Demand/Sales Forecasting in Indian Firms

Abhijeet Kamble (05329031)

Ajay Nalawade (05329007)
Ashutosh Deo (05329003)
Sagar Ranadive (05329006)
April 3, 2006

Abstract
Forecasting is a science of predicting the future by analyzing the
previous data. The science of forecasting helps firms to take strategic
decisions to achieve their goals. Therefore, there is a need for precision
in the demand forecasts.
This has led to development of various new tools and methods for forecasting in the last two decades. This report discusses the forecasting
methods used for prediciting demand/sales of Indian firms. We also
present a case study on two-wheeler Indian automotive industry.

Introduction

Forecasting implies predicting the future after studying and analysing the
past and present data. Forecasting is of special importance in industries,
where important strategic decisions need to be taken to optimize the efficiency of the resources and hence maximize the profits.

Sales Forecasting

The forecasting of sales is one of the most important information tools for
every management. In a company lot of units use the sales forecast for example top management, finance, production, human resources, purchasing
and marketing units. Top management unit allocates resources among functional areas and to control operations inside and outside of the company by
using the sales forecast. The companys finance unit uses the sales forecasting to decide on capital appropriation, to project cash flows, and to establish
operating budgets. Production uses it to decide how much the company has
to produce and in what time and to control inventories. Human resource
units use the sales forecasting (to plan personnel requirements and also as
an input in collective bargaining). Purchasing unit uses it to plan how much
1

materials the company needs and in which part of the year/month/week or

even day.
Marketing units of firms find sales forecasting very useful (to plan marketing
and sales programs and to allocate resources among the various marketing
activities). The meaning of sales forecasting grows as firms start to coordinate their own bussines in a world wide scale.
The bussiness world knows two types of sales forecasting methods and they
are subjective methods and objective methods, all of them have advantages
and disadvantages. Usually decision makers use one or another forecasting method and their decision will depend on technical knowledge, previous
sales data, and for what exactly the sales forecasting will be used.

Methods of sales forecasting

3.1

Qualitative Methods

Qualitative Forecasts consider the range of factors which influence the demand. These factors are then ranked in order of importance and each of
them in turn is analyzed to reveal future trends.
Qualitative methods of forecasting are: consumer expectations, sales force
composite, jury of executive opinion and Delphi technique.
3.1.1

Consumer Expectations

Consumers are frequently interviewed with the help of questionnaires concerning their buying habits, motives and intentions. The consumer feedback
is used to estimate the expected consumption or purchases of the product.
This method has following advantages:
Forecast estimates straight from buyers.
Information about projected product can be very detailed.
Insights give assistance in planning the market strategy.
Practical for forecasting new-products.
Disadvantages of this sales forecasting method are:
Company has to choose potential customers carefully and the number
of customers has to be small.
Works well with business to business goods, but not with consumer
goods.
Depends on how precisely the users make their evaluations.
Takes a lot of money, time and labour.
2

3.1.2

Delphi technique

If a more accurate forecast is wanted from a group of experts and the effects
of group dynamics has to be controlled, then a good technique to use is Delphi technique. Delphi is based on iterative approach and it uses anonymous
repeated feedback. The people involved in the feedback give their own forecast about the subject and the feedbacks are gathered into a summary. The
forecast is made with the available knowledge the person has. The summary
lists the median of each forecast figure. Also a summary of the speard of
estimates is put into the summary. If someones answer differs a lot from the
median, then the person gives an explanation why she or he made such and
estimate. The explanation is added into the summary and the summary is
given to all the participants of the forecast. They study the summary and
make a new forecast. These steps are repeated several times in the idea that
repeated measurements will lessen the range of estimates and they will come
together to a correct or true answer.
The advantages of Delphi are:
The effects of group dynamics is reduced
Statistical information can be used
The disadvantages are:
Can take a long time and is money consuming

3.2

Objective methods

Objective methods are those that use well-specified processes to analyze

the data. Ideally, they have been specified so well that other researchers
can replicate them and obtain the same forecasts. These have also been
called explicit, statistical or formal methods. These methods usually rely
on quantitative analytical approaches in developing the forecast.
3.2.1

Market test

Company does the market test by placing a product in a few representative

geographic areas to understand how well it performs and then projecting
that experience to the whole market. Usually companies do this test for a
new or improved version of an old product.
Advantages of this forecasting method are:
The method gives the best understanding of consumers reaction to
the product.
The method permits evaluation of the success of the total marketing
program.
3

 Has a good use for new and innovative products.

The disadvantages are:
Competitors of the company can easily find out what is going on.
Competitors start to do something about their products.
Market tests takes lots of money and time.
Often takes a long time to measure the exact level of initial and repeat
demand.
3.2.2

Time-series analysis

A company can use different time-series analysis methods for forecasting

sales. Time-series analysis makes a forecast for the future sales on the base
of historical data. The difficulty and style of these analyses can differ extensively. The most easy assessment might use last years figures as the next
years forecast, but if the company is new or is a growing company, then
this easy method cannot be used.
3.2.3

Moving average

Moving average is a quite simple method of sales forecasting, it considers the

previous years sales to be the forecast for the next year. Using this method
can lead to large errors if there is much variation in sales from one year to
the next. By using some kind of average of recent values, it is possible to not
care about the randomness of sales from one year to the other. For example,
company might average the last two years sales, the last three years sales, or
any number of other periods. The forecast would simply be the average that
resulted. The number of observations included in the average is typically
determined by trial and error. Differing numbers of periods are tried, and
number of periods that produce the most accurate forecasts of the trial data
is used to develop the forecast model. Once determined, it remains constant.
3.2.4

Exponential smoothing

Exponential smoothing is a type of moving average. However, instead of

weighing all observations equally in generating the forecast, exponential
smoothing weighs the most recent observations heaviest, for good reason.
The most recent observations contain the most information about what is
likely to happen in the future, and they should logically be given more
weight.

3.2.5

Decomposition method

The Decomposition method of sales forecasting is typically applied to monthly

or quarterly data where seasonal pettern is evident and the manager wishes
to forecast sales not only for the year but also for each period in the year. It
is important to determine what portion of sales changes represents an overall, fundamental change in demand and what portion is due to seasonality
in demand. The decomposition method attempts to isolate four separate
portions of a time series: the trend, cyclical, seasonal and random factors.
In using the decomposition method, the analyst typically first determines
the seasonal pattern and removes its impact to identify the trend. Then the
cyclical factor is estimated. After the three components are isolated, the
forecast is developed by applying each factor in turn to the historical data.
The advantages are:
Utilizes historical data.
Objective, inexpencive.
Disadvantages are:
Not useful for new or innovative products.
Factors for trend, cyclical, seasonal, or product life-cycle phase must
be accurately assessed and included.
Technical skill and good judgement required.
Final forecast difficult to break down into individual territory estimates.
Ignores planned marketing effort.
3.2.6

Panel Regression

A regression which includes both cross sectional and time series data. Panel
regressions are used when one wishes to explain differences across observations in a single period, but there are not enough observations in a single
period to generate reliable results. Thus one or more cross sections from
additional time periods are added to enlarge the sample. Panel regression
methods may handle models in which the dependent variable is continuous,
discrete, limited, or based on counts.

Case Study of Indian Two-wheeler Automotive

Industry

Automobile is one of the largest industries in the global market. Being the
leader in product and process technologies in the manufacturing sector, it
5

has been recognised as one of the drivers of economic growth. During the
last decade, well directed efforts have been made to provide a new look to
the automobile policy for realising the sectors full potential for the economy.
Two-wheeler segment is one of the most important components of the automobile sector. The two-wheeler industry has been in existence in India since
1955. It consists of three segments viz. scooters, motorcycles and mopeds.
According to the figures published by Society of Indian Automobile Manufacturers (SIAM), the share of two-wheelers in automobile sector in terms
of units sold was about 80 per cent during 2003-04. This high figure itself
is suggestive of the importance of the sector.

4.1

Growth Perspective

The composition of the two-wheeler industry has witnessed major changes

in the post-reform period. In 1991, the share of scooters was about 50
per cent of the total 2-wheeler demand in the Indian market. Motorcycle
and moped had been experiencing almost equal level of shares in the total
number of two-wheelers. In 2003-04, the share of motorcycles increased to 78
per cent of the total two-wheelers, while the shares of scooters and mopeds
declined to the level of 16 and 6 per cent respectively. A clear picture of the
motorcycle segments gaining importance during this period is exhibited by
the Figures 1, 2 and 3 depicting total sales, share and annual growth during
the period 1993-94 through 2003-04.

4.2

Demand Forecast

Estimations were based on Panel Regression, which takes into account both
time series and cross section variation in data. A panel data of 16 major
states over a period of 5 years ending 1999 was used for the estimation of
parameters. The models considered a large number of macro-economic, demographic and socio-economic variables to arrive at the best estimations for
different two-wheeler segments. The projections have been made at all India
and regional levels. Different scenarios have been presented based on different assumptions regarding the demand drivers of the two-wheeler industry.
The most likely scenario assumed annual growth rate of Gross Domestic
Product (GDP) to be 5.5 per cent during 2002-03 and was anticipated to
increase gradually to 6.5 per cent during 2011-12. The all-India and regionwise projected growth trends for the motorcycles and scooters are presented
in Table 1.
The above-mentioned forecast presents a long-term growth for a period of
10 years. The high growth rate in motorcycle segment at present is forecast
to stabilise after a certain point beyond which a condition of equilibrium

will set the growth path.

Table 1 suggests two important dimensions for the two-wheeler industry.
The region-wise numbers of motorcycle and scooter suggest the future market for these segments. At the all India level, the demand for motorcycles
will be almost 10 times of that of the scooterst, the same in the western
region will be almost 20 times. It is also evident from the table that motorcycle will find its major market in the western region of the country, which
will account for more than 40 per cent of its total demand. The south and
the north-central region will follow this. The demand for scooters will be
the maximum in the northern region, which will account for more than 50
per cent of the demand for scooters in 2011-12.
7

2-Wheeler
Segment

South

West

Regions
North-central

Motorcycle
Scooter

2835 (12.9)
203 (2.6)

4327(16.8)
219 (3.5)

2624 (12.5)
602 (2.8)

East and
North-east
883 (11.1)
99 (2.0)

All India
10669 (14.0)
1124 (2.08)

Table 1: Demand Forecast for Motorcycles and Scooters for 2011-12

4.3

Conclusion

There is a large market in semi-urban and rural areas of the country. Any
strategic planning for the two-wheeler industry needs to identify these markets with the help of available statistical techniques. Potential markets can
be identified as well as prioritised using these techniques with the help of
secondary data on socio-economic parameters. For the two-wheeler industry, it is also important to identify the target groups for various categories of
motorcycles and scooters. With the formal introduction of secondhand car
market by the reputed car manufacturers and easy loan availability for new
as well as used cars, the two-wheeler industry needs to upgrade its market
information system to capture the new market and to maintain its already
existing markets. Availability of easy credit for two-wheelers in rural and
smaller urban areas also requires more focussed attention. It is also imperative to initiate measures to make the presence of Indian two-wheeler
industry felt in the global market. Adequate incentives for promoting exports and setting up of institutional mechanism such as Automobile Export
Promotion Council would be of great help for further surge in demand for
the Indian two-wheeler industry.

References
[SCOTT] Long Range Forecasting From Crystal Ball To Computer; J.Scott
Armstrong; A Wiley Interscience Publication.
[FADA] Federation of Automobile Dealers
http://www.fadaweb.com/home.htm

Associations

of

India.

[MOTE] Managerial Economics: Concepts and Classes; Mote et al.;Tata

McGraw-Hill Publications;1985