Chapter Three: Forecasting

3.1. Demand management

3.2. Demand forecasting
3.3. Characteristics of forecasting
3.4. Principles of forecasting
3.5. Collection and preparation of data
3.6. Forecasting techniques
 Demand Management
Demand management is the function of recognizing
and managing all demands for products and/or
services. It occurs in the short, medium, and long
term.
In the long term, demand projections are needed for
strategic business planning of such things as
facilities.
In the medium term, the purpose of demand
management is to project aggregate demand for
production planning.
In the short term, demand management is needed for
items and is associated with master production
scheduling. This text is most concerned with the
latter.
 Cont
If material and capacity resources are to be
planned effectively, all sources of demand
must be identified.
These include domestic and foreign
customers, other plants in the same
corporation, branch warehouses, service parts
and requirements, promotions, distribution
inventory, and consigned inventory in
customers' locations.
Demand management includes several major
activities, all of which are primarily market
driven:
 Cont

Identifying all product and service demand in the

defined markets. This includes forecasting but also
involves possible segmenting of markets, classifying
customers, and identifying demand that does not add
value and therefore should be ignored. Includes
identifying customer desires for existing or possible
new product, or service design and features.
Identifying and understanding all aspects of the
market that will potentially impact customer demand.
This includes economic conditions and indicators,
governmental laws and regulations, and sources of
existing or potential competition, including possible
new competitors.
 Cont

synchronizing identified market demand with

company capabilities.
Setting priorities for demand when supply will
not cover all demand.
Making delivery promises.
Interfacing between manufacturing planning
and control and the marketplace.
Order processing.
 Cont

There are several other activities or components to

demand management. They include the following:
Setting and maintaining appropriate customer service
levels.
Planning for new product introductions and phase-out
of obsolete inventory.
Planning and managing interplant shipments and
distribution requirements planning
Establishing inventory target levels and maintaining
them.
Establishing performance metrics for demand and
using them to evaluate performance.
 Demand Forecasting
Demand shows market or customer requests. Demand has
the following characteristics.
1. Demand Patterns: A pattern is the general shape of a
time series. Although some individual data points will not
fall exactly on the pattern, they tend to cluster around it.
The pattern shows that actual demand varies from period
to period. There are four reasons for this: trend,
seasonality, random variation, and cycle.
A. Trend: The trend can be level, having no change from
period to period, or it can rise or fall.
B. Seasonality: It is usually thought of as occurring on a
yearly basis, but it can also occur on a weekly or even
daily basis.
 Cont

C. Random variation: it occurs where many factors affect

demand during specific periods and occur on a random
basis. The variation may be small, with actual demand
falling close to the pattern, or it may be large, with the
points widely scattered.
D. Cycle: Over a span of several years and even decades,
wavelike increases and decreases in the economy influence
demand.
2. Stable Versus Dynamic
The shapes of the demand patterns for some products or
services change over time, whereas others do not. Those
that retain the same general shape are called stable and
those that do not are called dynamic.
Dynamic changes can affect the trend, seasonality, or
randomness of the actual demand. The more stable the
demand, the easier it is to forecast.
 Dependent Versus Independent Demand
Demand for a product or service is independent when it
is not related to the demand for any other product or
service, or independent of internal activities of the firm.
Dependent demand for a product or service occurs
where the demand for the item is derived from that of a
second item.
Requirements for dependent demand items need not be
forecast but are calculated from that of the independent
demand item.
Only independent demand items need to be forecasted.
These are usually end items or finished goods but
should also include service parts and items supplied to
other plants in the same company (intercompany
transfers).
 Definitions of Forecasting
A forecast is a statement about the future value of a
variable such as demand. That is, forecasts are
predictions about the future. The better those
predictions, the more informed decisions can be.
Forecasts are a basic input in the decision processes
of operations management because they provide
information on future demand.
Forecasting is the process of estimating the
occurrence, timing or magnitude of future event.
Forecasts are statements about future specifying the
volume of sale to be achieved, material demand
required, equipments and other inputs needed to
meet the sales.
 Cont

They give operation managers a rational basis for

budgeting, capacity planning, sales, production and
inventory, personnel and material management.
Forecasting is the basis of planning ahead even though
the actual demand is quite uncertain thus, it involves
estimation of the future, and of particular interest here
is the expected demand of company’s product.
Therefore, forecast of future demand is the link
between company’s internal expectations with outside
environment that permits planning function to
commence activities.
A popular definition of forecasting is that it is
estimating the future demand of a product, service and
the resources necessary to produce an output.
 Characteristics of forecasting
The following are the characteristics of forecasts:-
Forecasting techniques generally assumes that the
same underling causal system that assisted in the
past will continue to exist in the future.
Forecasts are rarely perfect; actual results usually
differ from predicted values.
Forecasts for a group of items tends to be more
accurate than forecasts for individual item, because
forecasting errors among items in a group usually
are smaller than that of individual items.
Forecast accuracy decreases as the time period
covered by the forecast-time horizon increases.
 Principles of Forecasting
Forecasts have four major characteristics or principles. An
understanding of these will allow the more effective use
of forecasts. They are simple and, to some extent,
common sense.
1. Forecasts are usually wrong. Forecasts attempt to look
into the unknown future and, except by sheer luck, will be
wrong to some degree. Errors are inevitable and must be
expected.
2. Every forecast should include an estimate of error.
Since forecasts are expected to be wrong, the real question
is “by how much?” Every forecast should include an
estimate of error often expressed as a percentage (plus and
minus) of the forecast or as a range between maximum
and minimum values. Estimates of this error can be made
statistically by studying the variability of demand about
the average demand.
 Cont
3. Forecasts are more accurate for families or groups.
The behavior of individual items in a group is random
even when the group has very stable characteristics. For
example, the marks for individual students in a class
are more difficult to forecast accurately than the class
average. High marks average out with low marks. This
means that forecasts are more accurate for large groups
of items than for individual items in a group.
4. Forecasts are more accurate for nearer time periods.
The near future holds less uncertainty than the far
future. Most people are more confident in forecasting
what they will be doing over the next week than a year
from now. As someone once said, tomorrow is
expected to be pretty much like today.
 Collection and Preparation of Data

Forecasts are usually based on historical data manipulated

in some way using either judgment or a statistical
technique. Thus, the forecast is only as good as the data
on which it is based. To get good data, three principles of
data collection are important.
1. Record data in the same terms as needed for the
forecast. There is often a problem in determining the
purpose of the forecast and what is to be forecast. There
are three dimensions to this:
a. If the purpose is to forecast demand on production, data
based on demand, not shipments, are needed. Shipments
show when goods were shipped and not necessarily when
the customer wanted them. Thus, shipments do not
necessarily give a true indication of demand.
 Cont

b. The forecast period, in weeks, months, or

quarters, should be the same as the schedule period.
If schedules are weekly, the forecast should be for
the same time interval.
c. The items forecasted should be the same as those
controlled by manufacturing.
2. Record the circumstances relating to the data.
Demand is influenced by particular events, and
these should be recorded along with the demand
data
3. Record the demand separately for different
customer groups. Many firms distribute their goods
through different channels of distribution, each
having its own demand characteristics.
 Forecasting Techniques
There are many forecasting methods, but they
can usually be classified into categories.
Forecasting techniques generally may be
qualitative or quantitative.
Qualitative techniques are projections based
on judgment, intuition, and informed
opinions. By their nature, they are subjective.
Such techniques are used to forecast general
business trends and the potential demand for
large families of products over an extended
period of time.
 Cont

As such, they are used mainly by senior management.

Production and inventory forecasting is usually
concerned with the demand for particular end items,
and qualitative techniques are seldom appropriate.
When attempting to forecast the demand for a new
product, there is no history on which to base a
forecast.
Quantitative techniques are projections based on
historical or numerical data, whether it be from inside
or outside the organization.
 Qualitative Techniques
Four of the better-known qualitative forecasting
methods are: executive opinions, the Delphi method,
sales-force polling, and consumer surveys:
1. Executive Opinions
The subjective views of executives or experts from
sales, production, finance, purchasing, and
administration are averaged to generate a forecast about
future sales. Usually this method is used in conjunction
with some quantitative method, such as trend
extrapolation. The management team modifies the
resulting forecast, based on their expectations.
 Cont
Advantage
The forecasting is done quickly and easily without need
of elaborate statistics.
executive opinions may be the only means of
forecasting feasible in the absence of adequate data
Decision is fast
Bring together the considerable knowledge,
experiences, skills and talents of various mangers.
Disadvantages
Probably poor forecast due to lack of experience
Domination by one or few manager
Diffusing responsibility for the forecast over the entire
group may result in less pressure to produce a good
forecast.
 Cont
2. Sales force opinion
Some companies use as a forecast source
salespeople who have continual contacts with
customers. They believe that the salespeople who
are closest to the ultimate customers may have
significant insights regarding the state of the future
market.
Forecasts based on sales force polling may be
averaged to develop a future forecast. Or they may
be used to modify other quantitative and/or
qualitative forecasts that have been generated
internally in the company.
 Cont

Advantages
It is simple to use and understand.
Can see from different approaches
It uses the specialized knowledge of those closest to the
action.
It can place responsibility for attaining the forecast in
the hands of those who most affect the actual results.
The information can be broken down easily by
territory, product, customer, or salesperson.
Disadvantages
Time taking decision
Avoidance of responsibility
Influenced by majority high watch persons
 Cont

3. Consumer Survey
Some companies conduct their own market surveys
regarding specific consumer purchases.
Surveys may consist of telephone contacts, personal
interviews, or questionnaires as a means of obtaining data.
Extensive statistical analysis usually is applied to survey
results in order to test hypotheses regarding consumer
behavior.
Advantages
Tap information that may not be available else where
Enhance the quality and accuracy of forecasts
Disadvantages
Experience and knowledge is constructing
Expensive and time consuming
 Cont
4. Delphi Method
This is a group technique in which a panel of
experts is questioned individually about their
perceptions of future events.
The experts do not meet as a group; in order to reduce
the possibility that consensus is reached because of
dominant personality factors. Instead, the forecasts and
accompanying arguments are summarized by an
outside party and returned to the experts along with
further questions. This continues until a consensus is
reached.
This type of method is useful and quite effective for
long-range forecasting. The technique is done by
questionnaire format and eliminates the disadvantages
of group think.
 Cont
Advantages
There is no committee or debate.
The experts are not influenced by peer pressure to
forecast a certain way, as the answer is not intended
to be reached by consensus or unanimity.
Disadvantage
Low reliability is cited as the main, as well as lack
of consensus from the returns
It take time to take censuses
Coordination and interpretation is difficult
 Quantitative Techniques

Quantitative methods come in two main types:

Time-series methods and explanatory methods.
Time-series methods make forecasts based purely on
historical patterns in the data. Time-series methods
only use historical site visit data to make that forecast.
Explanatory methods use other data as inputs into the
forecasting data. Time-series methods are probably the
simplest methods to deploy and can be quite accurate,
particularly over the short term.
Most quantitative forecasting methods try to explain
patterns in historical data as a means of using those
patterns to forecast future patterns.
 Time series Analysis
There are four types of time series analysis. They are:
Simple moving average
Weighted moving average
Simple exponential smoothing
Trend equation
 Cont
A. Simple moving average is obtained by summing and
averaging values from a given number of periods
repetitively, each time deleting the oldest value and
adding the new value.
SMA= Ft= A1 +A2+------An
N
Summation of A divided by N.
Where
SMA= Simple moving average
Ft= Forecast fore period t
A= Actual demand in period t
N = Number of periods or data points in the moving
average
 Cont
Simple moving average is preferable if the demand for
the product is neither growing nor declining rapidly and
also does not have any seasonal characteristics .
Example: A food processor uses a moving average to
forecast next month’s demand. Past actual demand is
shown in the following table.
Month 1 2 3 4 5 6 7 8

Actual 105 106 110 110 114 121 130 128

Demand
 Cont
Required:
A) Compute a simple five month moving average to

forecast demand for month 9

B) Compute a simple six month moving average to

forecast demand for month 9

C) Compute a simple five month moving average to

forecast demand for month 10 if the actual demand

for month 9 is 123
D) Compute a simple five month moving average to

forecast demand for month 11 if the actual demand

for month 10 is 130
 Cont

Solution:
A) 128+ 130+121+114+110 = 603 = 120.6
5 5
B) 128+ 130+121+114+110+110 = 713 =118.83
6 6
C) 123+ 128+ 130+121+114 = 616 = 123.2
5 5
D) Therefore the five month moving average forecast
demand for month 10 is 123.2
In moving average as each new actual value becomes
available, the forecast is updated by adding the newest
value and dropping the oldest value and computing the
average.
Consequently the forecast move by reflecting only the
most recent values.
 Cont
Advantages
Easy of computation
Easy of understanding
Disadvantages
All values in the average are weighted
equally. The oldest value has the same weight
as the most recent value.
Example 2: Compute the three month moving
average forecast given demand for shipping carts
for the last five Demand
Period period
1 40
2 44
3 36
4 42
5 40

Solution: MA3= 36+42+40 = 39.33 units

3 3
 Cont
If the actual demand in period six turns out to be
41, the moving average forecast for period seven
would be:
Solution: MA3= 42+40+41/3 = 41
Exercise 1. Shoa Bakery makes a cakes and supply
to the market. The following data shows their daily
demand for the last four weeks.
The bakery is closed on Saturday. So Friday’s
production must satisfy demand for both Saturday
and Sunday.
Days 4 weeks ago 3 weeks ago 2 weeks Last weeks
Monday 200 400 300 400
Tuesday 200 100 100 300
Wednesday 300 400 300 500
Thursday 800 900 400 100
Friday 500 200 300 400
Saturday & Sunday 800 700 700 500
Total 2800 2000 2000 2200

Required: make forecast for this week on the following basis

 Cont
A) Daily, using a simple four week moving average?
B) Daily, using a simple three week moving average?
C) Weekly, using a simple four week moving average?
D) Weekly, using a simple three week moving average?
Solution
A) Monday: 200+400+300+400/4= 325
B) Monday: 400+300+400/3= 367 (rounded)
C) Weekly= 2200+2000+2700+2800/4=2425
D) Weekly = 2200+2000+2700/3=2300
 Cont

B. Weighted Moving Average

In weighted moving average, the weight is given in
such away that more weight is given to the most
resent value in the time series. Weights can be
percentages or any real numbers.
In weighted moving average, forecasts are
calculated by
Ft= WMA=W1A1+W2A2+---WnAn
Where
Ft= forecast in time t
WAM= Weighted moving average
W=weight
A=actual demand
 Cont

Example: A department store may find that in four

month period the best forecast is derived by using 40%
of actual demand for the most recent month, 30% of
actual demand for the two months ago, 20% of three
month ago and 10% for four month ago. The actual
demands were as follows:
Month Month 1 Month 2 Month 3 Month 4

Demand 100 90 105 95

 Cont
Required;
A) Compute Weighted 4-month MA for month
five
WAM=95x0.4+105x0.3+90x0.2+100x0.1= 97.5
units
B) Suppose that demand for month five actually
turned out to be 100. Compute forecast for
month six.
WAM=0.4x100+0.3x95+0.2x105+0.1x90=F6
=102.5 units
 Exercise 1

Bishoftu floor factory sells Macaroni and its monthly sales

for a seven months period were as follows:
Month Feb Mar Apr May June Jul Aug

Sales(in ‘00 units) 19 18 15 20 18 22 20

Required: Forecast September sales in units using three

months weighted moving average with 0.60 for Aug, 0.3
for July and 0.1 for June
 Cont

C) Simple exponential smoothing

With this the forecast is made up of last period forecast
plus a portion of the difference b/n the last period
actual demand and the last period forecast.
Mathematically: Ft= Ft-1 +α(At-1 –Ft-1)
Where:
Ft= Forecast for period t
Ft-1 =forecast for the previous period
α=Smoothing constant (0< α<1)
At-1= Actual demand for the previous period
 Cont

The difference between the actual demand and the

previous forecast (i.e. At-1-Ft-1) represents the
forecast error.
The smoothing constant α usually dictates how
much corrections will be made. It is a number b/n 0
and 1 and it is used to compute the forecast Ft.
Exponential smoothing is the most widely used of
all forecasting techniques, b/c :
Exponential forecasting model provide closer
forecasts to actual demand
Formulating an exponential smoothing model is
relatively easy.
The user can easily understand the model
 Cont
It requires little computation
It requires only three pieces of data
The most recent forecast
The actual demand of the previous period
The smoothing constant, α
1. The production manager Example
of Oxford company uses a
simple exponential smoothing technique (α=0.2) to
forecast demand. In May, the forecast was for 20
shipments and the actual demand was for 20 shipments.
The actual in June and July was 25 and 26 shipments.
Forecast the value for August.
Solution: First forecast the demand for June and July
Fjune= Fmay + α (Amay-Fmay)
= 20+ α(20-20)=20
FJul= 20+ α (Ajun-Fjun)
Fjul=20+0.2(25-20) =21
FAug= Fjul+0.2(Ajul-Fjul) = 21+0.2(26-21)=22,
therefore, the forecast for August is 22 shipments
 Cont

2. Demands during the past six months have been as

follows:
Month Demand

January 115

Feb 123

March 132

April 134

May 140

June 147
 Cont

Required: using simple exponential smoothing with

α=0.70, if the forecast for January had been 110, compute
exponentially smoothed forecasts for each month through
July.
Solution:
Ffeb= Fjan+0.7(Ajan-Fjan)= 110+0.7(115-
110)=110+0.7x5=113.5
Fmar=Ffeb+0.7(123-113.5)= 113.5+.6.65= 20.15
FApr= Fmar+0.7(Amar-Fmar)=120.15+0.7(132-120.15)
=120.15+0.7x11.85=128.445
FMay=FApr+0.7(AApr-FApr)= 128.445+ 0.7(134-
128.445)= 128.445+3. 8885=132.33
Fjun=Fmay+0.7(Amay140-FMay132.33)=138 (rounded)
Fjul=FJun+0.7(Ajun-Fjun)=138+0.7(147-138)=144.3
 Cont
D. Trend Projections
In this method, a trend line is fitted to the given time
series data and then projections are made into the future
by using this line. The trend line may be linear or
curvilinear in nature.
In-order to obtain the trend line, the historical data are
plotted on the graph, representing time scales on X-
axis.
Then a line is drowning through these points in such a
way that
i. The sum of deviations above & below the line are
equal.
ii. The sum of the squares of these vertical deviations is
minimum.
 Cont

In essence, the trend line is drown based on the

principle of least square. Such a line is represented by
y = bx +a
where
y = is the trend line to be predicted
b = the slope of the trend line
a = the y intercept
x =independent variable
b= ∑ xiyi − n( x)( y)
2 2

a=
∑ xi − n( x)
y − b(x)
 Example
Consider the following demand pattern of ABC
Company for iron ore.
Year Demand for iron sheet in ‘000 tons

1989 13
1990 17
1991 16
1992 16
1993 21
1994 20
1995 20
1996 23
1997 25
1998 24
1999 25
 Cont
Required:
A. Forecast the demand of the iron sheet for the year
2000 for ABC Company. Using 1984 as base year.
B. Forecast the demand of the iron sheet for the year
2000 for ABC Company. Using 1997 as base year.
Solution:
Here by using the square method we can develop the
estimating (regression) equation.
By using the last square method, we can develop the
line of the best fit.
y = bx +a
 Cont
The line of the best fit always passes through the
two points
Now in order to develop the equation of the line
we need to have a base year to see the deviation
of others from it. Most of the time the base year
is the middle observation.
In the case of even observations, we use the mean
of the two middle observations as base year by
multiplying the mean by any number. for
simplicity in calculation.
When we use 1994 as a base year.
Year Xi Yi xiyi
xi2
1989 -5 13 -65 25
1990 -4 17 -68 16
1991 -3 16 -48 19
1992 -2 16 -32 4
1993 -1 21 -21 1
1994 0 20 0 0
1995 1 20 20 1
1996 2 23 46 4
1997 3 25 75 9
1998 4 24 96 16
1999 5 25 125 25

∑xi = 0 ∑yi = 220 ∑xiyi = 128 ∑ xi2 = 110

The equation of the estimating line, when the base year is 1994 ⇒
x = 0. The value of b & a for trend line can be obtained as
follows:
b= ∑ xiyi − n( x)( y)
2 2
xi − n( x)
∑–11(0)
b = 128 (20) = 128 =1.1636
110 – 11(0)2 110
a=
y − b(x)
20 – 1.1636(0)
= 20
(y = bx + a)

The trend linear equation will be; y =

1.1636x + 20
 Cont
Therefore, the forecast for the year 2000 will be as
follows;
x =6 (because it is six years after the base year 1994
y = 1.1636(6) + 20
y = 26.9816 tons of iron sheet
B) if we take 1997 as a base
year
Year Xi Yi xiyi
xi2
1989 -8 13 -104 64
1990 -7 17 -119 49
1991 -6 16 -96 36
1992 -5 16 -80 25
1993 -4 21 -84 16
1994 -3 20 -60 9
1995 -2 20 -40 4
1996 -1 23 -23 2
1997 0 25 0 0
1998 1 24 24 1
1999 2 25 50 4

∑xi=-33 ∑yi = 220 ∑xiyi = -532 ∑ xi2 = 209

 Cont
y = ∑yi 220 =
20
n 11
b = -532 –(11)(-3)(20) = 128 = 1.1636
209 – 11(-3)2 110

= 20-1.1636(-3) = 23.4908
y − b( x )
 Cont
Then the equation will be
y = 1.1636x+23.4908
the forecast for the year 2000 will be;
x =3 (because it is 3 yrs. after the base year -
1997)
y =1.1636(3) +23.4908
26.9816 tons of iron sheet.
 Cont
N.B. The forecast for the year 2000 will be the
same even if the base year is changed i.e.
irrespective of the base year the forecast will
be the same.
Whenever there is a change in the base year,
there is no need of calculating the slope of the
equation.
We need to calculate the y – intercept only &
we can use the slope calculated in the
previous base year as it is.
 Example
y = 20 + 1.1636x when the middle year is
used as a base year, when the base year is
changed by adding /subtracting from the value
of x i.e. from 1994 to 1997 there is a 3 yrs.
time gap.
So the value of x will be x +3. The trend
equation by using 1997 as a base year is
y = 20 +1.1636 (x+3)
y = 20 +3.4908+1.1636x
y = 23.4908 +1.1636x
Explanatory/ Causal Models of Forecasting
The causal model considers two types of variables the dependent and
independent variables.
Casual methods are used when historical data are available and there
is relationship between the factors to be forecasted.
Example
Sales of Samsung TV might be related to
The company’s advertising budget, the company’s prices,
competitor’s prices and promotional strategies, and even the nation’s
economy and unemployment rates.
Real estate prices are usually related to
Property location -Square footage
Crop yield are related to
Soil conditions - Amounts and timings of water - Fertilizer
application
 Types of Casual Methods of Forecasting
Regression and correlation techniques are means of
describing the association between two or more
variables.
Regression can be defined as a functional relationship
between two or more correlated variables. It is
concerned about the first two issues, i.e. bringing out
the nature of relationship between any two variables
and measuring the rate of change in one (the
dependent) variable associated with a given change in
the other (independent) variable.
Regression means ‘dependence’ and involves
estimating the value of a dependent variable, Y, from
an independent variable X.
 Cont

Correlation: is concerned about evaluating the

strength of the relationship and quantifying the
closeness of such relationship.
Simple Linear regression and correlation: In
simple linear regression, only one independent
variable is used and the model takes the form
Y = a + bx
Where
Y = predicted (dependent) variable, demand
a = value of Y at X = 0
b = slope of the line