Demand Forecasting Methods

Introduction
o Business world is characterized by risk and uncertainty. One way to reduce it is to acquire knowledge
about the future demand prospects for the product.
o The information regarding the future demand for the product is obtained by demand forecasting.
o Demand Forecasting is predicting the future demand for firm’s product.
o Demand forecasting assumes greater significance where large scale production is involved.
o Uncertainty will not be a problem for small firm with small fraction of supply of the total demand.
 Steps in Demand Forecasting
• Specifying the objective – Short term or long term, demand for industry's product or firms
product, whole or only a segment of the market or firms market share.
•Determining the time perspective – Short period or long period.
•Making choice of method for demand forecasting – Statistical and Non – Statistical methods
•Collection of data and data adjustment – Primary or secondary data.
•Estimation and interpretation of results.
Methods of Demand Forecasting
Survey Methods
• Consumer Survey – direct interview
• Opinion Poll Methods
Statistical Methods
• Trend Projection
• Barometric Method
• Econometric Method
 Survey Methods
• Survey methods are generally used when the purpose is to make short run forecast of
demand for a product.
• Information is collected through a survey on consumers/users
• Survey are conducted by two methods:
◦ Consumer survey method
◦ Opinion poll method.
• The choice of any of these methods depends on the availability of primary data and
time and money that firms are willing to spend on the survey.
Consumer Survey Method
• Consumers are directly interviewed to collect the required data on their
demand for the product.
• Depending on the time and cost consumer survey methods are conducted by
three methods
• Complete Enumeration Method
• Sample survey
• End – use survey method
Complete enumeration Method
• Complete enumeration method is used when market size is small and all consumers can
to be contacted by the surveyors.
• The quantities indicated by the consumers are added together to obtain the probable
demand for the product.
• For example: Suppose majority of households in a city report the quantity(q), they are
willing to purchase of a commodity. Then, the total demand can be calculated as,
Dp= q1 +q2 + q3+…………+ qn , q1 , q2, q3 denote demand of individuals
Limitations :
• Consumers may not have a exact idea of their demand
• Answers might be hypothetical, not real
• Consumers response may be biased
• Their plans may change with a change in the factors not mentioned in the questionnaire
Sample Survey Method
•Sample survey method is used when population is large and expanded over a large area.
•From the population, only a sample of potential consumers will be selected for interview.
•Consumers are selected through sampling methods.
•Method of survey may be direct interview or mailed questionnaire.
Dp = HR/HS (H * AD )
Dp = Probable demand forecast, HS = census no. of households surveyed, HR = no. of
households reporting demand for the product, AD = average expected consumption by
the households.
Advantage : It is very simple, less costly, and less time consuming.
The end – use method
The end use method of demand forecasting consists of four distinct stages of estimation.
First stage – Identify and list all the possible users of the product in question.
Difficulty arises because data on the end users are rarely available.
If the data's are not available, managers need to have a thorough knowledge of the product and
its uses.
Second Stage – Fixing suitable technical ‘ norms’ of consumption of the product
Norms are expressed in physical terms, either per unit of production of the complete product, or per unit
of investment
Third Stage – Application of the norms.
Fourth Stage – Aggregate the product wise or use wise content of the item for which the demand
is to be forecast
Advantages – Future demand can be estimated at different types and sizes, It is possible to trace
pinpoint at any time in future as to where and why the actual consumption has deviated from the
estimated demand.
Opinion Poll Methods
The opinion poll methods aims at collecting opinions of those who are
supposed to process knowledge of the market like sales representatives,
sales executives, marketing experts etc…
• Expert opinion method
• Delphi method
• Market studies and experiments
Expert Opinion Method
• Sales representatives of the firms will assess the demand for a targeted product in the areas,
regions or cities they represent.
• Sales representatives being in touch with the customers will be able to know their future
purchase plans.
• The estimates of demand thus obtained are added up to get the overall probable demand for
the product.
Limitations:
• Estimates provided by the representatives are reliable only to an extend depending on their
skill and expertise.
• Demand estimation may involve a subjecting judgment, which may lead to over or under
estimation.
• Inadequate information as the sales representative have only narrow view of the market.
Delphi Method
• It is an extension of expert opinion method.
• Under Delphi method, the task of projecting demand for a product id assigned anonymously to a group of
market experts with a group leader.
• Each market experts makes their own demand projection individually and submit the result to the group leader.
• The group leader will interchange the reports among the experts. The groups leader seeks suggestion and
modifications made by the other experts. The suggestions made are again interchanged among the experts.
• The group will find a compromise estimate, which will be the final estimate of demand forecast
Advantages:
• Provides opportunity to experts to forecast the demand systematically.
• Overcomes the disadvantages of face to face discussion by the panel of experts.
• It offers a fairly reliable demand forecast.
Market Studies and Experiments
• Method of collecting necessary information regarding current and also future demand for a product is to carry
out market studies and experiments on consumers behavior under actual, though controlled market conditions.
• Under this method, firm first select some areas of representative markets.
• Second, they carry out experiments by changing prices, advertisement expenditure and other controllable
variables. The variables may be changed over time.
• After the changes are made, the consequent changes in the demand over a period of time is recorded.
• Market experiments can be replaced by consumer clinics or controllable laboratory experiments.
Limitations:
• Expensive
• Experimental methods are based on short term and controllable conditions which may not exist
• Changes in socio economics conditions may invalidate the results.
• Change in price may cause a permanent loss of customers to competitive brands.
Statistical Methods
• Statistical method uses historical ( Time – Series Data) and cross sectional data for
estimating long term demand.
• Method of estimation is based on theoretical relationship between the dependent and
independent variable.
• Estimates are relatively more reliable
• Smaller time and money cost
• Kinds of statistical methods used for demand projection include:
• Trend Projection
• Barometric Method
• Econometric Method
 Trend Projection Methods
• Trend projection method is the study of movement of variables through time. This method
requires a long and reliable time series data.
• Old firms obtain time – series data on sales from their own sales department and books of
account. New firms can obtain the necessary data from the older firms belonging to the same
industry.
• Techniques of trend projection based on time series data include:
• Graphical Method
• Least square method
• Box Jenkins method
Graphical Method
• Under this method, an annual sales data is plotted on a graph paper and a line is draw through the plotted
points.
• Free hand line is drawn through the mid values of variations, and another line drawn, which is a straight
trend line.
• This method is simple and inexpensive.
Trend Fitting Equation : Least Square Method
• Under this method, a trend line is fitted to the time series sales data with the aid of statistical techniques.
• It can be either by plotting the sales data or by trying different forms of trend equations for the best fit.
•Most common type of trend equations are : Linear Trend and Exponential Trend
Linear Trend
When the time series data reveals rising trend in sales, then a straight line trend equation of the following form
is fitted.
S = a + bT
S= Annual sales, T = time (in years) a and b are constants. B measure the annual increase in sales.

Exponential Trend
When total sales have increased over the past years at an increasing rate or at a constant percentage rate per
unit time.
Limitations :
• Does not bring out the measure of relationship between dependent and independent variables.
• Method cannot be used for short term estimates
• This method can be considered reliable only during the time period during which the past rate of change in
the dependent variables will persist in the future to assumption holds.
Box Jenkins Method
• This method is used only for short terms projections and predictions. Suitable for forecasting demand with only
stationary time series data.
• Stationary time series data is one that does not reveal a long term trend.
• According to this method, any stationary time series data can be analysed by the following three models.
• Auto regression Model
• Moving Average Model
• Autoregressive moving average model.
• Steps in Box Jenkins Approach
• First step: Eliminate trend from the time series data
• Second step : Make sure that there is seasonality in the stationary time series
• Final step: Use the models to predict sales in the intended period.
Auto regression Model
The behavior of a variable in a period is linked to the behavior of the variable in future periods
Yt = a1Yt-1 + a2Yt-2 + …… anYt-n + et
Moving Average Model
Estimates Yt in relation to the residuals (et) of the previous years
Yt = m + b1 et-1 + b2 et-2 + ………. + bp et-p + et
Autoregressive Moving Average Model
Moving Average Model combines with the Auto regression model to form the final of Box - Jenkins
model, also called Autoregressive moving average model.

Yt = a1Yt-1 + a2 Yt-2 + …… anYt-n + b1 et-1 + b2 et-2 + ………. + bp et-p + et

Box Jenkins method is a sophisticated and highly complicated method.
Barometric Methods
Load lag indicators
• The basis approach is to construct an index of relevant economic indicators and to forecast
the trends on the basis of movements in the index of economic indicators.
The lead lag indicators include:
• Leading Indicators – indicators which moves up or down ahead of some other series
• Coincidental Indicators – Indicators which moves up or down simultaneously with the level of
general economic activities.
• Lagging Indicators – Indicators that follow a change after some time lag
• Indicators are chosen on the basis of the following criteria:
• Economic Significance, statistical adequacy, conformity with overall movement in
economic activities, immediate availability of the series, smoothness of the series.
Diffusion index
• Diffusion Index is the percentage of rising indicators.
• A diffusion index copes with the problem of differing signals given by the indicators
• It is obtained by the ratio of the number of indicators, in a particular class, moving up or down
to the total number of indicators in that group.
• If three out of six indicators in the lagging series are moving up, the index shall be 50 %.
•However, there are problems of identifying the leading indicator for the variables under study.
•Limitations:
• It can be used only short term forecasting
• Leading indicator of the variable to be forecast is not always easily available.
Econometric Method
• The forecast made through econometric method is much more reliable than those made
through any other method.
• The econometric model may be a single equation regression model or it may consists of
simultaneous equations.
• The econometric models are briefly described under two basic methods:
• Regression method
• Simultaneous equation model.
Regression Method
• This method combines economic theory and statistical techniques of estimation. Economic theory is employed
to specify the determinants of demand and to determine the nature of relationship between demand for a product
and its determinants.
• Under this method, one needs to estimate the demand function for a product. Demand is a dependent variable,
all the other variable which determine demand are independent variables or explanatory variables.
• Demand for commodities sometimes depends on a single independent variable. In that case, the demand
function for those commodities are Single Variable demand functions.
• Demand for commodities depends on a number of independent variables. Such demand functions are called
Multi Variable demand function
•For Single Variable demand function Single Regression equation is used and for Multi Variable demand function,
Multi – variate Regression equation is used.
Simultaneous Equation Model
• Simultaneous equation enables the forecaster to take into account the simultaneous interaction between
dependent and independent variables.
• This method is a complete and systematic approach for forecasting.
• The first step in this model is to develop a complete model and specify the behavioral assumptions regarding the
variables in the model.
• The variables that are included in the model are called Endogenous and Exogenous variable.
• Endogenous variables are those determined within the model. These are included in the model as dependent
variable. They are also called controlled variables.
•Exogenous variables are those determined outside the model. They are also called uncontrolled variables.
• The second step is to collect the necessary data on both endogenous and exogenous variables.
•The third step is to estimate the model through some appropriate method.
•Finally, the model is solved for each endogenous variable in terms of exogenous variable.