MAD Calculator (Mean Absolute Deviation)

About the MAD Calculator (Mean Absolute Deviation)

Mean Absolute Deviation (MAD) measures forecast accuracy in the same units as demand — pieces, cases, or dollars. Unlike MAPE, which expresses error as a percentage, MAD gives you the average absolute number of units by which the forecast misses actual demand.

MAD is especially useful when comparing forecasts for items with similar demand volumes or when you need to translate forecast error into inventory buffer requirements. Many safety stock formulas use MAD as an input to calculate the amount of buffer stock needed.

This calculator takes pairs of actual and forecast values and computes MAD, along with the total absolute deviation and number of periods evaluated.

Use the result to compare operating scenarios, pressure-test assumptions, and rerun the model when volumes, rates, or service targets change.

Use the output to compare options, spot the main cost drivers, and rerun the math when lane assumptions or operating constraints change.

Use the output to compare options, spot the main cost drivers, and rerun the math when lane assumptions or operating constraints change.

When This Page Helps

MAD is the most practical forecast error metric for operations because it speaks the language of the warehouse — units. It directly feeds into safety stock calculations and helps planners understand how much buffer inventory is needed to compensate for forecast inaccuracy.

How to Use the Inputs

1. Enter actual demand values separated by commas.
2. Enter corresponding forecast values separated by commas.
3. Ensure both lists have the same number of values.
4. Review the MAD result in demand units.
5. Use MAD to calibrate safety stock levels.
6. Track MAD over time to measure forecasting improvement.

Example Calculation

Tips & Best Practices

• MAD is used as an input for safety stock: Safety Stock = z × MAD × 1.25 (approximate conversion from MAD to standard deviation).
• Lower MAD means less safety stock needed and lower inventory investment.
• MAD doesn't indicate bias direction — combine with Forecast Bias for a complete view.
• Compare MAD across items at similar demand levels for meaningful ranking.
• For items with very different scales, use MAPE or MASE for cross-item comparison.
• Track MAD by SKU family and update safety stock parameters as MAD changes.

MAD and Safety Stock

Many supply chain textbooks derive safety stock as z × σ × √L, where σ is the standard deviation of demand. Since MAD ≈ 0.8 × σ for normally distributed data, you can substitute σ ≈ 1.25 × MAD. This makes MAD a practical, easy-to-compute input for safety stock calculations.

Tracking MAD Over Time

Plot MAD monthly alongside changes in forecasting methods or data sources. A declining MAD trend confirms that your improvements are working. Be alert to sudden MAD increases that may signal data quality issues, demand pattern shifts, or supply disruptions.

MAD by Product Segment

Group SKUs by demand pattern (stable, trending, seasonal, intermittent) and track MAD for each group separately. Stable-demand items should have low MAD; intermittent items will naturally have higher MAD. Setting group-specific MAD targets is more meaningful than a single company-wide target.

Frequently Asked Questions

• What is Mean Absolute Deviation?

  MAD is the average of the absolute differences between actual demand and forecast demand across all periods. It measures forecast error magnitude in the same units as demand.

• How is MAD different from MAPE?

  MAD expresses error in absolute units (e.g., 50 pieces). MAPE expresses error as a percentage of actual demand. MAD is better for safety stock calculations; MAPE is better for cross-item comparison.

• How does MAD relate to standard deviation?

  For normally distributed errors, the standard deviation is approximately 1.25 × MAD. This conversion is used in safety stock formulas that require standard deviation as input.

• What is a good MAD value?

  A "good" MAD depends on the item's demand level. MAD of 5 is excellent for an item with demand of 100 (5% error) but poor for an item with demand of 10 (50% error). Always evaluate MAD relative to demand volume.

• Can MAD be zero?

  Yes, MAD equals zero when the forecast exactly matches actual demand in every period. This is practically impossible but serves as the theoretical best case.

• Should I use MAD or MSE?

  MAD treats all errors equally. MSE squares errors, penalizing large deviations more heavily. If large errors are particularly costly, MSE is more appropriate. For general use and safety stock, MAD is standard.