Unit-3: Sales Forecasting

SALES FORECASTING
In general terms, ‘forecast’ refers to the process of
prediction of some future outcome. Such definition
may be technically appropriate, but in managerial
perspective, the definition is very general.
According to Stults, “A sales forecast is an estimate
of the amount or unit sales for a specified future
period under proposed marketing plan or
program”.
 OBJECTIVES OF SALES FORECASTING
• To formulate appropriate production policy for meeting the
demand forecasted by the organisation.
• To ensure continuous availability of raw materials for supporting
the production process as per the sales forecast.
• To optimise the use of machinery.
• To determine appropriate pricing policy for specified time period.
• To make a robust estimate of the working capital required by the
business and also make provisions for the same.
• To fix the sales quotas and targets for different segments in
which the company operates.
• To make an estimate of the inventory requirements for the
finished, unfinished, and semi-finished
• To get an estimate of the cash inflow from the sales operations of the
business.
• To make provisions for the capital expenditure of the organisation.
• To plan and estimate the plant/production capacity of the organisation for
meeting the increasing future demands.
• To plan for adequate manpower required for production and distribution
activities.
• To make arrangements for procuring the necessary raw materials as per
future demand of the organisation.
• To plan for the dividend policy of the organisation.

• To minimise the selling cost so that the total cost of the product can be kept
at minimum.

• To initiate budgetary and planning activity.

• To forecast the future profits that the organisation is likely to earn.

•
 CRITERIA FOR EFFECTIVE FORECASTING

• Accuracy
• Availability of Statistical Indexes
• Durability
• Flexibility
• Plausibility
 USES OF SALES FORECASTING

• Sales Planning
• Marketing
• Inventory Control
• Supply Chain Management
• Financial Planning
• Price Stability
• Demand Forecasting
 METHODS OF SALES FORECASTING

1. Qualitative Methods of Forecasting→

Qualitative methods of forecasting rely on the
judgements of experts or the collective
opinions of stakeholders or the customer.
These methods can be categorised as follows:
1.Executive Opinion

2.Delphi Method

3.Historical Analogy Method

4.Survey of Buyer’s Expectations

5.Sales Force Composite Method

2. Quantitative Methods of Forecasting
• Trend Projection Method
• Econometric Method
• Naive/Ratio Method
• Exponential Smoothening
 REGRESSION MODEL TO FORECAST SALES

• Regression analysis is a statistical tool used to

calculate a continuous dependent variable
from various independent variables and is
commonly used for prediction and forecasting.
For example, if agriculturist knows that the
yield of rice and rainfall are closely related then
he will want to know the amount of rain
required to achieve a certain production. In this
situation, he will use the regression analysis.
• This technique is used for modelling and analysis of mathematical data using
the value of a dependent variable and one or more independent variables.
Dependent variable, also known as response variable, is the single variable
predicted by the regression model. Independent variable, also known as
predictor variable, is used to predict the values of dependent variable.
• Regression analysis is the process to calculate the unknown value of one
variable from the known values of the other variable.
• According to Blair, “Regression is the measure of the average relationship
between two or more variable in terms of the original units of the data.”
• According to Taro Yamane, “One of the most frequently used techniques in
economics and business research, to find a relation between two or more
variables that are related causally, is regression analysis.”
• Regression may be classified as follows:
• Simple Regression
• Multiple Regression
• Linear Regression
• Non-Linear Regression
• APPLICATION OF REGRESSION ANALYSIS
• Following are the basic uses of regression analysis:
• Used to know the relationship between different (one or more) variables.
• Used to find out the value of dependent variable from the value of independent
variables.
• Used to find out the coefficients of correlation (r) and coefficients of
determination.
• In corporate sector it is useful to check the quality.
• Also very useful to determine the statistical curve (demand, supply, etc.).
•
• REGRESSION LINES
• The lines of the best fit between two variables and stating the common average
relationship are known as regression lines. There are two regression lines for
two variables.
• When we take two variables X as dependent variable and Y as independent
variable, then the straight line is called as the regression line of X and Y and vice
versa is also true
 MULTIPLE REGRESSION

• Three or more variables, such as X1, X2, and X3,

are analysed in multiple regression analysis.
Here, X1 is taken as a dependent variable and
its relative movement is found out, according
to the movement in both independent
variables X2 and X3. Thus, the effect of two or
more independent variables on one
dependent variable is studied in the ‘multiple
regression analysis’.
ASSUMPTIONS OF MULTIPLE REGRESSION
ANALYSIS
• The dependent variable should be a random variable
whereas the independent variable need not be
random variable,
• The relationship between various independent
variables and dependent variable is linear, and
• The variances of the conditional distributions of the
dependent variable and various combinations of
values of the independent variables are all equal.
The conditional distributions for the dependent
variable follow the normal probability distribution
 MODELING TREND AND SEASONALITY

• When a researcher collects information and

plots a pattern or trend (based on that
information) then this process is known as
“trend analysis”. Trend analysis is used to
estimate and forecast the future and past
events. For example, a researcher can make an
estimate that how many ancient kings may rule
in a given time period (between two dates).
This calculation is based on various data, such
as average years in which other kings reigned.
 DETERMINATION OF TREND

• The secular trend is determined by four

methods which are as follows:
• a) Graphic Method
• b) Semi-Average Method
• c) Moving Average Method
• d) Least Squares Method
•
 SEASONAL VARIATIONS

• By seasonal variations in a time series we

mean such variations which occur regularly
and periodically with period less than one
year. For example, the sale of ice cream is
affected by weather conditions so that in
every summer the sale would be more than
what it is in winters. This is the seasonal effect
on the sale of ice cream and would be in every
season.
• MEASUREMENT OF SEASONAL VARIATIONS
• The following are some of the methods more
popularly used for measuring seasonal
variations:

• a) Simple Average Method

• b) Ratio to Trend Method
• c) Ratio to Moving Average Method
• d) Link Relative Method
•
 USING S CURVE TO FORECAST SALES OF A NEW PRODUCT

• New product revenue and sales forecasting

approaches can be as simple as “guesstimating”
the first year sales of a given product, escalating
it for future year forecasts by an annual growth
rate until a specific level of saturation is
reached. This simple approach assumes that
sales will continually grow throughout the
product’s lifecycle.
 New Product Sales Forecasting S Curve
Models: Product Diffusion
• The new product sales model explains this S-
curve shape based on diffusion theory.
Diffusion theory is actually a theory of
communication regarding how information is
dispersed within a social system over time.
Because people place different emphases on
how much they rely on media and interpersonal
communication for new ideas and information,
they “adopt” new products either earlier or
later in a product’s lifecycle.
• The “S-curve” model is what is known as a
“single purchase” model in that it forecasts
sales of products that are typically bought just
once, or infrequently, such as consumer
durables or industrial products such as
mainframe computers.
 Assumptions of Diffusion(S) models

• The product whose sales are being forecast by

the model is a product that is destined to be a
successful new product introduction.
Estimates of new product failure rates vary
from 33% to 60% or higher. The present model
is appropriate only for successful new product
entries
• When the user estimates their brand’s market
share within the product category, a number of
underlying assumptions about competitive
response underlie such an estimate
• The model requires that the user input an
estimate for the total market size for the
particular brand within the product category
(i.e., total number of adopters of the branded
product).
• The model recommended here has been
developed from theoretical work in the area
of diffusion processes and the customer new
product adoption process. Diffusion process
models attempt to forecast the market
penetration rates of innovative products (air
conditioners, cell phones, the internet, hybrid
gas/electric cars, a new brand of coffee, etc.)
over time
S curve model