DEMAND FORECASTING

What is Forecasting

Demand forecasting is the business process that

attempts to estimate sales and the use of products so that
they can be purchased, stocked, or manufactured in
appropriate quantities in advance to support the firm's
value adding activities.
A forecast is an objective estimate of future demand
attained by projecting a pattern of events of the past into
the future. (David Ross)
A prediction is a subjective estimate of what events will
happen in the future, based on extrapolating or interpreting
data that occurred in the past.
 Why Demand Forecasting

In general practice, accurate demand forecasts lead to

efficient operations and high levels of customer service, while
inaccurate forecasts will inevitably lead to inefficient, high cost
operations and/or poor levels of customer service.
Forecasts of future demand will determine the quantities
that should be purchased, produced, and shipped.
Forecasting helps firm to achieve fast to order cycle time.
Forecast demand also affords the firm opportunities to
control costs through leveling its production quantities,
rationalizing its transportation, and generally planning for
efficient logistics operations.
 Features of Forecast

 Assume the same underlying causal system.

 Forecasts are rarely perfect.

 Forecasts for groups of items are more accurate than

forecasts for individual items.

 Forecast accuracy decreases as the time period covered

by the forecast increases.
Elements of a Good Forecast

Timely

Reliable Accurate

e
f ul u s
n g Written to
ni s y
ea Ea
M
 Components of Forecast
Time Frame
Demand Behavior
 Trend
 A gradual, long-term up or down movement of
demand.
 Cycle
 An up-and-down repetitive movement in demand.
 Seasonal pattern
 An up-and-down repetitive movement in demand
occurring periodically.
 Random variations
 Movements in demand that do not follow a pattern.
Different Types of Demand Behavior
 Approaches to Forecasting

Qualitative
Use management judgment, expertise, and opinion to
predict future demand.

Time series
Statistical techniques that use historical demand data to
predict future demand.

Associative Techniques
Attempt to develop a mathematical relationship between
demand and factors that cause its behavior.
 Qualitative Methods

 Executive Opinions

 Sales force Opinions

 Consumer Survey

 Delphi Method
 Time Series Forecasting

Assume that what has occurred in the past will continue to

occur in the future.

Relate the forecast to only one factor – time.

Include
Naïve method
Moving average
Weighted moving average
Exponential smoothing
Trend Adjusted Moving Average
Linear Trend Model
 Naive Method

A naive forecast uses a single previous value of a time

series as the basis of a forecast.

The naive approach can be used with a stable series

(variations around an average), with seasonal variations, or
with trend.

It has virtually no cost, it is quick and easy to prepare

because data analysis is nonexistent, and it is easily
understandable.

The main objection to this method is its inability to provide

highly accurate forecasts
 Simple Moving Average
 The moving average method uses average demand for a
fixed sequence of periods.

 This method is used with stable demand with no

pronounced behavioral patterns.
n
∑ At-i
i=1
Ft = MAn = -------------
n

Ft = Forecast for time period t

MAn = n period moving average
At−i = Actual value in period t − i
n = Number of periods (data points) in the moving average
Longer-period moving averages react more slowly
 Weighted Moving Average

Place relatively more emphasis on recent data and

relatively less emphasis on less current data.
 Simple Exponential Smoothing

 Averaging method.
 Weights most recent data more strongly.
 Reacts more to recent changes.
 Widely used, accurate method.

Ft =Ft-1 + At-1 – Ft-1)

Where,

Ft = Forecast for period t

Ft−1 = Forecast for the previous period
α = Smoothing constant
At−1 = Actual demand or sales for the previous period
Effect of Smoothing Constant a
Values of a are restricted such that 0 << 1.
The choice of  is up to the analyst. In this form,  can be
interpreted as the relative weight given to the most recent data
in the series.
If = 0, then Ft = Ft-1
Forecast does not reflect recent data

If = 1, then Ft = At-1

Forecast based only on most recent data

For example, if an  of 0.2 is used, each successive forecast

consists of 20% "new" data (the most recent observation) and
80% "old" data, since the prior forecast is composed of
recursively weighted combinations of prior observations.
 Example#1
The chairperson of the Management Department at UMass
Lowell wants to forecast the number of students who will
enroll in the Operations Management course next semester in
order to determine how many sections to schedule, instructors
to allocate, textbooks to order, etc. The chair has compiled the
following enrollment data for the past eight semesters:
a. Compute a 3-semester simple moving average forecast for
semesters 4 through 8.
b. Compute a 3-semester weighted moving average for
semesters 4 through 8 using the weights 0.5, 0.3, and 0.2, with
the heavier weights on the more recent semesters.
c. Use exponential smoothing with parameter  = 0.2 and to
compute the forecast for semesters 2 through 8. Take F1 = 400
as the forecast for the first semester.
 Student Weighted Exponential
Semester Enrolled Moving Avg Moving Avg. Smoothing

1 400 400 400 400

400+.2(400-400)=
2 450 450 450 400.00
400+.2(450-400)=
3 350 350 350 410.00
(400+450+350)/3= (.2×400+.3×450+ 410+.2(350-410)=
4 420 400.00 .5×350)=390.00 398.00
(450+350+420)/3= (.2×450+.3×350+ 398+.2(420-398)=
5 500 406.67 .5×420)=405.00 402.40
(350+420+500)/3= (.2×350+.3×420+ 402.4+.2(500-
6 575 423.33 .5×500)=446.00 402.4)= 421.92
(420+500+575)/ (.2×420+.3×500+ 421.92+.2(575-
7 490 3=498.33 .5×575)=521.50 421.92)= 452.54
(420+500+575)/ (.2×500+.3×575+ 452.54+.2(490-
8 650 3=521.67 .5×490)=517.50 452.54)= 460.03
 Linear Trend Model
Analysis of trend involves developing an equation that will
suitably describe trend.
The trend component may be linear, or it may not.
A simple plot of the data can often reveal the existence and
nature of a trend.

Ft = a + bt
where,
Ft = 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
The coefficients of the line, a and b, can be computed from historical
data using these two equations:

b  ty  n t y
 t  n(t )
2 2

a  y  bt
where,
n = Number of periods
y = Value of the time series
 Example 2
Cell phone sales for a California-based firm over the last
10 weeks are shown in the following table.

a. Plot the data, and visually check to see if a linear trend line
would be appropriate.
b. Determine the equation of the trend line, and predict sales
for weeks 11 and 12.
 Week(t) Sales(y) t2 y2 ty
1 700 1 490000 700
2 724 4 524176 1448
3 720 9 518400 2160
4 728 16 529984 2912
5 740 25 547600 3700
6 742 36 550564 4452
7 758 49 574564 5306
8 750 64 562500 6000
9 770 81 592900 6930
10 775 100 600625 7750
Sum 55 7407 385 5491313 41358
Mean 5.5 740.7

41358 − 10 × 5.5 × 740.7

𝑏= = 7.51 𝑎 = 740.7 − 7.51 × 5.5 = 699.40
385 − 10 × ሺ5.5ሻ2

Ft= 699.40+7.51t
Associative Technique: Linear Regression Model

 Linear regression model develop a relationship between

demand and factors that cause its behavior.

y = a + bx a = intercept (at period 0)

b = slope of the line
x = Independent
variable
y = Dependent
variable

Σ xy – n(x)(y)
b= a = y - bx
Σ x² - n(x)2
r2

Correlation

 Coefficient of correlation, r
 Measure of strength of relationship.
 Varies between -1.00 and +1.00.

n xy   x  y
r
[n x 2  ( x) 2 ][n y 2  ( y ) 2 ]

 Coefficient of determination, r2
 Percentage of variation in dependent variable
resulting from changes in the independent variable.
 Example 3
Dan Ireland, the student body president at Toledo State
University, is concerned about the cost to students of
textbooks. He believes there is a relationship between the
number of pages in the text and the selling price of the book.
To provide insight into the problem he selects a sample of
eight textbooks currently on sale in the bookstore. Draw a
scatter diagram.
a.Determine the correlation between the two variables. Does
it appear that a relationship between these variables will
yield good predictions? Explain.
b. Obtain a linear regression line for the data.
c. Estimate the price of a book with 800 pages.
Book Page
Price($)
Introduction to History 500 84
Basic Algebra 700 75
Introduction to Psychology 800 99
Introduction to Sociology 600 72
Business Management 400 69
Introduction to Biology 500 81
Fundamentals of Jazz 600 63
Principles of Nursing 800 93
 Scatter Diagram of Number of Pages and Selling Price of Text

100

90
Price ($)

80

70

60
400 500 600 700 800
Page
Page (X) Price (Y) X2 Y2 XY
500 84 250000 7056 42000
700 75 490000 5625 52500
800 99 640000 9801 79200
600 72 360000 5184 43200
400 69 160000 4761 27600
500 81 250000 6561 40500
600 63 360000 3969 37800
800 93 640000 8649 74400
∑XY=39720
∑X=4900 ∑Y=636 ∑X2=3150000 ∑Y2=51606 0
Mean Mean
X=612.5 Y=79.5

8  397200  4900  636

r  0.61
8  3150000  4900 8  51606  636 
2 2
 397200  8  612.5  636
b  0.051
315000  8  612.5
2

a  79.5  0.051 612.5  48.26

The regression equation is,

Y(Price) = 48.26 + 0.051X(No. of pages)

The price of a 800 page book is,

Price = 48.26 + 0.051×800 =$89.06

 Forecast Accuracy

 D t  Ft
MAPD t 
 D F
t t
100
MAD t 
n D t

where
t = period number
Dt = demand in period t
Ft = forecast for period t
n = total number of periods
 Example 4

Compare the four forecasts of Example 1 using MAD and

MAPD and indicate which forecast is the most accurate.
Forecast Forecast
Student Forecast
Weighted Exponential ABS ABS ABS
Enrolled Moving
Moving Smoothing Dev(MA) Dev(WMA) Dev(ES)
in OM Average
Average
400 400 400 400 0.00 0.00 0.00
450 450 450 400.00 0.00 0.00 50.00
350 350 350 410.00 0.00 0.00 60.00
420 400.00 390.00 398.00 20.00 30.00 22.00
500 406.67 405.00 402.40 93.33 95.00 97.60
575 423.33 446.00 421.92 151.67 129.00 153.08
490 498.33 521.50 452.54 8.33 31.50 37.46
650 521.67 517.50 460.03 128.33 132.50 189.97
3835 401.67 418.00 610.12
MAD(MA)=401.67/5=80.21

MAD(WMA)=418/5=83.6

MAD(ES)=610.12/7=87.16

MAPD(MA) = (401.67/3835) x100 = 10.47%

MAPD(WMA) = (418/3835) x100 = 10.90%

MAPD(ES) = (610.12/3835) x100 = 15.91%