Vous êtes sur la page 1sur 49

FORECASTING

Premising Forecast
Establishment of, and the agreement by managers
and planners to utilize, consistent assumptions
critical to plans under consideration.
Planning premises are defined as the anticipated
environment in which plans are expected to operate.
They include assumptions or forecasts of the future
and known conditions that will affect the operation
of plans.
Plans themselves and forecasts of their future effects
often become premises for other plans.
Environmental Forecasting

The making of forecasts and their review by


managers compel thinking ahead, looking to the
future, and providing for it.
Preparation of the forecast may disclose areas
where necessary control is lacking.
Forecasting, especially when there is participation
throughout the organization, helps to unify and
coordinate plans.
Forecasting Techniques
Sales forecast(Estimate)
The sales forecast is a prediction of expected
sales, by product and price, for a number of
months or years based on a chosen marketing
plan & an assumed marketing environment.
The sales forecast is the key to internal planning.
Business and capital outlays and policies of all
kinds are made for the purpose of maximizing
profits from expected sales.
Forecasts vs. Potential

Expectations Possibilities
Firm/Brand Sales Forecast Sales Potential

Category Market Forecast Market Potential


(Industry)
Major Uses of Potential Estimates

1. To make entry / exit decisions


2. To make resource level decisions
3. To make location and other resource
allocation decisions
4. To set objectives and evaluate
performance
5. As an input to forecasts
Qualitative Approaches

Used when Historical data is not available.


The cost of using qualitative techniques can be
high because of the time commitment required
from the people involved.
A significant change in environmental conditions
affecting the time series may make the use of
past data questionable in predicting future values
of the time series.
1.Jury of executive opinion method
It merely combines and averages the views, many of
which may be little more than hunches, of top managers.
The final estimate is an opinion of the president, based
on a consideration of the opinions of other officers.
It involves the careful judgement of experienced
executives who have studied the underlying factors that
influence their company's sales.
The No. of experts used in judgemental forecasting
should be between 5 & 20.
Year Sugar Import Total Sugar Export
Production Availability Consumption

2005/06 19.2 0 23.2 18.5 1.1


2006/07 28.3 0 32.6 19.9 1.7
2007/08 26.3 0 37.3 21.9 4.9
2008/09 14.5 2.4 27.4 22.9 0.1
2009/10 18.9 4.0 27.3 21.3 0.2
2010/11 24.2 0 29.1 21.0 2.1
2011/12 26.0 0 31.5 22.0 -

All figures in million tonnes


Source : Indian Sugar Mills Association
There are 527 functioning mills in the country.
2. Delphi Method
The objective is to produce a relatively narrow spread
of opinions within which the majority of experts concur

A panel of experts on a particular problem area is


selected - usually from both inside and outside the
organization.
The experts are asked to make (anonymously, so that
they will not be influenced by others) a forecast as to
what they think will happen, and when, in various areas
of new discoveries or developments.
Cont
The answers are compiled, and the composite
results are fed back to the panel members.
With this information at hand (but still with
individual anonymity),further estimates of the
future are made.
This process may be repeated several times.
(iterations)
When a convergence of opinion or consensus
begins to evolve, the results are then used as an
acceptable forecast.
3. Scenario Writing
It consists of developing a conceptual scenario of
the future based on awell defined set of
assumptions.
The job of decision maker is to decide how likely
each scenario is (probability) & then make
decisions

Sales Crude @$ 60-75 Crude@$ 100-150 Crude@$150-200


Passenger Cars
Commercial Vehicles
2 Wheelers
Utility Vehicles
4. Sales force composite method
Bottom Up approach
Use of Spreadsheets is a must.
Obtain from salespeople and sales managers their
combined view as to expected sales & collate
them at head office to obtain the sales forecast
for the entire company.
Salespeople are given guidelines in the form of
company planning premises as to business
conditions in general.
Territory-Karnataka

April May June July August September October November December January February March

Distributor ABC
Distributor DEF
Distributor GHI
TOTAL

Distribut
or ABC
Model 01
Model 02
Model 03
Model 04
Model 05
TOTAL
Distribut
or DEF
Model 01
Model 02
Model 03
Model 04
Model 05
TOTAL
Quantitative Approach
5. Users' expectation method
Bottom Up approach
Use of Spreadsheets is a must.
Identifying all the potential buyers in each market.
Base forecasts on expected purchases by industrial
customers because their demand arises out of the
demand of end use products.
Used where a few large companies are dominant.
To determine multiplier factor or cascading effect on
intermediate/end user goods & services.
6. Trend Projection
This analysis is based on the assumption that "what is
past is prologue" and that a trend will continue unless
something happens to it.
Indian Telecom
Wireless Growth

wireless

Introduction of CPP gsm

wireless net
additions

Fastest growing market in the world


7.1 Simple Moving Averages
A simple moving average is used when there is neither a
trend nor a cyclical pattern.
Thus,the forecast is an average of the last nth observation:
Salest+1 = 1/n (St + St-1 + +St-n )
Where
Salest+1 = Forecast Sales
St = Sales in present period
St-1 = Sales in the period immediately past & so on
n= number of time periods over which moving average is
calculated
Simple Moving Average Forecast

3/00 4/00 1/01 2/01 3/01 4/01 1/02 2/02 3/02 4/02
Actual Rs sales in thousands 52 52 74 55 64 66 83 76 78
(deseasonalized)
Two-period moving average 52* 63 65 60 65 75 80 77
Three period moving average 59 60 64 62 71 75 79

*
52 + 52
=52
2

52 + 52 + 74
=59.3
3
7.2 Weighted moving averages
Qtr. Weight Sales Weighted
3/00 1 52 52
4/00 2 52 104
1/01 3 74 222
Total 6 378

So a 3 Qtr weighted moving average


=378/6= 63
More weight is given to the latest data & less
weight is given to the past data.
7.3. Exponential Smoothing
Using an exponential smoothing model, the forecaster
can allow sales in certain periods to influence the
forecast more than sales in other periods.
It is a special case of the weighted moving averages
method in which we select only one weight-the weight
for the most recent observation.
Used for short term averages (Daily Sales e.g.
Carbonated Drinks
Continuous Replenishment Program of Walmart (5
Days Moving Averages)
Nano Sales since launch in July 2009
Using Moving Averages to forecast price movements

As long as EMA(5) > EMA(13) > EMA(26), Price will go up


As long as EMA(5) < EMA(13) < EMA(26), Price will go Down
Stock Prices ( Todays actual price*0.91 + todays forecasted price*0.09)
8. Time Series Analysis
Used to analyze chronologically arranged sets of data in
order to explain dips and surges from the long term
trend.
The data may be in terms of weekly,monthly, quarterly, or
annual sales volume.
Time Series analysis attempts to determine 4 sets of
factors which influence sales
The 4 factors can be combined additively
Sales = T+C+S+R
or multiplicatively
Sales = T * C * S * R
Trend(T) It is the long term movement of the series which
can be attributed to factors such as population growth.
Cyclical (T) : It is a recurrent fluctuation of sales above
and below the trend line,generally referred to as business
cycle. Cycles do not have a uniform frequency and may
be different for different industries (5-8 years)
Seasonal (S) It is a consistent pattern of sales movement
within the year.e.g cold remedies are more in demand
during the monsoon and winter months.Fertilizers are in
demand in 2 seasons, Kharif and Rabi
Residual (R) It is the unexplainable component of the
sales data
Types of Sales Fluctuations
An Example of Time Series Analysis

Year Sales
1991 130
1992 160
1993 150
1994 180
1995 190
1996 210
1997 200
1998 210
Ex. Commercial Vehicle Sales Expected to Rise 15-25%

Rolling Ahead
Total CV sales in India
Focus on infrastructure, normal monsoon
to give boost
Forecast according
to trend 560597

V Shape Recovery

U Shaped Recovery
Jubilant Foodworks,master franchisee of Dominos in India
offered shares in an IPO in 2009 @ Rs 265/Share.
The shares have touched a high of Rs 950 in 2 years and trades
@ 33times P/E in July 2011
9. Developing Regression Models

Plot Sales Over Time


Consider the Variables that Are Relevant to
Predicting Sales
Collect Data
Analyze the Data
Examine the correlations among the independent variables
Run the regression
Determine the significant predictors
Cereal Sales Data (Monthly)
Format for Reporting a Regression
Model Based Forecast
10. Intention to Buy Surveys (Chain Ratio Method)

Total Market Potential or estimating current


demand for a newly launched product

It involves multiplying a base number by several


adjusting percentages.
Brand: Buddy

Product Category School Bags

Key Markets South & West for current


launch

Launch : April 2009

TG- School kids (4-14 yrs), Sec A,B.

The Buddy range has character bags for kids


from age 4- 7yrs and non character bags for
kids from age 8- 14 years .
What People say is not always what people do?
Consumer surveys of buyer intention are always heavily discounted
to allow for this fact

Purchase Intent % Response Rule of thumb reduction % of market deemed


for forecasting purposes likely to actually buy
Definitely would Buy 27% Multiply by 0.8 27% * 0.8 = 21.6%
Probably would Buy 43% Multiply by 0.3 43% * 0.3 = 12.9%
Might or might not buy 22% Count as zero
Probably or definitely 8% Count as zero
would not buy
Totals 100% 21.6% +12.9%
=34.5%
11. Life Cycle Modelling
Most products and services follow the standard life cycle
stages
Modelling the entire life cycle is a difficult task.
If the marketer knows which part of the life cycle he is
in,he can model that stage alone.
Sales and Product Life Cycle Q= ab^cx + d
Gompertz Growth Curve (a>0,1>b>0, c<1)
12. New Product Forecasting
Large costs associated with new consumer product
introduction and testing.
Consumer goods manufacturers have instituted a series
of methods for investigating how the market is likely to
react to the new product offering.
The major methods of market testing are
Sales Wave Research
Pretest markets or purchase labs
Controlled Test Marketing
Test Markets
Test market results and marketer actions
Repurchase Rate
High Low
Commercialize the Redesign/Drop
Product Product
High
Trial Rate

Redesign/Increase
advertising, Drop the Product
use sales promotion,
sampling
Low
13. Correlation Analysis
The measurement of the relationship between
company sales and one or more other factors.e.g
Sales & Advertising in (Rs.)
Fertiliser sales in (tonnes) & rainfall in (mm)
Carbonated drink sales in cases & temperature (0C)
during summer months.
14. Lead-Lag Co-relation
Changes in macroeconomic variables occur
before/after changes in the economy,they are
termed lead/lag indicators
Such correlation, whether direct or with a lag or
lead of a given time period, can give a company
useful and highly reliable basis for sales
forecasting.e.g.
6 month lag between Credit demand & lowering/
raising interest rates
CV sales & GDP growth rates
15. Econometric Models

These are essentially large scale,multiple


equation regression models.
They never forecasted as well as advertised.
Thank You

Vous aimerez peut-être aussi