Philip Ribbins, founder of Orchestr8, has a passion for improving supply chains and for helping companies improve the ROI (return on investment) from their supply chains. In a recent interview, he expressed some of his concerns.
Ribbins began by saying, “A Gartner analyst told me not long ago that machine learning was now being applied to the analysis and maintenance of base data” that drives inventory and supply chain activity.
“Base data is base data,” he said; Ribbins’ voice clearly indicating that his concern was more like alarm. “Inventory and supply chain managers need to really know and understand their base data. They need to own it and take responsibility for it.”
Our conversation wandered back to a time when mainframe computers were just coming of age—used primarily by the military, large government agencies, and giant enterprises. We talked about man’s sometimes-misguided faith in what computers can do.
Sometime in the early 1980’s (if I recall correctly), some high-ranking bureaucrat at the U.S. Department of Commerce’s National Weather Service made an audacious statement something along these lines: “Give us a computer powerful enough, and we will be able to predict the weather with 100 percent accuracy.”
Today, millions of people each carry around with them in their smartphones, tablets and laptops more computing power than existed in the entire world on the day that statement was made. Yet, despite the all the computing power available, weather forecasting is still not 100 percent accurate.
Is it improved? Surely. Do we know and understand more about the factors that affect our weather? Indeed, we do.
However, what meteorologists have discovered, as computing power has grown exponentially over the years, is that we don’t even know or understand all of the factors that drive weather systems. Moreover, since we cannot quantify—or even know—all of the factors, our computing power still fails to deliver on accuracy.
Ribbins made clear that “attempts to cover over” the core failure to comprehend “the basics of supply chain management” with new, and increasingly complex technologies—whether machine learning, or more and more complex forecasting algorithms, or so-called “big data”—has already demonstrated its nearly complete inability to deliver effectively on the ROI promises that accompany it.
“The promises made” by technology vendors selling their “fancy tech” wares remain largely unfulfilled, Ribbins observes. “Machine learning adds complexity” without delivering significantly “improved outcomes.”
What supply chain managers and executives “really need to know, understand and manage are the crucial effects of lead time and batch sizes” across their supply chains, declares Ribbins. Without a thorough comprehension of the effects of these two basic factors, supply chain managers can never understand what is driving the actual inventory levels.
One of the reasons supply chain managers relying on traditional MRP (material requirements planning), MRP II, ERP (enterprise resource planning), and sophisticated forecasting software cannot know what is driving actual inventory levels across their supply chains is that there are simply too many factors employed by these systems. Almost nobody can keep them straight in their minds and thoroughly comprehend the interactions these factors have on one another through the systems multilevel and recursive calculations.
Ribbins summarized the current state of most supply chain managers’ and planners’ inability to take effective control of their supply chains and manage toward a process of ongoing improvement in the following three point:
Demand-driven approaches to supply chain management—such as array of tools implemented in Orchestr8—provide supply chain managers and planners with a return to simplicity. The impacts of lead-times and batch sizing can be seen and clearly understood when users are introduced to the underlying demand-driven concepts. Supply chain segmentation evolves organically as differences in lead-time and variability are recognized and implemented in a conscientiously applied program.
Sales and operations planning (S&OP) begins to become thoroughly and rationally integrated into a whole system of ongoing improvement augmented by dashboards and feedback loops implemented in Orchestr8. The common practices of finger pointing and blaming swiftly dissolve in the new environment and, almost without effort, everyone from the production floor to the C-suite can suddenly agree on correct courses of action and priorities.
In coming articles, we will talk more about how reducing complexity and returning to a full knowledge of the basics can deliver positive affects on supply chain performance and lead to dramatically improved ROI. In the subsequent articles, we will talk about how to begin dissolving the inherent disconnects typically found between your S&OP meetings and actual supply chain execution. We will also cover how to integrate forecasts while remaining truly demand-driven.
Watch for the upcoming articles!