The Importance of a Stable Forecast in Demand Planning

In the world of supply chain planning, achieving a balance between forecast accuracy and operational stability is crucial. While many organizations strive for highly accurate, frequently updated forecasts, using these forecasts within the operational horizon inadvertently leads to significant operational challenges.

Constant adjustments to forecasts create a volatile operational environment, resulting in inefficiencies such as frequent changes in production schedules, inventory levels, and procurement orders. Maintaining a stable forecast in this range far outweighs the need to have the latest 'precisely right' demand signal, contributing to a more resilient and efficient supply chain.

Balancing Accuracy and Stability

Forecast accuracy is undeniably important; however, excessive focus on short-term adjustments can be detrimental. When forecasts are continuously updated to reflect minor demand changes, it can lead to a "bullwhip effect" where small variations in demand cause large fluctuations in orders and inventory levels. This instability can strain resources, increase costs, and reduce overall supply chain efficiency.

Forecasting in DDMRP

In the DDMRP model, which is primarily about operational and tactical planning, we ensure that the forecast is stable by using an average of the forecast and historical sales, known as a Blended Average Daily Usage (ADU). For highly seasonal items, we weight this average to use more of the forecast over a shorter horizon. We use machine learning to improve the forecast where promotions or weather patterns are important, but it becomes part of the average.

This method works because we measure demand volatility and set a buffer (red zone) to absorb changes in actual demand. When actual demand exceeds the average expected demand, known as a spike, we react immediately to replenish and start prioritizing this big order. These buffers adjust over time as demand changes; however, if the buffer breaks, resulting in a stock-out, the planner receives instant feedback and can adjust the buffer size accordingly.

The Result of Using a Stable Forecast

Based on the feedback we have received from our +180 clients, this approach results in a much more stable plan with lower inventories, higher service levels, fewer changes over time, and fewer expedites, all of which enhance operational efficiency.

Operational vs. Strategic Forecasting

While the average method works operationally, strategic planning requires working with the actual forecast. The impact of demand changes (as supply shortages or operational issues) needs to be rapidly assessed, and plans must be put in place to meet these forecasted sales or understand the scale of the potential lost revenue. Scenarios such as planning more shifts, organizing extra space, and increasing prices to slow demand need to be assessed, and decisions need to be made.

Adjustments should be made at aggregated levels to speed up decision-making. Working with one forecast and collaborating effectively with different departments is critical. Hence, forecast entry needs to be in different units of measure for the different stakeholders: operations in units, sales in dollars, and warehouse in volume.

In addition, forecast adjustments should be made only when a significant impact on the feasibility of supply plans is identified. For instance, adjusting for causal effects such as promotions or market trends is less impactful if the forecast is already very volatile.

Measurement of Improvements

Traditionally, forecast measures focus on accuracy, but forecast stability and bias are just as important. Forecast stability, accuracy, and bias improvements can be effectively measured by comparing statistical forecasts to final enriched forecasts. This helps identify areas for improvement and ensures that the forecasting process continuously evolves.

Conclusion

While accurate forecasting remains a goal, stability in forecasts is equally important to maintain operational efficiency and reduce unnecessary fluctuations. By focusing on stable forecasts in the operational range, organizations can achieve a more resilient supply chain, better resource utilization, and improved service levels and profitability. This balanced approach to forecasting ensures that while the organization remains responsive to market changes, it does so without compromising its operational integrity.

b2wise

Demand Driven Institute

Institute of Business Forecasting & Planning

#forecasting