A Remedy for Your Forecast Headache!

by Dave Garwood

Looking for relief from that constant, pounding forecast headache? Read on!

Forecasting product sales is never easy --- and never accurate! One of our previous articles (see the archives) discussed how to shift the focus from forecast accuracy to minimizing demand variation. This article suggests some ways to avoid the need to forecast altogether!

Here are a few suggestions:

1. Establish a forecasting policy that identifies the criteria to determine which items will and won't be forecasted. For example, if the frequency of demand is less than 2 per year and the total annual volume is less than 10 per year, classify those products, options or service parts as "unforecastable." Stop forecasting them. Make them to order.

2. Forecasts are needed only when it takes longer to buy material and produce the product than the customer is willing to wait for delivery. If the lead time to buy all the material, make components and make the product can be reduced to less than the competitive customer delivery time, the product can be made-to-order, not to forecast. At first, this may sound like an overly simplistic solution; fact is, many companies have been successful at reducing lead times by 50 percent or more. Lead time reductions of this magnitude may very well make it possible to make fewer items to forecast. Usually most of the total lead time (frequently called time fence), often 70-80%, is purchase lead time. This is the key area to focus on lead time reduction to eliminate forecasting.

3. Many products are made by custom assembling or finishing from components. PCs are a good example. The forecasting challenge for these products is predicting which options will be needed, thus which components must be available to custom finish the product. For example, a different bracket may be required for a horizontal rather than vertical mounting. If 20 different mounting arrangements are offered, 20 different brackets are required. A product mix forecast of the 20 options is required. If the bracket can be redesigned by adding slots instead of holes, an extra flange or standardizing on one material, one bracket may be used for all mountings. This eliminates the need for a product mix forecast. Redesign the product to make more parts common to all options, eliminating the need for forecasting options.

4. Change production from make-to-forecast to finish-to-sales order. Many times, a wide variety of items can be manufactured from a single "basic" or semi-finished item. If the deliverable product can be made after receiving the customer order, the need for forecasting a wide variety of deliverable products has been eliminated. Changing the manufacturing process, redesigning the product, or both, may be required to make this solution work.

5. Today, many, companies are using an ERP, MRP II or whatever they call it, planning system. Inherent in these systems is planned orders for every item that they intend to buy. Take advantage of this information. Customer's planned orders are the supplier's future demand. This is probably the best forecast the supplier will ever get. Today, more than ever, customers are recognizing their responsibility for effective supply chain performance. The Internet has made communications of this data very feasible and economical. Work with the planning people in your customer's plants to get this information, eliminating the need for the supplier to forecast future requirements.

These alternatives won't work in all cases in every company. But they do work in many cases. Don't dismiss them too quickly. Implementation usually requires action from several departments. Solving the forecast accuracy problem can't be easily shuttled off to a single department. Product engineers, manufacturing engineers, sales/marketing, production planning and purchasing working together can make giant strides toward eliminating the need to forecast. The rewards are huge -- better response to meeting customer needs while reducing inventory and costs.

Look at the items you are forecasting. Ask what would have to be done to eliminate the need to forecast them. Use the five suggestions above as a guide. Your forecast headache just might disappear!