Whether your Plastics Manufacturing Business is volatile or stable, continued growth in the top and bottom lines is expected. Presumably, everyone knows the myriad of reasons and impact that accurate demand planning has on carrying costs attributed to overstocked inventory, Return on Capital, Cash Flow, and Earnings. If underestimated, it can obviously adversely affect revenues and customer satisfaction.
Demand forecasting requires:
- Collaboration with at least:
- Continuous refinement as
- Customer requirements change
- Economy changes including Micro and Macro factors and resin prices
- Product mix refinements and new launches
Art is the collaboration aspects necessary to help from accountability and to refine. Science is determining baseline forecasts (“Normal” volumes) and numerically accounting for spikes in demand.
It is the marrying of these two that results in a final Demand Forecast and the ensuing continuous refinement. Note that there is great variance in how often the forecast is adjusted and the timeframe used, but forecasting beyond three months seems to lose value with exceptions for longer running production orders.
The example below also includes references to Machine Learning models (see reference to Auto-regressive Integrated Moving Average (ARIMA), Auto-regressive Tree (ARTXP) or Mixed (Best Fit Model). In another blog, we may discuss the impact Machine Learning might have in product pricing to help guide clients through price changes due in elastic markets or when resin price increases will dictate eating margin.
Ellipse Solutions has documented the process of creating a Demand Forecast in our Introduction to Demand Forecasting in Microsoft Dynamics AX 3 part series!
For more on how Machine Learning and Demand Planning from Microsoft, see: https://docs.microsoft.com/en-us/dynamics365/operations/supply-chain/master-planning/introduction-demand-forecasting
One final comment… All models are wrong… but, some are less wrong.