Joe Shedlawski, CPIM | September/October 2012 | 22 | 5
Alleviate variability with more effective demand management
Whether in the supply chain, customer service, or financial performance functions, a root cause of most disruptions to business operations is some type of variability. To achieve sustained success—particularly in view of the frequently changing external environment—enlightened business leaders are taking sharp aim at controlling internal variables.
I was fortunate to see a chief executive officer of a large consumer goods supplier deliver a keynote address before a conference of academics and professionals. His topic dealt with succeeding in uncertain markets. His solution focused on taking control of variability through the proper implementation and use of a fundamental business tool: sales and operations planning (S&OP).
Uncertainty is ubiquitous. Much of it comes from external factors—regulations, exchange rates, the economy, the weather. Variability refers to the types of uncertainties a business can control through improved analysis, communication, and decision making. The variability of manufacturing output, for example, can be improved through proper operator training and machine maintenance. Process simplification, using lean and six sigma approaches, can reduce lead time variability. Standardizing communications practices can eliminate confusion and the resultant variability in interpretation.
I recommend converting from a forecast to a demand plan—and making sure to engage all key stakeholders in an assessment of capabilities, constraints, and customer information. These activities comprise the essence of the demand management portion of S&OP, and many organizations will show double-digit improvements in key business metrics as a result of being able to better control variability. Choosing the priorities
While any organization should set aggressive goals for variability reduction or elimination, the best opportunity for most will be found in the area of demand management. It used to be that demand management consisted of recording the statistical forecast—which was produced solely from historical data—and then comparing it against actual demand, usually to the dismay of those who were left holding the responsibility for the accuracy of the forecast. Natural responses to such a stressful situation include seeking out even more historical data or predicting even further out into the future—neither of which improve forecast quality.
It’s no wonder that the greatest amount of variability in most businesses today is found in the demand plan. Statistical forecasting is seen by forward-thinking professionals as only one of several inputs into a process of true demand management. One absolute truth about a forecast is that its accuracy deteriorates as the time horizon increases because there is more uncertainty in the distant future. The other truth is that detailed forecasts have higher error rates than aggregated forecasts. Recognizing these truths, and also that most improvement opportunities lie beyond the four walls of the organization, can set a priority plan for improvement.
Priority one should be to get everyone on the same page. Establish a shared understanding of lead times, capacities, risk tolerance, financial constraints, and flexibility to respond to change. This will help manage demand within tolerance whenever possible and stimulate fast action when conditions fall outside of normal limits.
Priority two requires an understanding of the difference between precision (how many decimal places in the measure) and accuracy (how well the prediction matches actual demand.) Data usually are overabundant, but those chosen to support improving the demand model should focus on improving accuracy to the levels of tolerance that were established in priority one.
The third concern is a matter of engaging all supply chain partners—from suppliers to customers—in the improvement effort, making sure that tolerance windows are understood and acted upon. The reasons for unplanned demand variation must be identified through root cause analysis, shared, and applied. Typically, this involves determining the leading indicators upon which a longer-range, aggregate forecast can be gauged, solidifying a demand plan at the volume level of detail, and refraining from a mix level of detail outside of the cumulative lead time. Certainly, the degree of uncertainty will rise as the time horizon lengthens; however, by following this process, an organization can effectively reduce the level of ambiguity at any point in time.
Joe Shedlawski, CPIM, is a consultant, coach, and educator with JFS Associates; an adjunct professor in the MBA program at Misericordia University; and an associate of the R.A. Stahl Company, specializing in sales and operations planning. He may be contacted at firstname.lastname@example.org.
Editor’s note: The editors of APICS magazine would like to thank Joe Shedlawski, CPIM, for his contribution to the “Sales and Operations Planning” department as a guest author.