Demand Management &

Forecasting
Role of demand management in
MPC system

MPC
Boundary

Resource
Planning

Sales &
Operations
Planning

Demand
Management

Master
Production
Schedule
(MPS)

Demand Management in MPC

System

Marketplace

Demand Management &

Forecasting
Independent
Demand
Demand from the market/customer is
independent demand
Demand is not related to the demand for
other assemblies or products, instead it is
from outside sources
Independent demand - must be forecasted.
Wagons, bicycles, finished goods, services,
etc.
Independent demand can be influenced

Demand Management &

Forecasting
Dependent Demand

Generally related to production of an end

product
Can be calculated instead of forecasted
Demand of components in the assembly of a
product is often dependent demand

1/8

Demand Management &

Forecasting
Forecasts and
Plans
What we actually plan to deliver to
customers each period is the output of the
process
This is based on marketing quotas, special
sales incentives, etc.
These amounts will be based on inputs from
many different sources and not just
quantitative forecasts.
A manufacturing manager cannot be held
responsible for not getting a forecast right

Demand Management &

Forecasting
Forecasts
and
Plans
A manufacturing
manager can and should
be held responsible for making their plans
Providing the means for making as good a
set of executable plans as possible and
then providing the information to execute
them.
When conditions change the control
function should change the plans and
The new plans should be executed
faithfully

Demand Management &

Forecasting
MPC Environment

Engineer
Make

to order (ETO)

to order (MTO)

Assemble

to order

(ATO)
Make to stock (MTS)

Demand Management &

Forecasting
MPC Environment

Lead Time

Short

Long

Finished Goods

Make-to-Stock (MTS)

Components/Subassemblies
Raw Materials

Assemble to Order (ATO

Make to Order (MTO)

Suppliers

Engineer to Order

Customer Order Decoupling

Point

Demand Management &

Forecasting
MTS
Environment

Focus is maintenance of finished goods

inventory
TRACKING of demand by location throughout
the supply chain is an important activity
Key issue is HOW, WHEN,& HOW MUCH, to
REPLENISH STOCK at a specific location
Firms employ distribution centers, warehouses,
and even vendor-managed inventory inside
their customers location

Demand Management &

Forecasting
Managers require
information on the
MTS Environment
INVENTORY STATUS in the various locations
Managers require information on
relationships with transportation providers,
and estimates of demand by location and
item
Satisfying customers requires to BALANCE
the level of inventory against the level of
service to the customers.
IMPROVEMENTS can be made by having
better knowledge of demand, rapid
transportation alternatives, speedier
production, more flexibility

Demand Management &

Forecasting
ATO Environment

The primary task of Demand Management is to

DEFINE THE CUSTOMERS ORDER in terms of
alternative components and options.
One of the capabilities required for success is
ENGINEERING DESIGN
The inventory that defines customer service is
the inventory of COMPONENTS not finished
products
The number of finished products is usually
substantially greater than the number of
components that are combined to produce the
finished product

Demand Management &

Forecasting
MTO/ETO
Environment
MOVING THE CUSTOMER decoupling point to
raw material or even suppliers reduces the
scope of dependent demand information
The task of DM in this environment is to
COORDINATE INFORMATION on customers
product needs with engineering
DM now includes determining HOW MUCH
ENGINEERING CAPACITY will be required to
meet future customer needs
In these environments, suppliers capabilities
may limit what we are able to do, so
COORDINATION with them is essential

Demand Management &

Forecasting
Communicating with
S&OP
MTS Environment
Demand Forecasts
ATO Environment
Demand Forecasts,
Product Family Mix
MTO/ETO Environment
Demand Forecasts
Engineering Details

Demand Management &

Forecasting
Communicating with
MPS

Interaction between DM and MPS are frequent

and detailed
Details vary significantly between the three
environments
Types of uncertainties also differ between the
three environments

Demand Management &

Forecasting
Producing Forecast
Subjective
Judgmental
Sales force surveys
Delphi techniques
Jury of experts

Experimental
Customer surveys
Focus group sessions
Test Marketing
Simulation

Demand Management &

Forecasting
Producing Forecast

Objective
Time Series
Uses past to predict the future

Causal Methods

Demand Management &

Forecasting
Time Series
Forecasting

The typical problem:

Generate the large number of short-term, SKU
level, locally disaggregated demand forecasts
required for production, logistics, and sales to
operate successfully

Predominant use is for:

Forecasting product demand of
Mature products over a. .
Short time horizon (weeks, months, quarters,
year) . . .

Demand Management &

Forecasting
Time Series
Forecasting

Using models to assist in the forecast where . . .

Demand of items is independent
Special situations are treated differently

New product introduction

Old product retirement
Short life-cycle products

Demand Management &

Forecasting
Basic Components
Level
Value where demand hovers around

Trend
Persistent movement in one direction
Typically linear but can be exponential,
quadratic, etc.

Cyclical Movements
Periodic movement not tied to calendar

Demand Management &

Forecasting
Basic Components
Seasonal Variations
Movement that is
periodic to the calendar
Hourly, daily, weekly,
monthly, quarterly, etc.

Random Fluctuations
Irregular and
unpredictable
variations, noise

Demand Management &

Forecasting
Time
SeriesForecast
Methods
Cumulative
All history matters
equally
Pure stationary demand

Nave Forecast
Most recent dictates next
Random Walk, Last is
Next

Moving Average
Only include the last M
observations
Compromise between
cumulative and nave

Demand Management &

Forecasting
Time Series Methods

Simple Exponential Smoothing

Why should past observations all be
weighted the same?
Value of observation degrades over
time
Introduce smoothing constant ()

Exponential Smoothing with

trend (Holts Method)
Similar to exponential smoothing
Holts Method -smoothing constants
for level () and trend () terms

Demand Management &

Forecasting
Moving Averages
(MA)
Simple Moving Average
(SMA)
Week
1
2
3
4
5
6

Sales Three Period SMA

17
21
19
23 SMA
18 SMA
16SMA

Demand Management &

Forecasting
Moving Averages
(MA)
Weighted Moving Averages (WMA)
In SMA each observation in the calculation
receives the same weight.
In WMA each observation in the calculation

receives different weight.
Usually most recent observation in the calculation
receives the largest weight.
For example WMA forecast for week 4 =

Demand Management &

Forecasting
Exponential
Smoothing

where:
=forecast for time period t+1
= forecast for time period t
= actual value for time period t
= smoothing constant,
01

Demand Management &

Forecasting
Exponential Smoothing
Period

Demand

Forecast

50

40

46

0.3 * 50 + (1 - 0.3) * 40 = 43.0

52

0.3 * 46 + (1 - 0.3) * 43 = 43.9

48

0.3 * 52 + (1 - 0.3) * 43.9 = 46.33

47

0.3 * 48 + (1 - 0.3) * 46.33 = 46.83

0.3 * 47 + (1 - 0.3) * 46.83 = 46.88

Demand Management &

Forecasting
Example ES
What value of should be used in the
exponential smoothing model?
If the raw data has more randomness then
small value of should be used in the model.
If the raw data has less randomness then
large value of should be used in the model.

Demand Management &

Forecasting
Period
1
2
3
4
5
6
7
8
9
10

Demand
10
11
9
11
10
8
22
20
21
19

=0.2
=0.8
Forecast Forecast
10
10
10.00 10.00
10.20 10.80
9.96
9.36
10.17 10.67
10.13 10.13
9.70
8.43
12.16 19.29
13.73 19.86
15.18 20.77

Demand Management &

Forecasting
Adjusted ES with
Trend
where:

trend factor for the next period

trend factor for the current period

= smoothing constant for trend, 0 1

Demand Management &

Forecasting
Example Adjusted
ES
Lets assume
= 0.3, = 0.6, =30,
=27 (to start the process initial value is
assumed) and
=0 (to start the process initial value is
assumed)

Demand Management &

Forecasting

Period
1
2
3
4
5
6
7
8
9
10

Unadjusted
Trend
Demand Forecast,
30
27
0
34
27.9
0.54 28.44
37
29.73 1.31 31.04
40
31.91 1.83 33.75
44
34.34 2.19 36.53
48
37.24 2.62 39.85
51
40.47 2.98 43.45
55
43.63 3.09 46.72
58
47.04 3.28 50.32
50.33 3.29 53.61

Demand Management &

Forecasting
Example Adjusted
ES
Moving Averages and Simple Exponential
Methods can only forecast only one period in
future
Adjusted Exponential Smoothing can be used
to forecast any number of period in future

From the example and

Therefore,

Demand Management &

Forecasting
Seasonal
Demand
For each season,
can compute a seasonal
index:
Seasonal Index = (Period average demand)/
(Average Demand for all periods)
Index can be used as a multiplier for future
seasons
Use only deseasonalized data to forecast
Deseasonalized demand = (actual seasonal
demand)/(seasonal index)
Base forecast is deseasonalized demand
Calculate seasonal forecasts by applying
seasonal indices to base forecast

Demand Management &

Forecasting

Demand Management &

Forecasting
Calculating Seasonal Indices:
Average quarterly demand = 400/4 = 100
units
Seasonal indices
Quarter 1: 128/100 = 1.28
Quarter 2: 102/100 = 1.02
Quarter 3: 75/100 = 0.75
Quarter 4: 95/100 = 0.95

Demand Management &

Forecasting
Applying Seasonal Indices:
Suppose the same company forecast an
annual demand of 420 units next year.
The average quarterly demand = 105
units. Applying seasonal indices:
Quarter 1: 1.28 x 105 = 134.4 units
Quarter 2: 1.20 x 105 = 107.1 units
Quarter 3: 0.75 x 105 = 78.75 units
Quarter 4: 0.95 x 105 = 99.75 units

Demand Management &

Forecasting
Evaluating Forecast
Basic rule assume the forecast is incorrect. The
key issue: How incorrect is it and what do we do
about it?
The error can be used to:
Evaluate and possibly change forecasting
methodology
Apply buffer stock or capacity to account for
possible error
Systematic error in which the actual demand is
consistently above or below the forecasted demand
When exists, evaluate forecast to improve accuracy

Demand Management &

Forecasting
Evaluating Forecast
Month
1
2
3
4
5
Total

Forecast,
100
100
100
100
100
500

Actual, A
106
6
93 -7
102
2
101
1
95 -5
497
-3

Error=A-F

Demand Management &

Forecasting
Mean Absolute
Deviation

MAD = (Sum of absolute deviations)/(Number

of deviations)
A positive number that indicates the average
value of forecast error during the time of
evaluation

Demand Management &

Forecasting
Mean Absolute Deviation (MAD)

Demand Management &

Forecasting
Mean Absolute
Deviation
Note there is no bias over the six months
shown
The MAD
Sum of absolute deviations =
5+6+2+4+3+4 = 24
MAD = 24/6 = 4
This means while the forecast method
showed no bias, it was in error by an average
of 4 units per month