Supply chain leaders continue to rank demand variability as one of the top challenges to achieving their goals and objectives. For five years running, respondents to the “Gartner Supply Chain User Wants and Needs Study” have listed it as the top challenge in efficiently and effectively managing supply chains. For a discussion of this, see Dwight Klappich’s article in the current issue of Supply Chain Quarterly.
So what is variability? Variability is basically the difference between what we expect from something and what actually happens. It is the statistical distribution of outcomes one can expect from a process. It has become axiomatic in recent years that this statistical distribution is widening, meaning supply chain managers have to anticipate, plan for and react to a widening array of demand and supply scenarios.
Variability in all forms is the enemy of supply chain efficiency. If there was no variability in consumers’ tastes, in product portfolios, in manufacturing, transportation, and distribution lead times, or in any of the other of the processes for buying materials, producing products and distributing them to consumers, then supply chains would run like clockwork with very little need for sophisticated management. Obviously, that is not reality; in fact, variability increases every year as supply chain structures continuously morph in a competitive environment based on customer-centricity, responsiveness and efficiency.
Thus, the role of supply chain management is to reduce variability while at the same time putting in place synchronized and responsive processes for managing variability. Companies that do this effectively are able to gain significant advantage over their competitors.
Variability vs. Volatility?
Variability and volatility are terms that are often used interchangeably. My mentor, Ken Sharma, who was a pioneer in the field of supply chain management, spent a good deal of his professional life advising companies on how to manage demand variability and volatility. Back in the 1990s, he described variability as a change in demand outside of order lead time and demand volatility as a change in demand inside of order lead time. He took much of his foundational thinking from the seminal work of Jay Forrester, first published in 1961. He made a distinction between variability and volatility to explain how one must plan for variability through response buffers and capacity and at the same time respond to volatility through synchronized business processes and tradeoff analysis. In other words, plan for variability across a planning horizon and respond to volatility as you execute the plan. For this discussion, I will use Ken’s definitions, which draw a clear distinction between the terms variability and volatility.
As indicated in the Gartner surveys, much of the attention is on demand variability and volatility, since that is the starting point for all subsequent upstream activities and has the ability to fan out and drive variability into these upstream activities through manifestation of the classic bullwhip effect. For more on the resurgence of the bullwhip effect, read my guest blog post in Jay and Barry’s OM blog.
Following are some of the ways that supply chain leaders are managing demand variability and volatility for competitive advantage. In general, demand variability is effectively managed through robust and synchronized planning processes, and demand volatility is effectively managed through robust and synchronized response capabilities.
- Drive the supply chain from real demand. In any discussion of how best to manage demand variability and volatility, it is important that we define and understand real demand. (For more on this topic, see my earlier blog post). Lora Cecere of Supply Chain Shaman, has been advising clients for years on how to become demand-driven. Most companies describe themselves as demand-driven, but many still don’t truly understand what this means. Leading companies today are using advanced analytics in collaboration with their downstream partners to understand real demand and then drive operations from this demand signal. For more on how companies are doing this, see the webinar “Building a Shelf-Connected Supply Chain that Even the Weather Can’t Beat.”
- Consider different demand outcomes using probability and scenario analyses as part of planning processes, starting with the S&OP process.
- Incorporate automated segmentation and classification capabilities along with dynamic demand response capabilities into the forecasting process. For more on this, see the webinar “Beyond ‘Best Pick’ Forecasting for Supply Chains.”
- Continuously evaluate response buffer strategies to ensure they are aligned with customer segmentation and associated demand variability. For example, in configurable product industries, forecasts can be significantly improved through postponement strategies in which end item forecasting is replaced with pooled component forecasting. For an example of this, see my Supply Chain Quarterly article on segmentation from earlier this year.
- Provide end-to-end and linked demand-supply visibility to deal with demand volatility as plans are executed. This visibility allows you to see deviations from the demand plan as they occur.
- Synchronize demand management processes with supply management processes. This provides the foundation for tradeoff analyses as demand fluctuates inside of lead times. For example, diversion (shifting to an alternate, or diverting supply from one customer to another) and time compression (alternate transportation mode) are supply capabilities that must be quickly considered in the face of unexpected demand changes. Likewise, on the front-end of the chain, it’s important to integrate allocation management with order promising to perform tradeoff analyses between channels, business units, and customers.
- Use the synchronization described above to employ demand shaping strategies. This includes price, lead time, and product content tradeoffs to drive customer behavior to support available supply.
- Incorporate a learning framework to gain insights into variability and volatility. For example, simple forecast waterfall analysis as part of the S&OP process can provide useful insights into how much the supply chain is getting whipsawed by week-to-week and month-to-month changes.
- Close the loop between variability management and volatility management. This means a robust planning process integrated with a robust response capability. This creates a closed-loop control process in which the plan is the objective function and the response capability provides the means to achieve the objective function in the face of demand volatility. This closed-loop approach is critical to achieving supply chain velocity, which is a topic I will discuss in a future blog.
- Continuously improve processes to reduce lead times and process variability. Lead time reduction in all processes provides the means to deal with increasing levels of variability and volatility. (If lead times were zero, variability and volatility would be meaningless since you could instantaneously respond to everything). Likewise, it is important to continuously improve supply chain process capabilities to reduce variability so that the domain of outcomes that need to be planned is lower, thus reducing all sorts of buffering including inventory and capacity.
We are seeing leaders adopt many of these strategies. Where do you stand in this journey?