1.0 INTRODUCTION
This note introduces you to forecasting in production and
operations management (POM). Planning is
an integral part of a manager’s job, and if
uncertainties becloud the planning horizon, managers
will find it
difficult to plan effectively. Forecasts
help managers by reducing some of the
uncertainties, thereby enabling them to develop more meaningful plans.
In a nutshell, a forecast is statement
about the future.
2.0 OBJECTIVES
By the end of this note, you should be able to:
(i) Describe at least four qualitative forecasting techniques and
the advantages and limitations of
each.
(ii) Compare and contrast qualitative and quantitative approaches
to forecasting.
(iii) Identify the five basic demand patterns that combine to
produce some series.
(iv) Choose an appropriate forecasting technique for a given
decision problem.
3.0 MAIN CONTENT
3.1 Introduction to Forecasting in POM.
Customer demand is the backbone of all enterprises. Occasionally,
however, customers appear unexpectedly,
without prior notice. This sudden situation
very often throws organisations off balance to the extent that the
quality of their products, response time
and customer service are badly affected. But this shouldn’t be allowed to happen. A
well-managed enterprise will make efforts
to forecast demand, which normally allows it to be reasonably prepared
when the demand actually occurs. Broadly
speaking, well-managed businesses strive
to manage demand, and this normally includes:
• Planning for demand
• Recognizing and accounting for all sources of demand
• Pre-processing of demand.
From the foregoing therefore, it is important that organisations
have effective approaches to
forecasting. In addition, forecasting should be an integral part of their business planning. Figure 3.1 is an
illustration that forecasting is an
integral part of business planning. The figure shows that the major
inputs from various market conditions,
economic outlook and other factors such as legal, political, sociological and cultural forces
are processed through forecasting models
or methods to develop demand estimates. You must however note that these demand estimates are not the sales
forecasts. They are just the starting point for management teams to develop sales
forecast. The sales forecasts in turn
become inputs to both business strategy and production resource forecasts. Actually, when managers plan, they are merely
trying to determine in the present, what
causes of action they will take in the future. The first step in planning is therefore forecasting or better
still, estimating the future demand for
products and services and the resources necessary to produce these
outputs. Estimates of the future demand
for products are usually referred to as sales
forecasts. These are the starting point for all the other forecasts in
POM. Can you now guess why forecasting
is so essential to POM? Anyway, let us look at
this together: Operations managers need long-range forecasts to make
strategic decisions about products,
process, and facilities. They will also need shortrange forecasts to assist them in making decisions
about product issues that span only the
next few weeks.
Figure 3.1
Forecasting as an Integral Part of Business Planning INPUTS
3.2 Importance of Forecasting in POM.
Some of the reasons why forecasting is very essential in POM are
given below:
1. New Facility Planning: It usually takes as long as five years
to design and build a new
factory or design and implement a new production process. Such strategic activities in POM
require long-range forecasts of demand
for existing and new products so that operations managers can have the necessary lead time to build
factories and install processes to
produce the products and services when needed.
2. Production Planning: Usually, demands from products continue
to vary from month to month and from one
season to the other. Hence production
rates need to be scaled up or down to meet these varying demands. We should also note that it can take
some months to alter the capacities of
production processes. Therefore, operations managers need medium-range forecasts so that they can have
the lead time necessary to provide the
production capacity to produce there variable monthly demands.
3. Work Force Scheduling: Demands for products and services
may actually vary from week to week. In
order to remain on an efficient or
profitable level of operation, the work force must, out of necessity
be scaled up or down to meet these
demands by using various methods, such
as reassignment, overtime, layoffs, or hiring. In this regards, operations managers need short-range forecasts
so that they can have the lead time
necessary to provide work force changes for the production of weekly demands.
3.3 An Overview of Demand Measurement:
We need to realize right from here, that demand management is a
shared responsibility. Usually, a master
planning team, composed of experts in
Marketing, Finance and Operations is responsible for taking care of,
and coordinating demand management
activities. This team has at least three
important roles to play. These are to:
* Account for all sources of demand: historical demand patterns,
sales force estimates, actual orders and
direct selling, within – company (i.e.
division-to-division) demands, and economic influences.
* Influence demand, e.g. through special promotions
* Evaluate the impact of any demand management plan on capacity
and cash flow.
Time Horizon in Forecasting
From our previous discussion, you will have observed that
forecasts can be made over any time
horizon. However, the shorter the period being considered, the more accurate is the forecasts, since one
is more certain of the variables involved. Descriptions of forecast elements
over three time horizon include
short-range, medium-range and long-range. Each time frame is discussed below with examples of some of the things
usually forecasted:
3.4.1 Short Range
A short-range forecast is one for a time span of a few weeks, up
to say about three months. It would
include forecasting such items as:
• purchase transactions;
• cash requirements;
• work scheduling;
• workforce levels;
• job assignments; and
• production levels.
3.4.2 Medium Range
The medium-range forecast covers between about three months and up
to one year. Items usually included here
are:
• capacity plan;
• operating cash budgets;
• production plans;
• sales plans; and
• Subcontractor needs.
3.4.3 Long Range A long-range forecast
usually spans a year up to about five years, and would include:
• new investments;
• capital expansion plans;
• facility location;
• new product development
• strategic plans;
• acquisition;
• implementing new technology; and
• research and development programmes.
3.5 Importance of Sales Forecast
What we have been stressing all along is that an estimate of
demand, typically in the form of a sales
forecast, is critical to the successful functioning of most businesses. It is one of the most important
pieces of data used by management and
takes a central stage in most companies’ planning efforts. Its importance spreads across the following areas: as shown
in Table 3.1.
Table 3.1: Different Areas
of Application of Sales forecasts within an
organization 
3.6 Sales
Forecasting Methods. There are two
main classes of forecasting methods: Qualitative (or subjective) and Quantitative (or objective). The
qualitative or subjective methods rely primarily
on judgment to produce sales forecasts. The quantitative or objective methods, in contrast, involve the application
of statistical techniques of varying
degrees of sophistication. The different techniques under each main
class are shown in Figure 3.2. We will
consider these methods at some length in the
sections that follow.
Figure 3.2: Classification of Sales Forecasting Methods
. 
3.5.1 Subjective or
Qualitative Methods
The subjective methods are based on assumptions, or intuitive estimates
of those in the firm that are familiar
with the market. This may include sales
personnel, purchasing representatives or management people who all have close contact with customers. Some of there
techniques may involve several levels of
sophistication. An example here is an opinion survey that has been scientifically conducted. Others are merely
intuitive hunches about future events.
The accuracy of a particular subjective approach depends on the good judgment, honesty and philosophy of the individuals
concerned. We shall attempt to examine
each of the subjective techniques indicated in Figure 3.2
3.5.1.1 Users’ Expectations
The users’ expectations method is also known as the buyers’
intentions methods since it relies on
responses from customers with regard to their
expected consumption or purchase of a product. The customers may be surveyed in person, over the telephone, or by
mail. In some particular situations, the
respondents in a users’ expectations survey do not necessarily have to be the ultimate consumers. Rather,
the firm may find it advantageous to
secure the reactions of wholesalers and retailers that serve the
channel.
Advantages
The users’ expectations method offers several advantages. These
include the following:
(i) The forecast is based on estimates obtained directly from
firms whose buying actions will actually
determine the sales of the product.
(ii) The way through which the information was obtained i.e.
projected product use by customers,
allows preparation of forecasts in great detail
e.g. by product, by customer, or by sales territory.
(iii) The method may often provide some insight into the buyer’s
thinking and plan. Therefore, it could
be helpful in planning the marketing
strategy.
(iv) It is particularly useful to solicit opinions from
prospective buyers about a new product
that is just coming to the market.
Disadvantages of Users’ expectations are as enumerated below:
(i) The method is limited to situations in which the potential
customers for the product are few and
well defined. The method could be difficult to
adopt and can actually result in grave errors when there are many customers that cannot be easily
identified.
(ii) The method also depends on the sophistication of the potential customers in appreciating their needs. Here,
we should remember that buyer intentions
are subject to change, thus the method does not work particularly well for consumer goods.
(iii) It is often difficult to determine the firmness of
intentions to purchase, particularly
when the person being interviewed is not literate or uncooperative.
(iv) The method requires a considerable expenditure of money, time
and manpower.
3.6.1.2 Sales Force Composite
The sales force composite is a specific judgmental forecast for
which opinions are solicited from line
sales personnel and sales managers. Each person states how much he or she expects to sell during the
forecast period. The usual technique is
to ask sales people to forecast sales for their districts and have these estimates reviewed by the regional
sales manager and then by the head
office sales manager. This method is based on the belief that those
closest to the sales people have the
best knowledge of the market.
Advantages
(i) A primary advantage of the sales force composite method is
that it uses the specialised knowledge
of the people closest to the market.
(ii) It has also been argued that the size of the sample used to
develop the forecast tends to produce
estimates that are fairly accurate.
(iii) The method lends itself to the easy development of customer;
product, territory, or sales force
breakdowns. These are particularly important in
controlling the sales effort.
Disadvantages
(i) Sales representatives are often seen to be notoriously poor
estimates. For instance, they tend to be
overly optimistic when the economy is
booming and overly pessimistic when things are not so good.
(ii) Salesmen usually are not trained forecasters and are
ill-informed on the factors influencing
sale.
(iii) The approach makes no provision for bringing the
systematic consideration of
uncontrollables into the analysis.
(iv) The approach does not provide for discovery of important
facts through statistical analysis of
historical data.
3.6.1.3 Jury of Executive Opinion
The jury of executive opinion method is about the oldest and
simplest method of making sales
forecast. The method either formally or informally polls the top executives of the company for their
assessment of sales possibilities. The
separate assessments are then combined into a sales forecast for the
company. This is sometimes often done by
simply averaging the individual judgments.
Disparate views are resolved through group discussions. In some cases,
the process amounts to little more than
group guessing. In other cases however, it
involves the careful judgment of experienced executives who have studied
the underlying factors influencing their
company’s sales.
Advantages
(i) Ease and quickness with which it can be made.
(ii) Does not require elaborate statistics.
(iii) The method brings together a variety of specialised
viewpoints. The resulting “collective
wisdom” reflects the thinking of the top people in the company.
(iv) When there is an absence of adequate data or experience, such
as with innovative products, the jury of
executive opinion method may be the only
means of sales forecasting available to the company.
Disadvantages
(i) The forecasts are based on opinions rather than on facts and
analysis.
(ii) Averaging opinions reduces responsibility for accurate
forecasting.
(iii) The method is expensive because of the large amounts of
highly paid executives’ time it
consumes.
(iv) The forecast may not properly weight the expertise of those
most informed.
3.6.1.4 Delphi Technique
This method is used to achieve consensus within a committee. The
Delphi technique uses repeated
measurements and controlled feedback instead of
direct confrontation and debate among the experts preparing the
forecast. The way this method is
employed is illustrated by Figure 3.3. The following steps are involved. First, each individual prepares
a forecast using whatever facts, figures
and general knowledge of the environment he or she has at his or her disposal. Second, the forecasts made are
collected, and the person supervising
the process prepares an anonymous summary. Third, the summary is distributed to each person who participated
in the initial phase. Usually, the
summary indicates each forecast figure, the average and some other
summary measure of the spread of the
estimates. Those whose initial estimates fell
outside the mid range of responses are asked to express their reasons
for these extreme positions. The
explanations offered are then incorporated into the summary. Those participating in the exercise
are asked to study the summary and
submit a revised forecast. The process is then repeated.
3.2: Operation of Delphi Process
(a) The range of responses will decrease, and the estimates will
converge with repeated measurements
b) The total group response or median will move successively
toward the “correct” or “true” answer.
Advantages
(i) The strategy of forcing those whose forecasts lie at the ends
of the distribution to justify their
estimates seems to have benefits in that
“informed” experts have greater opportunity to influence the final forecast.
(ii) Those who might have a deviant opinion, but with good reason,
can defend that position, rather than
going in to group pressure.
(iii) The method can result in forecasts that most participants
have ultimately agreed to in spite of
their initial disagreement.
Disadvantages
(i) The process of iteration and feedback in the Delphi often
takes a long time
(ii) The method can also be very expensive.
3.6.2 Objective of Quantitative Methods
As we have already noted, the objectives or quantitative methods
of forecasting are statistical in
nature. They range in complexity from relatively simple trend extrapolations to the use of sophisticated
mathematical models. A lot of
organisations are tending toward the use of advanced methods in which
the computer correlates a host of relationships.
Let us now go into the treatment of the
quantitative techniques earlier shown in Figure 3.
3.6.2.1 Market Test
Market testing is a relatively recent phenomenon in demand
estimation and is mostly used to assess
the demand for new products. The essential feature of a market test is that it is a controlled
experiment, done in a limited but carefully
selected part of the marketplace, whose aim is to predict the sales or
profit consequences, either in absolute
or relative terms, of one or more proposed
marketing actions. It therefore goes beyond estimating the potential
sales of a new product.
It is necessary for us to note that market testing methods differ
in the testing of consumer and
industrial products. For instance, when testing consumer products, the company wants to estimate the
major determinants of sales, such as
trial, first repeat, adoption, and purchase frequency. The major methods
of consumer goods market testing include
sales-wave research, simulates store
technique, controlled test marketing and test markets.
However, we are not going into their details here. You will learn
more about them under Marketing
Research. Test marketing is not
typically used in the case of industrial products. For instance, it will be too expensive to produce
a sample of airplanes; ships etc, let
alone put them up for sale in a select market to see how well they will
sell. Marketing research firms have
actually not built the test-market systems that
are found in consumer markets.
Therefore, goods industrial manufacturers have to resort to other methods to research the
market’s interest in a new industrial
product. The most common method adopted is product-use test. A
second common market test is to
introduce the new industrial product at trade shows. A new industrial product can also be tested
in a distributor and dealer display
rooms. The details of these methods are under Marketing Research.
Advantages
(i) Market testing can indicate the product’s performance under
actual operating conditions.
(ii) It can also show the key buying influences and the best
market segment
(iii) It provides ultimate test of consumers’ reactions to the
product
(iv) It allows the assessment of the effectiveness of the total
marketing programme
(v) It is very useful for new and innovative products.
Disadvantages
(i) It allows competitors know what the firm is doing; hence
they may jam the experiment by creating
artificial situations so that the
results of the test may not be meaningful.
(ii) It invites competitive reaction
(iii) It is expensive and time consuming.
(iv) Often takes a long time to accurately assess level of initial
and repeat demand.
3.6.2.2 Time Series
This approach to sales forecasting rely on the analysis of
historical data to develop a prediction
for the future. The depth and sophistication of these analyses often vary widely. At one extreme,
the forecaster might just forecast next
year’s sales to be equal this year’s sales. This forecast might be
reasonably accurate for a mature
industry that is experiencing little growth.
However, if there is some
growth, the forecaster might allow for it by predicting the same percentage increase for next year that the
company experience this year. Still
further along the continuum, the forecaster might attempts to break
historical sales into basic components
by isolating that portion due to trend, cyclical, seasonal and irregular influences.
The first component, trend (T), is the result of basic
developments in population, capital
formation, and technology. It is found by fitting a straight or curved line through pass sales. The second
component, cycle (C), captures the
wavelike movement of sales. Many sales are affected by swings in general economic activity, which tends to be
somewhat periodic. The cyclical
component can be useful in medium-range forecasting. The third
component, season (S), refers to a
consistent pattern of sales movement within the year.
The term season, describes any recurrent hourly, weekly, monthly,
or quarterly sales pattern. The seasonal
component may be related to weather factors,
holidays, and trade customs. The seasonal pattern provides a norm
for forecasting short-range sales. The
fourth component, erratic events (E),
includes strikes, blizzards, fads, riots, fires, war scares, and other
disturbances.
These erratic components are by definition unpredictable, and
should be removed from past data to see
the more normal behaviour of sales. Time
series analysis consists of decomposing the original sales series, Y, into the components, T, C, S, and E. Then these
components are recombined to produce the
sales forecast. The following is an example.
A company sold 12,000 notes of its main product this year. It now
wants to predict next year’s December
sales. The long-term trend shows a 5% sales
growth rate per year. This alone suggests sales next year of 12,600.
(i.e. 12,000 x 1.05). However, a
business recession is expected next year and will probably result in total sales achieving only 90% of
the expected trend-adjusted sales. Therefore,
sales next year will more likely be 11,340 (i.e. 12, 600 x 0.90). If sales were the same each year, monthly sales
would be 945 (i.e. 11,340 – 12).
However, December is an above-average month for that particular
product, with a seasonal index of 1.30.
Therefore, December sales may be as high as
1,228.5 (i.e. 945 x 1.30). No erratic events such as strikes or new
product regulations are. Therefore, the
best estimate of new product sales next
December is 1,228.5.
A newer time-series
technique called exponential smoothing is now available. This is being used by a firm with hundreds of
items in its product line, and wants to
produce efficient and economical short-run forecasts. In its simplest form, exponential smoothing requires only
three pieces of information: this
period’s actual sales, Qt; this periods smoothed sales, Q t; and a
smoothing parameter, a . The sales
forecast for next period’s sales is then given by: Qt + 1 = a Qt + (1 - a ) Q t
Where:
Q t + 1 = sales
forecast for next period
a = the smoothing constant, where 0 = a = 1
Qt = current sales in period t
Q t = smoothed sales in period t.
Example:
Suppose the smoothing constant is 0.3, current sales are N600,
000, and smoothed sales are N500,
000.
Then sales forecast is:
Q t + 1 = 0.3 (N600, 000) + 0.7 (N500, 000)
= N180, 000 + N350, 000
= N530, 000.
You will observe that the sales forecast is always between (or at
an extreme of) current sales and
smoothed sales. Another technique under
time series analysis is the method of moving
averages. This is conceptually simple. Let us consider the forecast that
next year’s sales will be equal to this
year’s sales. Such a forecast might be subject
to large error, if there is a great deal of fluctuation in sales from
one year to the next. To allow for such
randomness, we might want to consider making use of some kind of recent values. For example, we
might average the last two years sales,
the last three years’ sales, etc. The forecast would simply be the average that resulted. The term moving average is
used because a new average can be
computed and used as a forecast as each new observation becomes
available. Table 3.2 presents 15 years
of historical data for a manufacturer of shirts, together with the resulting forecast for a
number of years using two-year and
four-year moving averages.
Table
3.2: Annual and Forecasted sales for a manufacturer of shirts.
As earlier explained, the
calculation of moving averages is relatively
simple. For instance, the
entry 4305 for 1976 under the two-year moving average method, for example, is the average of the
sales of 4,200 notes in 1974 and 4,410
notes in 1975. In the same vein, the forecasts of 5520 notes in 1989 represent the average of the number of notes
sold in 1987 and 1988. You may attempt
to verify other forecast in the table.
Advantages
(i) The time series approach to sales forecasting provides a
systematic means for making quantitative
projections of sales.
(ii) The method is objective in the sense that two analysts
working on the same data series using
the same forecasting technique and the same
model should produce the same forecast.
Disadvantages
(i) It is not useful for new or innovative products
(ii) Factors for trend, cyclical, seasonal, or product life-cycle
phase must be accurately assessed and
included (iii) Technical skill and good
judgement required.
(iv) Final forecast may be difficult to break down into individual
territory estimates.
3.6.2.3 Statistical Demand Analysis.
Statistical demand analysis is a set of statistical procedures
designed to discover the most important
real factors affecting sales and their relative
influence. The factors most commonly analysed are price, income,
population and promotion.
The method consists of expressing sales (Q) as a dependent
variable and trying to explain sales as
a function of a number of independent demand variables X1, X2, …, Xn; that is: Q = f (X1, X2, …, Xn) By making use of multiple regression analysis,
various equation forms can be
statistically fitted to the data in the search for the best predicting
factors and equation.
Let us make use of the work of Palda (1964), who tried to measure
cumulate advertising effects of a
vegetable product. He found that the following demand equation gave a fairly good fit to the
historical sales of the product in question
between the years 1908 and 1960:
Y = - 3649 + 0.665X1 + 1180 log X2 + 774 X3 + 32X4 – 2.83X5
Where:
Y = Yearly sales in thousands of dollars
X1 = yearly sales (lagged one year) in thousands of dollars
X2 = yearly advertising expenditures in thousands of dollars
X3 = a dummy variable, taking on the value 1 between 1908 and 1925
and 0 from 1926 on
X4 = year (1908 = 0, 1909 = 1, etc)
X5 = disposable personal income in billions of current
dollars.
It was found that all the five independent variables accounted for
94% of the yearly variation in the sale
of the commodity under investigation between 1908 and 1960. How can we use this demand equation
as a sales forecasting equation for the
five independent variables? It follows thus:
Sales in 1960 should be put in X1;
The log of the company’s planned expenditures for 1961 should be
put in X2;
0 should be put in X3;
The numbered year corresponding to 1961 should be put in X4;
and
Estimated 1961 disposable personnel income should be put in
X5.
The result of multiplying these numbers by the respective
coefficients and summing them gives a
sales forecast (Y) for 1961.
Advantages
(i) It has great intuitive appeal
(ii) Requires quantification of assumptions underlying the
estimates. This makes it easier for
management to check the results
(iii) It provides a means of discovering factors affecting sales
which intuitive reasoning may not
uncover.
(iv) The method is objective in the results can be reproduced by
different analysts using the same model
and variables.
Disadvantages
(i) It presumes that historical relationships will continue into
the future, hence the analysts may have
a false sense of security in this regard.
(ii) It requires technical skill and expertise
(iii) Some managers are reluctant to use the method due to its
sophistication.
4.0 CONCLUSION
In this note, you
have learned that planning is an integral part of a manager’s job. If uncertainties cloud the planning
horizon, managers will find it difficult
to plan effectively. Forecasts help managers by reducing some of
the uncertainties, thereby enabling them
to develop more meaningful plans.
5.0 SUMMARY
Forecasts are vital inputs for the design and the operation of the
productive systems because they help
managers to anticipate the future. Forecasting
techniques are generally classified as qualitative or quantitative. Qualitative techniques rely on judgement, experience, and
expertise to formulate forecasts;
quantitative techniques rely on the use of historical data, or
associations among variables to develop
forecasts. Some of the techniques are simple, while others are complex. Some work better than others,
but no technique works all the
time.
0 comments:
Post a Comment