2. General considerations
Demand forecasting is very important for
smooth/efficient functioning of any org. and
even an economy as a whole.
This forecasting becomes more important in
advanced countries where demand conditions
are more uncertain than supply conditions and
where mostly S>D.
As competition begins to intensify, demand
forecasting assumes significance.
3. General considerations continued
Six factors involved in demand forecasting:
How far ahead: short-run ( unto one year), and
long run( up to 5, 10 or 20 years).
Short run can also mean operating within the
limits of already available resources and long run
means extending or reducing the limits of
resources.
4. General considerations continued
Three levels of demand forecasting:
i) Macro level (economy as a whole),
ii) micro level (industry) and,
iii) firm level- imp. far managerial level.
General vs. specific forecast : firms need specific
product/area-wise forecasts.
Problems and methods for new and old products vary
as sales trends and competition characteristics are not
available for new products.
5. General considerations contd.
Distinctive patters of demand: Important to classify
goods into categories like producers goods, consumers
goods, services etc.
Finally, special factors peculiar to each product and
market must be taken into account. This involves
psychological/ sociological considerations of people
about the product and its future. It is the basis of
branding.
6. Methods of forecasting:
1. Opinion survey
No easy method or simple formula
Opinion survey or Survey of buyers intensions
usually for a year ahead. It is a passive method.
It may turn out to be biased as respondents may
not give realistic and rational responses.
It is quite useful when bulk of the sales is made
to industrial producers.
7. 2. Delphi method
A variant of opinion poll.
Attempts to involve large number of experts
Questions them repeatedly till a consensus is arrived at among
participating experts.
The identities of different experts, especially holding contrary
views are not revealed to avoid “halo effect” till there is
consensus.
Originally developed at Rand Corporation by Olaf Helmer,
Dalkey and Gordopn in late 1940’s.
Used successfully, especially for technological forecasting.
But it assumes panelists to be rich in experience/ knowledge and
objective in their analysis.
8. 3. Hunch method or Expert opinion
Involves field experts like dealers, distributors
and suppliers, officers of trade associations as
also industry analysts, special marketing
consultants etc.
Collects their assessments and arrives at
forecasting by applying varied statistical
methods of analysis.
A simple and quick method.
9. 4. Collective opinion/ sales-force
polling
Salesmen required to estimate expected sales in their
respective territories/sections
These estimates are reviewed to avoid biases of
optimism and pessimism of salesmen
The revised estimates further examined in the light of
factors like proposed changes in prices, product
designs, advertisement programs, expected changes in
competition, changes in purchasing power, income
distribution etc.
Simple and based on first hand information.
But subjective and relevant to short periods
10. 5. Naive models
Based on historical observations of sales
Ignores casual relationships of variables
Consider Y as actual sale value and Y’ as forecast
t t+1
value
Three models: _
i) Y’t+1 = Yt , ii) Y’t+1=Yt +(Yt-Yt-1), iii) Y’t+1= Yt x (Yt / Yt-1).
Consider the data below:
12. 6. Smoothing techniques
These techniques are a higher form of naïve models. Its typical
forms are: a) moving averages and b) Exponential smoothing.
Moving average are updated as new information is received.
Exponential smoothing is popular for short run forecasting. It
uses weighted average of past data as basis for forecast. Heavier
weights are accorded to more recent information. It is effective
when there is randomness and no seasonal fluctuations in data.
The formula for exponential smoothing:
Y’t+1= αYt +(1- α)Yt -1
13. 7. Analysis of time series and trend
projections
Firms, industry, and economy data available for
some years are used in this analysis to forecast
demand.
A number of statistical tools are available for
this analysis.
The trend projections are also arrived at by
analyzing the data with the help of statistical and
graphic methods
14. 8. Use of economic indicators
Construction contracts sanctioned for building
materials, say cement
Personal income for demand of consumer goods
Agricultural income for the demand of
agricultural inputs, implements, fertilizers etc.
Automobile registration for car accessories/
petrol demand
15. 9- 10. Controlled experiments and
judgmental approach
Controlled experiments are undertaken by
varying some variable while holding others
constant. This method uses a host of statistical
methods, especially the regression analysis
Judgmental approach means using judgment to
choose the method and tools for demand
forecasting as per the specific product case
16. Engle’s Law of Consumption
Dr. Engle was a German statistician.
He made a study of family budgets around the middle
of the nineteenth century
He arrived at the following major conclusions:
i) As income increases the percentage expenditure on
food decreases and vice versa
ii) The percentage expenditure on clothing, etc. remains
more or less constant at all levels of income
17. Engle’s law………
iii) The percentage expenditure on fuel, light, rent,
etc. also remains practically the same at all levels
of income.
iv) However, the percentage expenditure on what
may be called comforts and luxuries of life
increases with increase in income and vice versa.