McGraw-Hill/Irwin Copyright 2009 by The McGraw-Hill Companies, Inc. All Rights Reserved.

Chapter 3
Forecasting
3-2
Forecast
Forecast a statement about the future value
of a variable of interest
We make forecasts about such things as weather,
demand, and resource availability
Forecasts are an important element in making
informed decisions
3-3
Two Important Aspects of Forecasts
Expected level of demand
The level of demand may be a function of some
structural variation such as trend or seasonal variation
Accuracy
Related to the potential size of forecast error
3-4
Features Common to All Forecasts
1. Techniques assume some underlying causal system
that existed in the past will persist into the future
2. Forecasts are not perfect
3. Forecasts for groups of items are more accurate than
those for individual items
4. Forecast accuracy decreases as the forecasting
horizon increases
3-5
Elements of a Good Forecast
The forecast
should be timely
should be accurate
should be reliable
should be expressed in meaningful units
should be in writing
technique should be simple to understand and use
should be cost effective
3-6
Steps in the Forecasting Process
1. Determine the purpose of the forecast
2. Establish a time horizon
3. Select a forecasting technique
4. Obtain, clean, and analyze appropriate data
5. Make the forecast
6. Monitor the forecast
3-7
Forecast Accuracy and Control
Forecasters want to minimize forecast errors
It is nearly impossible to correctly forecast real-world
variable values on a regular basis
So, it is important to provide an indication of the
extent to which the forecast might deviate from the
value of the variable that actually occurs
Forecast accuracy should be an important
forecasting technique selection criterion
3-8
Forecast Accuracy and Control (contd.)
Forecast errors should be monitored
Error = Actual Forecast
If errors fall beyond acceptable bounds, corrective
action may be necessary
3-9
Forecast Accuracy Metrics
n
!
"
=
t t
Forecast Actual
MAD
( )
2
t t
1
Forecast Actual
MSE
!
!
=
"
n
n
!
"
#
=
100
Actual
Forecast Actual
MAPE
t
t t
MAD weights all errors
evenly
MSE weights errors according
to their squared values
MAPE weights errors
according to relative error
3-10
Forecast Error Calculation
Period
Actual
(A)
Forecast
(F)
(A-F)
Error
|Error| Error
2
[|Error|/Actual]x100
1 107 110 -3 3 9 2.80%
2 125 121 4 4 16 3.20%
3 115 112 3 3 9 2.61%
4 118 120 -2 2 4 1.69%
5 108 109 1 1 1 0.93%
Sum 13 39 11.23%
n = 5 n-1 = 4 n = 5
MAD MSE MAPE
= 2.6 = 9.75 = 2.25%
3-11
Forecasting Approaches
Qualitative Forecasting
Qualitative techniques permit the inclusion of soft information such as:
Human factors
Personal opinions
Hunches
These factors are difficult, or impossible, to quantify
Quantitative Forecasting
Quantitative techniques involve either the projection of historical data or
the development of associative methods that attempt to use causal
variables to make a forecast
These techniques rely on hard data
3-12
Judgmental Forecasts
Forecasts that use submective inputs such as
opinions from consumer surveys, sales staff,
managers, executives, and experts
Executive opinions
Salesforce opinions
Consumer surveys
Delphi method
3-13
Time-Series Forecasts
Forecasts that project patterns identified in
recent time-series observations
Time-series - a time-ordered sequence of
observations taken at regular time intervals
Assume that future values of the time-series can
be estimated from past values of the time-series
3-14
Time-Series Behaviors
Trend
Seasonality
Cycles
Irregular variations
Random variation
3-15
Trends and Seasonality
Trend
A long-term upward or downward movement in data
Population shifts
Changing income
Seasonality
Short-term, fairly regular variations related to the calendar or
time of day
Restaurants, service call centers, and theaters all experience
seasonal demand
3-16
Cycles and Variations
Cycle
Wavelike variations lasting more than one year
These are often related to a variety of economic, political, or even
agricultural conditions
Random Variation
Residual variation that remains after all other behaviors have
been accounted for
Irregular variation
Due to unusual circumstances that do not reflect typical behavior
Labor strike
Weather event
3-17
Time-Series Forecasting - Nave Forecast
Nave Forecast
Uses a single previous value of a time series as the
basis for a forecast
The forecast for a time period is equal to the
previous time periods value
Can be used when
The time series is stable
There is a trend
There is seasonality
3-18
Time-Series Forecasting - Averaging
These Techniques work best when a series
tends to vary about an average
Averaging techniques smooth variations in the data
They can handle step changes or gradual changes in
the level of a series
Techniques
Moving average
Weighted moving average
Exponential smoothing
3-19
Moving Average
Technique that averages a number of the most
recent actual values in generating a forecast
average moving in the periods of Number
1 period in value Actual
average moving period MA
period for time Forecast
where
MA
1
1
t
=
! =
=
=
= =
!
=
! "
n
t A
n
t F
n
A
F
t
t
t
n
i
i t
t
3-20
Moving Average
As new data become available, the forecast is
updated by adding the newest value and
dropping the oldest and then recomputing the
the average
The number of data points included in the
average determines the models sensitivity
Fewer data points used-- more responsive
More data points used-- less responsive
3-21
Weighted Moving Average
The most recent values in a time series are
given more weight in computing a forecast
The choice of weights, w, is somewhat arbitrary and
involves some trial and error
!
F
t
= w
n
A
t"n
+ w
n"1
A
t"(n"1)
+ ... + w
1
A
t"1
where
w
t
= weight for period t, w
t"1
= weight for period t "1, etc.
A
t
= the actual value for period t, A
t"1
= the actual value for period t "1, etc.
3-22
Exponential Smoothing
A weighted averaging method that is based on
the previous forecast plus a percentage of the
forecast error
!
F
t
= F
t"1
+#(A
t"1
+ F
t"1
)
where
F
t
= Forecast for period t
F
t"1
= Forecast for the previous period
# =Smoothing constant
A
t"1
= Actual demand or sales from the previous period
3-23
Other Forecasting Methods - Focus
Focus Forecasting
Some companies use forecasts based on a best
current performance basis
Apply several forecasting methods to the last several
periods of historical data
The method with the highest accuracy is used to make
the forecast for the following period
This process is repeated each month
3-24
Other Forecasting Methods - Diffusion
Diffusion Models
Historical data on which to base a forecast are not
available for new products
Predictions are based on rates of product adoption and
usage spread from other established products
Take into account facts such as
Market potential
Attention from mass media
Word-of-mouth
3-25
Techniques for Trend
Linear trend equation
Non-linear trends
Parabolic trend equation
Exponential trend equation
Growth curve trend equation
3-26
Linear Trend
A simple data plot can reveal the existence and
nature of a trend
Linear trend equation
!
F
t
= a + bt
where
F
t
= Forecast for period t
a = Value of F
t
at t = 0
b =Slope of the line
t =Specified number of time periods from t = 0
3-27
Estimating slope and intercept
Slope and intercept can be estimated from
historical data
!
b =
n ty " t y
# # #
n t
2
" t
#
( )
#
2
a =
y "b t
# #
n
or y "bt
where
n = Number of periods
y = Value of the time series
3-28
Trend-Adjusted Exponential Smoothing
The trend adjusted forecast consists of two
components
Smoothed error
Trend factor
!
TAF
t +1
= S
t
+ T
t
where
S
t
= Previous forecast plus smoothed error
T
t
= Current trend estimate
3-29
Trend-Adjusted Exponential Smoothing
Alpha and beta are smoothing constants
Trend-adjusted exponential smoothing has the
ability to respond to changes in trend
!
TAF
t +1
= S
t
+ T
t
S
t
= TAF
t
+" A
t
# TAF
t
( )
T
t
= T
t#1
+ $ TAF
t
# TAF
t#1
#T
t#1
( )
3-30
Techniques for Seasonality
Seasonality is expressed in terms of the amount that
actual values deviate from the average value of a series
Models of seasonality
Additive
Seasonality is expressed as a quantity that gets added or
subtracted from the time-series average in order to
incorporate seasonality
Multiplicative
Seasonality is expressed as a percentage of the average (or
trend) amount which is then used to multiply the value of a
series in order to incorporate seasonality
3-31
Seasonal Relatives
Seasonal relatives
The seasonal percentage used in the multiplicative seasonally
adjusted forecasting model
Using seasonal relatives
To deseasonalize data
Done in order to get a clearer picture of the nonseasonal
components of the data series
Divide each data point by its seasonal relative
To incorporate seasonality in a forecast
Obtain trend estimates for desired periods using a trend
equation
Add seasonality by multiplying these trend estimates by the
corresponding seasonal relative
3-32
Techniques for Cycles
Cycles are similar to seasonal variations but are of
longer duration
Explanatory approach
Search for another variable that relates to, and leads, the
variable of interest
Housing starts precede demand for products and services
directly related to construction of new homes
If a high correlation can be established with a leading
variable, it can develop an equation that describes the
relationship, enabling forecasts to be made
3-33
Associative Forecasting Techniques
Home values may be related to such factors as home
and property size, location, number of bedrooms, and
number of bathrooms
Associative techniques are based on the development
of an equation that summarizes the effects of predictor
variables
Predictor variables - variables that can be used to
predict values of the variable of interest
3-34
Simple Linear Regression
Regression - a technique for fitting a line to a
set of data points
Simple linear regression - the simplest form of
regression that involves a linear relationship between
two variables
The object of simple linear regression is to obtain an
equation of a straight line that minimizes the sum of
squared vertical deviations from the line (i.e., the least
squares criterion)
3-35
Least Squares Line
!
y
c
= a + bx
where
y
c
= Predicted (dependent) variable
x = Predicted (independent) variable
b =Slope of the line
a = Value of y
c
when x = 0 (i.e., the height of the line at the y intercept)
and
b =
n xy " x y
# # #
n x
2
" x
#
( )
#
2
a =
y "b x
# #
n
or y "bx
where
n = Number of paired observations
3-36
Standard Error
Standard error of estimate
A measure of the scatter of points around a
regression line
If the standard error is relatively small, the predictions
using the linear equation will tend to be more accurate
than if the standard error is larger
!
S
e
=
y " y
c
( )
2
#
n "2
where
S
e
= standard error of estimate
y = the value of each data point
n = number of data points
3-37
Correlation Coefficient
Correlation
A measure of the strength and direction of relationship between
two variables
Ranges between -1.00 and +1.00
r
2
, square of the correlation coefficient
A measure of the percentage of variability in the values of y that
is explained by the independent variable
Ranges between 0 and 1.00

!
r
2
=
n xy
"
( )
# x
"
( )
y
"
( )
n x
2
"
( )
# x
"
( )
2
n y
2
"
( )
# y
"
( )
2
$
%
&
&
&
'
(
)
)
)
3-38
Simple Linear Regression Assumptions
1. Variations around the line are random
2. Devaiations around the average value (the
line) should be normally distributed
3. Predictions are made only within the range of
observed values
3-39
Issues to consider:
Always plot the line to verify that a linear
relationships is appropriate
The data may be time-dependent.
If they are
use analysis of time series
use time as an independent variable in a multiple
regression analysis
A small correlation may indicate that other
variables are important
3-40
Using Forecast Information
Reactive approach
View forecasts as probable future demand
React to meet that demand
Proactive approach
Seeks to actively influence demand
Advertising
Pricing
Product/service modifications
Generally requires either and explanatory model or a subjective
assessment of the influence on demand