Step 3: Forecasting & Scenario testing

Model picture was taken from “Responding to the Lehman Wave: Sales Forecasting and Supply Management during the Credit Crisis”, by Robert Peels et al, published as BETA Working Paper Series, nr 297.

The model is used with the end market demand that in general is published by statistical bureaus or branch organizations. These data are often referred to as key economic indicators. Where relevant and known other factors that have a relevant influence on the demand, such as Active de-stocking, will be used. 

When the model is ready it will allow to distinguish between the various drivers that determine demand for an upstream company, such as: 

  • * End market demand, which determines maybe 80% of demand.
  • * Stock building and reduction in the chain, which can be 5% or much more if the market is volatile.
  • * Events, such as the Lehman crisis, a flooding, a war or a disaster.
  • * Segment shifts, when a product range gains share in the end market, e.g. because customers switch to a more sustainable end product.
  • * Product penetration, which is similar to segment shift, but more directed to a component in a product. E.g. a new type of chip in laptops.
  • * Market share, which is dedicated to a single company.* Import & export, outside the region in scope. 

 

The resulting demand curve for the past is compared with actual sales data and the model is tweaked if necessary. The results and their strategic impact are discussed.  The output is: understanding your recent market share growth and gaining supply chain visibility.

For many end markets a forecast is available in the open literature and will be obtained. This kind of macro forecasts of economic indicators in general is much more reliable than a forecast by e.g. surveying your direct customers. This end market forecast will be combined with the own vision of the team on the future growth of the end markets  and with your vision on your future market share development, e.g. based on innovation and/or new product introductions.  By entering these two factors into the model we will develop a number of demand scenarios for the future. The output is: one or more future demand curves.