Methods of Demand

Forecasting
Dr. Rashmi Ahuja
Forecasting has fascinated
people for thousands of
years…..
 Tell us what the future holds, so we may know that
you are gods. (Isaiah 41:23)
What can be forecasted ???

 Forecasting is required in many situations:

 deciding whether to build another power
generation plant in the next five years requires
forecasts of future demand;

 scheduling staff in a call centre next week requires

forecasts of call volumes;

 stocking an inventory requires forecasts of stock

requirements. 
Introduction
All organizations conduct their activities in an
uncertain environment. The major role of forecasting
is to reduce this uncertainty.

In order that corporate management can set

reasonable targets for its objectives, it must have
available the relevant forecasts, both for the short and
long terms.

Corporate planners in all areas will utilize an array of

forecasts in constructing the various portions of the
business plan.
Decisions that need forecast….

 Which markets to pursue?

 What products to produce?
 How many people to hire?
 How many units to purchase?
 How many units to produce?
 And so on……
 Some things are easier to forecast than others.

 The predictability of an event or a quantity depends on

several factors including:
 how well we understand the factors that contribute to it;
 how much data is available;
 whether the forecasts can affect the thing we are trying to
forecast.
 Eg.
 For example, forecasts of electricity demand can be
highly accurate because all three conditions are usually
satisfied.

 When forecasting currency exchange rates, only one of

the conditions is satisfied: there is plenty of available
data. However, we have a limited understanding of the
factors that affect exchange rates, and forecasts of the
exchange rate have a direct effect on the rates
themselves.
Common Characteristics of
Forecasting
 Forecasts are rarely perfect

 Forecasts are more accurate for aggregated data

than for individual items

 Forecast are more accurate for shorter than longer

time periods
Forecasting Steps

 What needs to be forecast?

 Level of detail, units of analysis & time horizon required
 What data is available to evaluate?
 Identify needed data & whether it’s available
 Select and test the forecasting model
 Cost, ease of use & accuracy
 Generate the forecast
 Monitor forecast accuracy over time
Subject of Forecast

 Macro forecasts
 Gross domestic product
 Consumption expenditure
 Producer durable equipment expenditure
 Residential construction
 Industry forecasts
 Sales of an industry as a whole
 Sales of a particular product within an industry
 Firm-level forecasts
 Sales
 Costs and expenses
 Employment requirements
Prerequisites of a Good Forecast

A good forecast should

 be consistent with other parts of the business.
 be based on adequate knowledge of the relevant past.
 take into consideration the economic and political
environment.
 be timely.
Forecasting techniques
 No. of techniques available.
 Imp. to choose one that gives accurate forecast with low cost
and minimum use of other resources.
 Imp. Factors for manager to consider :-
 Desired level of accuracy.
 Availability of data.
 Length of forecast period.
 Associated cost & Benefits.
Two kinds of Techniques
Qualitative Techniques
based on judgments,
opinions, intuition,
emotions, or personal
experiences. [surveys &
interviews]
subjective in nature.
They do not rely on any
rigorous mathematical
computations.
The qualitative methods
rely on the opinion of
different groups of people
who are associated with the
product to predict the future
demand.
Quantitative Techniques
These are the methods that
make use of statistical tools
to predict the future
demand of the product.
Objective in nature.
They rely heavily on
mathematical
computations.
They forecast future
demand from past data and
extrapolate it to make
forecasts of future levels.
 If there are no data available, or if the data
available are not relevant to the forecasts,
then qualitative forecasting methods must be
used.

 Quantitative forecasting can be applied when

two conditions are satisfied:
 numerical information about the past is available;
 it is reasonable to assume that some aspects of the
past patterns will continue into the future.

 New products  which techniques ???

 Existing products  which one ???
Qualitative Techniques
Consumer Surveys
 Survey consumers for advanced information about their
buying intentions.
 Consumer surveys can be used to ask consumers
directly about their buying behaviour, using a number
of ways : face-to-face, direct mail, internet, telephone,
spot survey at store checkout points.
 Eg. Airlines asking randomly selected group of
consumers abt their travel plans, services prices, views
regarding competitors etc.
 Consumers are asked about their reaction to hypothetical
changes in prices, income, prices of substitutes and
complements, advertising expenditure, consumer
expectations reg. future prices, inflation etc.
 Three ways for consumer surveys :-
• Complete survey of all consumers(complete
enumeration) :
• Few consuming units out of relevant population
(sample survey) :
• Industries using as end Product (End-use method)
Delphi Method
• The Delphi method originally developed by Rank
Corporation in 1969 for forecasting military events,
has become a useful tool in other areas also.

• This method attempts to arrive at a consensus by

questioning a group of experts in this area.

• It helps to capture the knowledge of diverse experts

while avoiding the disadvantages of traditional group
meetings.
Steps
 A panel of experts is assembled.
 Forecasting tasks/challenges are set and distributed to
the experts.
 Experts return initial forecasts and justifications. These
are compiled and summarised in order to provide
feedback.
 Feedback is provided to the experts, who now review
their forecasts in light of the feedback. This step may
be iterated until a satisfactory level of consensus is
reached.
 Final forecasts are constructed by aggregating the
experts’ forecasts.
 Jury of Executive opinion
Method
 A sales forecasting method, wherein the executives from different
departments come together and forecast sales for the given period,
on the basis of their experience and specialization.

 The jury method is based on the judgments, the top executives of

Marketing, HR, Finance, Production department come
together and give their opinions on sales trend. The final
forecast is arrived by averaging the opinions given by all at the
meeting.

 This method is based on assumption that executives are well

informed about the industry outlook and co’s marketing position,
program and capabilities etc. All should support their estimates with
factual material and explain their rationales.
Advantages
 Quick and Easy way to turn out a forecast.
 This is a way to pool the experience and
judgement of well informed people.
 This method may be used when co. is so young
that it has not yet accumulated the experience
to use other forecasting techniques.
 This method may be used when adequate sales
and marketing statistics are missing or this data
is not available in reqd. form for use of more
sophisticated techniques..
Disadvantages
 Finding based on opinion.
 Increases workload of key executives.
 Difficult to break forecast by this method into
estimates of probable sales by products, by time
intervals, by markets, by customers etc.
 Sales Force composite Method
 The Sale Force Composite Method is a sale forecasting
method wherein the sales agents forecast the sales in their
respective territories, which is then consolidated at
branch/region/area level, after which the aggregate of all
these factors is consolidated to develop an overall
company sales forecast.
 The sales force composite method is the bottom-up
approach where the sales force gives their opinion on
sales trend to the top management.
 Can be more accurate because it is from the sales person
who are directly in contact with the consumers.
Advantages of Sales
composite force method
 The intimate knowledge and experience of the sales force
in their respective territories can be used efficiently.
 The responsibility to forecast sales rests on the shoulders
of the sales agent and thus could be held accountable if
anything goes wrong.
 Since the sales agents forecast the sales by themselves,
put more efforts to achieve them.
 This method is more reliable because of a large population
sample and moreover, it can be readily broken down into
product-wise, month-wise, area-wise forecast.
 Limitations
 Since the sales agents are not the experts in forecasting, they
cannot employ the sophisticated forecasting techniques properly
and neither they have complete data to have a fact-based
forecasting.

 Also, the salesman often gets heavily influenced by the

conditions existing in his territory, due to which he either
becomes more optimistic or more pessimistic about the future
sales.

 Sometimes, the sales agent intentionally gives fewer sales

forecast, so that they can fetch more incentives or bonus from
the management on exceeding the sales targets
Quantitative Techniques

attempts to predict the

 Time series Model -

future by using historical data over

time .
“What will happen in the future is a function of
what happened in the past”

 Casual method -incorporates factors that may

influence the quantity being forecasted into the
model . Eg Predict sales of cola: temperature,
season, day of week, humidity etc
Time Series Analysis

 Forecast based on the analysis of time-series data.

 Anything that is observed sequentially over time is
a time series.
 Time Series data : values arranged by days, weeks,
years. [Eg. Monthly sales data,quarterly GDP etc.]
 Plot past data, examine the trend, forecast the
future.
 Assumption:
o Past trend will remain in the future
o when changes in a variable shows a clear pattern over
time, this is appropriate.
Examples of time series data include:

 Daily IBM stock prices

 Monthly rainfall
 Quarterly sales results for Amazon
 Annual Google profits
Figure shows the quarterly Australian
beer production from 1992 to the
second quarter of 2010.
 The value of the variable for the future as a
function of its values in the past.
Dt+1 = f ( Dt , Dt-1, Dt-2, .....)
 There exists a function whose form must be
estimated using the available data..
 The most common technique for estimation of
equation is regression analysis.
Components of Time series data

 Trend
 Seasonal component
 Cyclical
Trend component - A trend exists when there is a long-term
increase or decrease in the data.
Seasonality : A seasonal pattern occurs when a time series is
affected by seasonal factors such as the time of the year or the day
of the week. Seasonality is always of a fixed and known frequency.
 A cycle occurs when the data exhibit rises and falls that are
not of a fixed frequency. These fluctuations are usually due
to economic conditions, and are often related to the
“business cycle”.
 These components are isolated and then
mathematical model is developed and forecast is
predicted.
 Generally more suitable for long term forecasts
in most companies.
Model Description
Naïve Uses last period’s actual value as a
forecast.
Simple Mean Uses an average of all past data as a
(Average) forecast
Simple Uses an average of a specified number of
Moving the most recent observations, with each
Average observation receiving the same emphasis
(weight)
Weighted Uses an average of a specified number of
Moving the most recent observations, with each
Average observation receiving a different emphasis
(weight)
Exponential A weighted average procedure with
Smoothing weights declining exponentially as data
become older
Trend Technique that uses the least squares
Projection method to fit a straight line to the data
Naïve Forecasting

Next period forecast = Last Period’s actual:

Ft 1  At
Simple Average (Mean)

Next period’s forecast = average of all historical data

At  At 1  At  2  .............
Ft 1 
n
Moving Average

Next period’s forecast = simple average of the last N periods

At  At 1  .........  At  N 1
Ft 1 
N
The Effect of the Parameter N ????
Weighted Moving Average

Ft 1  C1 At  C2 At 1  .........  C N At  N 1
where
C1  C2  .........C N  1
Exponential Smoothing

Ft 1  At  1    Ft
where
0  1
Time Series Problem

Determine forecast for

periods 11
 Naïve forecast
 Simple average
 3- and 5-period moving
average
 3-period weighted moving
average with weights 0.5,
0.3, and 0.2
 Exponential smoothing with
alpha=0.2 and 0.5
Time Series Problem Solution
 Casual methods

 Simple and multiple regression models

 Dep. Variable = f ( no of. indep. Variables….)

Forecast Accuracy

 Forecasts are rarely perfect

 Need to know how much we should rely on our chosen forecasting
method
 Measuring forecast error:
Et  At  Ft
 Note that over-forecasts = negative errors and under-forecasts =
positive errors
 Tracking Forecast Error
Over Time
 Mean Absolute Deviation
(MAD):  actual  forecast
 A good measure of the actual MAD 
n
error in a forecast

 Mean Square Error (MSE):

 Penalizes extreme errors

� actual - forecast
2

MSE 
n
Auto sales at Carmen’s Chevrolet are shown below.
Develop a 3-week moving average.

Week Auto Sales

1 8

2 10

3 9

4 11

5 10

6 13

7 -
Carmen’s decides to forecast auto sales by weighting the
three weeks as follows:

Weights Applied Period

3 Last week

2 Two weeks ago

1 Three weeks ago

6 Total
 Exponential smoothing is used to forecast
automobile battery sales. Two values of α are
examined i.e. α=0.8 and α=0.5. Which is
preferable? (Assume the forecast for January was
22 batteries.) Actual sales are given below

Month Actual Battery Sales Forecast

January 20 22

February 21  
March 15  
April 14  
May 13  
June 16