Mean Absolute Deviation

Measuring Consistency: The Mean Absolute Deviation (MAD)

In statistics, calculating the "average" (mean) of a dataset is often the first step, but it only tells half the story. It doesn't tell you how spread out the data is. That's where dispersion comes in. The Mean Absolute Deviation (MAD) is one of the most intuitive ways to measure this variability.

What is MAD?

The Mean Absolute Deviation answers a simple question:
"On average, how far is each data point from the mean?"

A low MAD indicates that the data points are clustered closely around the average (consistent data).
A high MAD indicates that the data points are spread far apart (volatile data).

Step-by-Step Calculation Formula

Let's calculate the MAD for a simple dataset: {2, 5, 9, 12}.

Step 1: Find the Mean (Average).
(2 + 5 + 9 + 12) / 4 = 28 / 4 = 7.

Step 2: Find the Deviation for each number.
Subtract the mean from each value:
2 - 7 = -5
5 - 7 = -2
9 - 7 = 2
12 - 7 = 5

Step 3: Take the Absolute Value.
We ignore negatives because distance is always positive.
|-5| = 5
|-2| = 2
|2| = 2
|5| = 5

Step 4: Find the Mean of the Deviations.
Add them up and divide by the count (n=4):
(5 + 2 + 2 + 5) / 4 = 14 / 4 = 3.5.

The MAD is 3.5.

MAD vs. Standard Deviation

You might ask, "Why not use Standard Deviation?" Constructing Standard Deviation involves squaring the differences before averaging them.
- Standard Deviation penalizes outliers heavily (because squaring a large number makes it huge). It is preferred for complex mathematics and normal distributions.
- MAD treats all deviations linearly. It is often considered more "robust" if your data has extreme outliers that you don't want to skew the result too much. It is easier to interpret for real-world scenarios like forecasting errors.

Real-World Example: Supply Chain

Imagine you manage a warehouse.
- Product A sales per day: {10, 10, 10, 10}. Mean = 10, MAD = 0.
- Product B sales per day: {0, 0, 20, 20}. Mean = 10, MAD = 10.
Both products sell roughly the same amount per week. However, Product B is highly unpredictable (high MAD), so you need to keep more safety stock inventory to avoid running out on the busy days. Product A is predictable, so you can run "lean."