The Lean manufacturing, PPT Plus,
"Back-to-Basics" e-Training Package 
Lean Manufacturing Solutions

Increase the effectiveness of your
Lean Manufacturing Training Program
Lean Manufacturing Simulation Game 

Never Rely on Demand History Alone (Cont.)

My examples probably won't be appropriate for your busi­ness, but they do illustrate the fact that there are many ways to confirm whether an apparent trend is true or false. You have nothing to lose and everything to gain by finding the answers that will help you make an intelligent forecast.

But what about cases where you have no demand history, such as new part numbers? This can be very difficult because demand and pipeline fills are often based not on actual need, but by customer perception. Nobody needed a Pet Rock or a Mood Ring, but they sold like crazy. Pipeline fills are generally based on the distributor's perception that they have a market for that product.
In some cases, forecasts may be linked reliably to other known factors, such as a tool company basing volume on the number of sales/service trucks, or replacement parts for computers based on the number of each model sold.

But you must be extremely cautious of forecasting in this fashion, driven by numbers alone. Be very sure that you've asked a lot of questions and that the true demand drivers are understood. Also acknowledge that while the forecast quantity may be helped by these methods, the timing may not. Do not forget to find out when these new items will first appear in a catalog, announcement, or other advertis­ing medium.

In another example from my industry to illustrate why the numbers alone don't always tell the whole tale, you might think that a good method for us to use would be to multiply registrations by a replacement rate. While that approach sounds viable, and may be a good start, I have never personally seen that method alone produce consistently good forecasts.

We know the number of United States registrations, but we have foreign customers as well, and those registration numbers are not know. Aggregate replacement rates may be very accurate, but there are tremendous differences by item due to factors like original equipment problems, vehicle weight, vehicle type (Four Wheel Drive Trucks vs. Sports Cars), and wear rates due to geometric design differences. Unfortunately, you won't know the individual item replacement rate until after it starts selling, when it's too late, and that rate will change as the vehicle ages.

While the examples from my industry will not likely help you out with your product offering, they do illustrate the point that there is much more to the process than meets the eye. Once again, the numbers alone fail us. A thorough understanding of what drives customer demand for your products will help you determine if a numbers alone approach is right for you.
However, the same principles that were discussed for forecasting product lines can be used for individual items as well. It is very helpful to review the demand of the most similar items you already have in the line, assess the differences, and make the most likely adjustments. The key is to fully understand what drives the customer de­mand for the product.

In Closing

I hope that everyone may gain from the understanding that demand history alone is not enough to consistently produce accurate forecasts. This is why many statistical models and systems fail to achieve the hoped-for results. Forecast success in the real world lies less with pure statistics and more with intelligent use of Sales and Marketing input and understanding your business and customers. Understand­ing the demand drivers and using them in your forecasts will certainly help prevent mistakes and help you realize true accuracy improvements. My best wishes to you and your success.