Mean Absolute Percentage Error (MAPE)

Definition

Mean Absolute Percentage Error (MAPE) is a forecasting accuracy metric that calculates the average of the absolute percentage differences between forecast values and actual observed values across a set of periods, products, or transactions.

What is Mean Absolute Percentage Error (MAPE)?

MAPE is used to show how far a forecast deviates from actual results in percentage terms. Because it expresses error relative to the actual value, it is easy to interpret across products or time periods with different absolute scales. A lower MAPE indicates that forecasts are, on average, closer to actual outcomes. This makes the metric common in demand planning, inventory management, sales forecasting, and supply chain performance reporting.

Its popularity comes from interpretability, but it must be used carefully. Percentage based error behaves poorly when actual values are very small or zero, and it can overemphasize errors in low volume items.

How to Calculate MAPE

The standard formula is the mean of absolute error divided by actual value, expressed as a percentage. For each observation, calculate the absolute value of actual minus forecast, divide by actual, and multiply by one hundred. Then average those percentages across all observations in scope.

For example, if actual demand is 100 and the forecast is 90, the absolute percentage error is 10 percent. If actual demand is 50 and the forecast is 60, the absolute percentage error is 20 percent. MAPE would average those errors across the selected dataset.

How MAPE Is Used in Planning

Planners use MAPE to compare forecast performance across products, locations, time buckets, planners, or forecasting models. It helps show whether a forecasting process is improving over time and whether one method produces more reliable predictions than another. In procurement and supply chain settings, that matters because poor forecasting drives excess inventory, shortages, unstable replenishment, and unnecessary expediting.

MAPE can also support segmentation. High volume stable items may be held to a different forecast accuracy standard than promotional, intermittent, or new launch items because the predictability of demand is inherently different.

Limitations of MAPE

MAPE becomes problematic when actual values are zero because division by zero is undefined. It can also become very large when actual values are close to zero, which makes the metric unstable for intermittent demand. Another limitation is that it treats over forecasting and under forecasting symmetrically even though the business consequences may differ.

For these reasons, organizations often use MAPE alongside other measures such as bias, weighted error metrics, or service impact indicators rather than relying on it alone.

MAPE in Procurement and Inventory Decisions

Forecast accuracy affects procurement order timing, safety stock, capacity reservation, and supplier scheduling. A rising MAPE may indicate that replenishment parameters are becoming unreliable and that inventory policy needs review. Procurement teams also use forecast accuracy discussions in supplier collaboration because unstable demand signals can increase supplier cost and reduce responsiveness.

Frequently Asked Questions about Mean Absolute Percentage Error (MAPE)

What does a lower MAPE mean?

A lower MAPE means the forecast is, on average, closer to the actual values in percentage terms. If one forecasting process has a MAPE of 8 percent and another has a MAPE of 18 percent for comparable data, the first process is generally more accurate. However, the interpretation is only meaningful when the data scope, time buckets, and item characteristics are comparable.

Why is MAPE not suitable when actual demand is zero?

MAPE divides the forecast error by the actual value. If the actual value is zero, the calculation is undefined because division by zero is not possible. Even when actual demand is very small rather than zero, the percentage error can become disproportionately large and misleading. That is why planners often use alternative metrics for intermittent or low volume demand patterns.

Is MAPE enough to judge forecast quality by itself?

No. MAPE shows average percentage error, but it does not indicate whether the forecast is consistently too high or too low, and it does not reflect the operational importance of specific errors. A forecast could have an acceptable MAPE while still creating stockouts on critical items. Most mature planning teams therefore review MAPE together with bias, service impact, and item segmentation.

How does MAPE affect procurement decisions?

Forecast accuracy influences the quality of purchasing and replenishment decisions because order quantities, supplier schedules, and safety stock settings are all built on expected demand. If MAPE is high, procurement may place orders too early, too late, or in the wrong quantity. The result can be excess inventory, missed sales, supplier disruption, or increased expedite cost.