Why You Should Embrace Uncertainty in Demand Forecasting – Part 2
Last week in our first post about uncertainty in demand planning, we said that in reality most companies are already part way down the path towards a probabilistic supply chain model.
The evolution is happening naturally as the global business landscape becomes more interconnected and complex, along with the effects of multichannel marketing, demand shaping and the internet. You’ve likely already recognized the limits of deterministic planning if your company has recently launched (or relaunched) an S&OP initiative, or you’ve started to migrate from traditional platforms like SAP APO to advanced demand analytics tools, or trained your planners on supply chain performance trade-off competencies.
Here is one straightforward opportunity to focus on in the near future: Migrate away from top-down demand forecasting.
Despite the added complexities in today’s supply chains, traditional SCP systems like SAP APO typically apply the traditional “top-down” approach to forecasting based on aggregated data. This approach aggregates demand to smooth out variability, which makes it easier to generate a high-level forecast, but the Item-Location level forecast quality is poor because demand signal details are dismissed along with the “noise”. So this approach of aggregated planning and then applying slicing and dicing rules, only works for simple and highly predictable businesses with few fast-moving commodity items and single-channel distribution. When it comes to long-tail items, forecasting metrics such as WMAPE become almost meaningless or even misleading. When dealing with intermittent demand, they don’t do anything to measure the uncertainty inherent in lumpy demand.
Uncertainty is best managed using a new breed of planning tools that employ adaptive probabilistic algorithms. They deal with the volatility and demanding response times, particularly as required by online and multichannel markets. These tools use adaptive modeling techniques that allow you to manage the supply chain statistically and with a high level of automation. Tinkering is no longer needed. Planners are called on to intercede only for exceptions that fall outside the boundary limits of statistical uncertainty.
Supply Chain Insights founder Lora Cecere has repeatedly advised her clients and community members to consider these newer, “best-of-breed” tools to mitigate business risk and future-proof their supply chains. Her blog post, Three Reasons Why SAP Supply Chain Planning Is a Risk to Your Business is just one example. Gartner and Nucleus Research have made similar recommendations.
An added benefit
An additional benefit from this approach is improved planner morale. Current approaches to forecasting not only “hit a ceiling” of diminishing returns, they also “hit a wall” due to an inability to cope with increased business complexity. Deterministic top-down systems may have worked in simpler times, but as companies grew, added product lines, acquired other business or went multichannel, those systems required unprecedented amounts of effort. Overwhelmed planners and working weekends ensued.
Planning systems that embrace uncertainty inevitably change the working environment for planners. Because the system understands and is able to handle much of the inherent uncertainty, planners are free from dealing with hundreds of mini-crises that beset a deterministic approach. Instead they can focus on adding business value to the S&OP process and dealing with those few truly unusual events.
When planners are taught to embrace uncertainty using a probabilistic approach, they feel much more in control. They know that over time the known variables provide a level of certainty, and for the rest they can devise controllable contingencies. It eliminates the spirit-crushing defeat of never being able to reach goals. As an automotive industry client of ours said recently: “Nothing affects team morale more than our ability to meet service requirements.”
Happy planners offer a second benefit for the company. In most regions, there is a distinct shortage of talented demand and supply chain planners. So employing a methodology that improves planner productivity reduces the need for additional planners, and an improved work life increases employee retention.
