Table of Contents
What it is
The Mean Absolute Percentage Error, commonly referred to as the MAPE, is defined as a statistical forecasting method that measures how accurate a prediction is. The MAPE is used as part of a regression analysis which mathematically sorts how independent variables are forecast to impact a dependent variable.
Why it’s used
The MAPE checks the accuracy of your demand forecast. In addition, it’s scale-independent, meaning it can be used to compare series on different scales.
Incorrect demand forecasts are a barrier to creating accurate and timely demand plans, which is a breakdown of the supply chain.
MAPE will be expressed as a percentage; the lower the MAPE value, the more useful the forecast is.
What it does not do
Understanding that a higher MAPE doesn’t mean the employee who created it did anything wrong is vital. If MAPE is the only measure used, you’re missing out on other valuable performance indicators.
Mean Absolute Percentage Error isn’t actually absolute; it’s not the ultimate and final measurement, but it remains a valuable tool for tracking performance indicators. So don’t do away with MAPE usage; add additional tools to your toolbox.
Additionally, setting a goal of MAPE 0% is missing the point of coming up with a reasonable percentage of error in the forecast. It’s also a generally impossible goal in real terms. Nevertheless, MAPE is the most commonly used measure to forecast error because it’s expressed as a percentage, so it’s relatively easy to understand.
Also read: Starting a Furniture Business: 4 Key Steps to Take
Make MAPE work for you
MAPE is commonly used to look at a sales forecast and predict the likely percentage of error. Forecast error is expressed as the gap between the forecast quantity and the actual quantity. It answers the question of whether demand met or exceeded the supply released on the market. Or whether the forecast caused too little supply to meet demand.
Drilling down to what MAPE reveals, any tendency for the forecast to regularly be higher or lower than the actual value. That’s forecast bias. Bias means estimates are wrong more often than they’re right, and the cause is not generally employee error, but mistakes in the system used to formulate the estimates.
Keep in mind that you can get sidetracked trying to set a “good, better, best, poor” level of forecast accuracy. For example, to arbitrarily say a MAPE of 6% is really good and 25% is really bad. Step back, remind yourself of the big picture and think about your ultimate goal which is delivering your goods in the way the market will best receive them. Forecasting is important for planning and reviewing how well the plan is performed.
MAPE is valuable for the early identification of changes in forecast accuracy. It also helps identify systemic changes.
In our post-pandemic world, markets want to rebound, but it’s proving to be a bumpy ride. Keeping an eye on MAPE and how the error percentage points to market changes allows you to stay nimble.
Increase the accuracy and value of all of your forecasts by folding in the use of bias and the Forecast Value Add (FVA). The complexity is worth the resulting gains.
