MENU
codes

How to provide a forecast accuracy measure that represents a non-weighted average of products?

October 2, 2013 • Featured, How to, Modeling

Our Supply Chain Manager has requested that we provide a forecast accuracy measure that represents a non-weighted average of our products, that is, Mean Absolute Percent Error (MAPE), not weighted Mean AbsoluteĀ Percent Error (wMAPE). He says that our customers score us in the same manner. They do not weight each SKU based on volume/Sales; we either fill the order or we donā€™t, regardless of quantity.

In this case, your customers need to be educated. Do they need the same fill rate for each SKU, which may be suboptimal, or a higher fill rate for high-volume (value) SKUs and somewhat lower for others? Bear in mind, we have limited amount of forecasting resources and, thus, cannot afford to pay the same amount of attention to each SKU. MAPE is fine if the objective is improvement in the fill rate across the board. If high-volume SKUs matter most, then WMAPE will be the best metric because it highlights how high-volume SKUs are doing. If WMAPE is high, it may very well mean that high-volume SKUs are not doing well.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

« »