8 Forecasting metrics
Why is forecasting metrics important for practitioners and businesses? Because insightful metrics provide valuable (and actionable!) feedback on what can be done to improve both the forecast and overall business performance. Simply put, you can only improve what you can measure. So, choosing the right metrics should be the first step of any improvement journey.
Unfortunately, many practitioners struggle to decide which forecasting metrics to utilize. Or do not understand or use them properly. Choosing the best metric(s) to assess the quality of a forecasting process (or model) is anything but straightforward. And it will have a profound impact on the resulting forecasts and on the subsequent business decisions! Depending on the metric selected, you might promote biased forecasts, give too much importance to extreme values (chapter 9), overlook critical products (chapter 11), or focus on the wrong forecasting horizon (chapter 6).
Measuring forecast accuracy (or error) is not a simple task, because there is no one-size-fits-all indicator. Each indicator will avoid some pitfalls only to present others.
We will first discuss in this chapter how to compute forecasting KPIs. In chapter 9, we will discuss the pros and cons of each metric. We will finally identify a set of metrics that will be the best compromise to support (most) demand-planning processes.