Measuring Forecast Performance

Measuring forecast performance is one of the most important elements of the demand forecasting process. As you know, what gets measured gets fixed, and what gets fixed gets rewarded. You cannot improve your demand forecast accuracy until you measure and benchmark your current forecast performance. It is not unusual to encounter companies that have never truly measured the accuracy of their demand forecasts on an ongoing (e.g., weekly, monthly) basis. Some measure forecast accuracy quarterly, but many still do not measure forecast performance as part of the weekly or monthly demand forecasting process. Those that do only measure forecast accuracy at the aggregate level, with little focus on the stock-keeping unit (SKU) detail or internal mix within the aggregates. It is not uncommon to find that many companies have virtually no idea that their lower-level product forecasts at the product group and the SKU detail have extremely high forecast error (or very low forecast accuracy). This is usually attributed to the way they calculate forecast accuracy (or error) in my experience with several companies. They normally do not measure forecast error in terms of absolute values; thus, when they sum those error values to the aggregate levels, the plus and minus signs wash each other out, making the accuracy look much better than the lower-level detail. In fact, most senior-level managers rarely use or understand the term forecast error. As a result, ...

Get Demand-Driven Forecasting: A Structured Approach to Forecasting, 2nd Edition now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.