’s business-analytics division, said businesses and governments alike are grappling with the challenge of making sense of this data deluge to turn it into new opportunities, increased performance, and faster, better decision-making. Gartner reports business analytics is a top 10 priority for companies in 2011. ‘The power of business analytics is transforming this information into a strategic asset. Although having the best, most complete and up-to-date information is useless if you can’t make sense of it,’ Powers said. ‘Data unanalyzed is data wasted. Therefore, businesses and governments need two very important things to make this happen: The right technology, and employees with the right expertise and skill sets.'”

Clearly the need for supply chain analytics is on the rise. To remain competitive in a globalized world, costs need to be controlled and efficiencies need to be implemented. As I have pointed out in numerous posts, cost reductions and lean processes create tensions in supply chains that also require increased agility. Supply chain analytics can help companies discover the sweet spot that keeps a supply chain cost effective and efficient yet still maintains the agility it needs to succeed in a rapidly changing world.

Related Posts: