Comments on: Stop Saying Forecasts Are Always Wrong

By: Klaus Spicher

Klaus Spicher — Wed, 07 Mar 2018 19:47:46 +0000

Thanks for your paper, stressing a future view on forecasting. I would like understanding the difference meanings of using “business impact” and “business effects”. (The dictionary is not helpful.) And how is the stated higher correlation between “error and business impact” vs “accuracy and business effect” identified? Analysis or opinion?
And how does a forecast range (and knowing the probabilities of outcomes) and probabilistic forecasting support buying the right quantities of materials – or other decisions based on demand forecasting? Thanks for answers!

By: Charles Chase

Charles Chase — Wed, 21 Feb 2018 01:39:19 +0000

Eric, great article.

Those same upper/lower forecast ranges should be used as part of the inventory safety stock calculations. If so, then inventory safety stock should be able to cover the variation in the forecast. That said, the more accurate (lower error) the forecast the lower safety stock is required. Their is a direct correlation between accurate forecasts with low errors and the reduction of inventory safety stock.

As you mentioned we focus too much on forecasting the actual number versus staying within the upper/lower forecast range. If the actuals fall anywhere within the upper/lower forecast range then the forecast is considerer very accurate.

By the way we never say that the Financial Plan, Supply Plan, Inventory Plan, or any plan for that matter is always wrong, and many times not as a result of a poor forecast. In many cases, very accurate forecasts are ignored because they don’t meet the needs or purpose of the departments who consume the forecast.

It’s time that someone call them out on this topic.

By: Charles Chase

Charles Chase — Tue, 20 Feb 2018 23:57:56 +0000

Eric, great article.

To expand on your points regarding why ranges around forecasts are so important. For example, if we have an upper/lower forecast range at a 95% confidence level we are actually saying that 9 out of 10 future forecasts will fall into that range. If they do, then the forecast should be considered accurate. Those same upper/lower forecast ranges should be used as part of the inventory safety stock calculations. If so, then inventory safety stock should be able to cover the variation in the forecast. That said, the more accurate (lower error) the lower safety stock required. Their is a direct correlation between accurate forecasts with low errors and the reduction of inventory safety stock.

As you mentioned we focus too much on forecasting the actual number versus staying within the upper/lower forecast range.

By the way we never say that the Financial Plan, Supply Plan, Inventory Plan, or any plan for that matter is also always wrong, and many times not as a result of a poor forecast. In many cases, very accurate forecasts are ignored because they don’t meet the needs or purpose of the departments who consume the forecast.

It’s time that we call out the truth.