Forecasting Methods: Product Segmentation Matrix (E)
by Laurus Nobilis
 

Posted: Nov 2009

Previous: Segmentation Of Products

 
The SKUs have different share in sales volume, therefore they participate in overall business in a different way. The forecast errors on high volumes can cause greater damage on business than slow mover SKU. If the stock out is caused by low forecast accuracy on fast mover it makes a huge impact on sales volume and profitability. In case that low forecast accuracy is causing overstocking it holds too much working capital on inventory and causing extra warehousing costs. 

The historical low accuracy on certain SKU is indication that attention should be paid to that product. If the historically inaccurate SKU is large  contributor to the volume, the priority for more active forecasting is even higher. In other words, combination of size and accuracy matters when it comes to segmented approach to demand forecasting. This approach enables demand planner to have more strategic approach to demand forecasting. How the segmentation looks like?

The products should be assessed on 6-12 months basis, by taking into account sales volume and average forecast error. The combination of these two parameters are placing the product on segmentation matrix.

The segmentation matrix is grouping the products to one of four segments. The other name for segmentation matrix is the Zoo Chart, due to similarities of behavior of products with the characteristics of certain animals. :
 

Turtles are SKUs with low volumes and low forecasting error. Their impact to business is not significant to require focus, unless there is some kind of marketing promotion ongoing. The products should be forecasted with statistical methods, without active forecasting.

Jack Rabbits are SKUs with small volume, but with high forecasting error. These SKUs should be mostly planned statistically. The focus should be only on SKUs under the promotion. For the other it is required only to run statistical analysis. As additional measure, the safety stock can be increased, in order to cover forecasting error.

Work Horses are SKUs with high volume and low forecasting error. In other words, these products are large in volume and reliable. They should be planned statistically.

Race Horses are SKUs with large volume and high error. These products are risk group for the business, since they are large in contribution to sales and profit, but they are unpredictable. This mean that they can cause losses due stock outs in case of increased sales versus forecast, or write offs in case of decreased sales. Race Horses should be forecasted with active approach.

Active forecasting should be applied to selected categories of significant impact to the business. Active approach involves combination of statistical approach and involvement of stakeholders into demand forecasting. The stakeholders are the sales managers and marketing managers. Their input should help demand planner to incorporate their opinion into statistical forecast.
 

The product segmentation matrix ( zoo chart ) can be useful approach to dement planner. The zoo chart gives insight into big picture of basic performance characteristics of SKUs; their sales contribution and forecast accuracy. As the result of increased forecasting accuracy the whole supply chain will benefit. The inventories will be within normal range, the stock outs will reduced, warehouse and distribution will be optimized and the service level toward customer will be increased.

Recommended Reading:

Demand Forecasting 
Forecasting Accuracy