Mean Absolute Percentage Error (MAPE)

What is Mean Absolute Percentage Error (MAPE)?

Mean Absolute Percentage Error (MAPE) is a metric used to measure how accurate a forecast was, expressed as the average percentage difference between predicted and actual values. In workforce management, it’s most commonly used to evaluate how well a call volume forecast held up against what actually happened.

How it’s calculated

For each forecasted interval, the absolute difference between the predicted and actual value is divided by the actual value, then converted to a percentage. Those percentage errors are averaged across all intervals in the period being measured, producing a single MAPE figure. A MAPE of 5% means the forecast was off by an average of 5% across the measured intervals; a MAPE of 25% indicates a considerably less reliable forecast.

Why WFM teams track this specifically

A forecast that looks reasonable on a chart can still be systematically wrong in ways that hurt staffing decisions, consistently under-forecasting Monday mornings, for instance, while being accurate on average across the week. MAPE gives a single, comparable number that lets a WFM team track forecast accuracy over time, compare different forecasting methods or tools against each other, and catch a degrading model before it causes a staffing problem.

Use cases

• Evaluating forecasting models: comparing MAPE across different forecasting approaches or software to choose the most accurate one.
• Ongoing forecast monitoring: tracking MAPE weekly or monthly to catch a forecast that’s drifting out of line with reality.
• Vendor and tool evaluation: using MAPE as a benchmark when assessing a new WFM platform’s forecasting claims.

A limitation worth knowing

MAPE can behave oddly when actual values are very close to zero, since dividing by a tiny number inflates the percentage error disproportionately. For contact centers with intervals of genuinely very low call volume, such as overnight hours, MAPE alone can be misleading, and many WFM teams pair it with other accuracy metrics for those specific periods.