Every month our client forecasts sales of every product in each month for the next two years. The forecasts are based primarily on historical sales patterns and planned marketing activities.
The people involved in the process are the demand manager, the supply manager, and a team of product managers. The demand manager has overall responsibility for the entire process, and for the accuracy of the forecasts. The product managers are responsible for individual products in various markets, and in the context of this system their role is to produce sales forecasts for their products. The supply manager is responsible for ensuring that supply matches demand, without shortfall or excessive over-stocking.
At the start of each month the demand manager loads the latest sales report, inventory list and shipping schedule into the system. Then he clicks a button to prepare a new blank forecast. This blank forecast contains all products, along with their sales histories. The demand manager then notifies the product managers that the new forecast is ready for them. This all takes about ten minutes.
Next the product managers prepare their individual product forecasts. For each product the manager selects an appropriate statistical forecast algorithm from a library of pre-defined algorithms. For example, for one product she might choose a 3-month run rate, and for another she might forecast a 5% growth rate. It takes the system about ten minutes to generate forecasts for all products. Then the product manager reviews the figures, and optionally overrides individual monthly figures if she anticipates a variation from the statistical forecast.
Sometimes a product manager wants to apply a custom forecast algorithm to a particular product. To do this she exports the blank forecast to Excel, does her custom forecasting there, and then imports the result back into the system. This way the system usually does the heavy lifting of the bulk forecasting, but the manager always has the flexibility to do things differently if she wants to.
When all the product managers have completed their forecasting, the supply manager gets to work. He opens the forecast and clicks a button to run the supply algorithm, which calculates monthly ordering quantities. For each product, in each month, the algorithm takes into account the opening inventory, the product managers’ sales forecasts, historical demand patterns, the shipping schedule, and other factors, and calculates the optimal order quantity. This takes about twenty minutes.
The system highlights any numbers that fall outside configured tolerances, such as inventory insufficient to satisfy forecast demand. This can happen when demand depletes stock within a product’s lead time window and can’t be replaced, or when the supply manager manually overrides a system-generated figure.
At every step in the process the system records details of how all figures were calculated, so that everyone can see where the numbers came from. And all system-generated numbers can be overridden manually.
Finally, when the supply manager has produced the optimal order quantities, the demand manager exports the order data to a file, and loads the file into their ordering system.