What factors go into the Mean Absolute Deviation Percentage (MADP) calculation?

Mean Absolute Deviation Percentage (MADP) is a measurement of the items demand fluctuations.

The MADP components include:

• Old Average (Demand History)
• Index (Seasonality)
• Demand (Actual Shipped + Lost - Promo)
• Error Rate (Proprietary)
• Old MADP
• Alpha Factor (Proprietary)

MADP Ranges:
 

Example: Demand History 8, 12, 8, 12, 8, 12
Average 10
MADP .20 or 20%

Example: Demand History 6, 14, 6, 14, 6, 14
Average 10
MADP .40 or 40%

MADP is in principle calculated as shown in the next example:

Example:
Demand History 8, 14, 6, 12
Average is then (8 + 14 + 6 + 12) / 4 = 10
MADP is then (2 + 4 + 4 + 2) / (4x10) = .30

The differences between the demands and the average 2, 4, 4, 2, are added regardless of their signs: ie,
the actual differences (Demand – Average) are –2, +4, -4, +2, but the sum of their absolute values is 12.

MADP is revised at the end of each demand history period (with exponential smoothing).

New MADP = Old MADP + Alpha ??(Error Rate – Old MADP)
Example: Old Average = 100
Index = 1.0
Demand = 130
Error Rate = +.30
Old MADP = .20
Alpha = .1
New MADP = .20 + .1 (.30 - .20) = .21

Example: Old Average = 100
Index = 1.0
Demand = 70
Error Rate = -.30
Old MADP = .20
Alpha = .1
New MADP = .20 + .1 (.30 - .20) = .21