SCA

Forecasting

By
Prof. Goutam Kumar Kundu
What is Forecasting?
SCA

• Process of predicting a future event

• Underlying basis of all business decisions
 Production
 Inventory
 Personnel
 Facilities
• May involve predicting future by
 using mathematical modeling on historical data
 Or subjective or intuitive prediction
 Or both
Forecasting Time Horizons
SCA

• Short-range forecast
 Up to 1 year, generally less than 3 months
 Purchasing, job scheduling, workforce levels, job assignments, production
levels
• Medium-range forecast
 3 months to 3 years
 Sales and production planning, budgeting
• Long-range forecast
 3+ years
 New product planning, facility location, research and development
Distinguishing Differences
SCA

• Medium/long range forecasts

 deal with more comprehensive issues and support management
decisions regarding planning and products, plants and processes
• Short-term forecasting
 usually employs different methodologies than longer-term forecasting
 Short-term forecasts tend to be more accurate than longer-term
forecasts
Types of Forecasts
SCA

• Economic forecasts
 Address business cycle – inflation rate, money supply, housing starts, etc.
• Technological forecasts
 Predict rate of technological progress
 Impacts development of new products
• Demand forecasts
 Predict sales of existing product
• For OM emphasis is on demand forecasting
Strategic Importance of Forecasting
SCA

• Human Resources
 Hiring, training, laying off workers
• Capacity
 Capacity shortages can result in undependable delivery, loss of
customers, loss of market share
• Supply-Chain Management
 Good supplier relations and price advance
 Seven Steps in Forecasting
SCA

• Determine the use of the forecast

• Select the items to be forecasted
• Determine the time horizon of the forecast
• Select the forecasting model(s)
• Gather the data
• Make the forecast
• Validate and implement results
In reality
SCA

• Forecasts are seldom perfect

• Most techniques assume an underlying stability in the system
• Product family and aggregated forecasts are more accurate than
individual product forecasts
Forecasting Approaches
SCA

Qualitative Methods
• Used when situation is vague and little data exist
 New products
 New technology
• Involves intuition, experience
Forecasting Approaches
SCA

Quantitative Methods
• Used when situation is „stable‟ and historical data exist
 Existing products
 Current technology
• Involves mathematical techniques
 e.g., forecasting sales of color televisions
 Overview of Qualitative Methods
SCA

• Jury of executive opinion

 Pool opinions of high-level executives, sometimes augment by statistical
models
• Delphi method
 Three types of participants – decision makers, staff personnel and respondents
 Panel of experts, queried iteratively
 Staff assists decision makers
 Respondents are people ( from different locations) whose opinions are valued
• Sales force composite
 Estimates from individual salespersons are reviewed for reasonableness, then
aggregated
• Consumer Market Survey
 Ask the customer
 Jury of Executive Opinion
SCA

• Involves small group of high-level managers

• Group estimates demand by working together
• Combines managerial experience with statistical models
• Relatively quick
• „Group-think‟ disadvantage
Delphi Method
SCA

• Iterative group process,

continues until consensus is Decision Makers
reached (Evaluate
• 3 types of participants
responses and
make decisions)
 Decision makers
 Staff
Staff
 Respondents (Administering
survey)

Respondents
(People who can
make valuable
judgments)
Sales Force Composite
SCA

• Each salesperson projects his or her sales

• Combined at district and national levels
• Sales reps know customers‟ wants
• Tends to be overly optimistic
Consumer Market Survey
SCA

• Ask customers about purchasing plans

• What consumers say, and what they actually do are often different
• Sometimes difficult to answer
Overview of Quantitative Approaches
SCA

1. Naive approach
2. Moving averages
Time-Series
3. Exponential smoothing Models
4. Trend projection

Associative
5. Linear regression Model
Time Series Forecasting
SCA

• Set of evenly spaced (weekly, monthly, quarterly and so on)

numerical data
 Obtained by observing response variable at regular time periods
 Daily sales, quarterly profit etc
• Forecast based only on past values
 Assumes that factors influencing past and present will continue influence
in future
Components of Demand
SCA

Trend
component
Demand for product or service

Seasonal peaks

Actual
demand

Average
demand over
Random four years
variation
| | | |
1 2 3 4
Year
Naive Approach
SCA

• Assumes demand in next period is the same as demand in most

recent period
 e.g., If May sales were 48, then June sales will be 48
• Sometimes cost effective and efficient
• At least provides a starting point
Moving Average Method
SCA

• Uses historical data to forecast

• Assumption – demand will stay fairly steady over time
• 4 month moving average – sum demand during past 4 months and
divide by 4
• With each passing month latest month added to the trend and earliest
dropped.
Moving Average Method
SCA

• MA is a series of arithmetic means

• Used if little or no trend
• Used often for smoothing
 Provides overall impression of data over time

∑ demand in previous n periods

Moving average = n
Moving Average Example
SCA

Actual 3-Month
Month Shed Sales Moving Average
January 10
February 12
March 13
April 16
May 19
June 23
July 26
Moving Average Example
SCA

Actual 3-Month
Month Shed Sales Moving Average
January 10
February 12
March 13
April 16 (10 + 12 + 13)/3 = 11 2/3
May 19 (12 + 13 + 16)/3 = 13 2/3
June 23 (13 + 16 + 19)/3 = 16
July 26 (16 + 19 + 23)/3 = 19 1/3
Weighted
SCA
Moving Average

• Used when trend is present

 Older data usually less important and more emphasis on recent values
• Weights based on experience and intuition (arbitrary)

∑ (weight for period n)

Weighted x (demand in period n)
moving average = ∑ weights
 Weights Applied
Weighted Moving Average Period
SCA
3 Last month
2 Two months ago
1 Three months ago
6 Sum of weights

Actual 3-Month Weighted

Month Shed Sales Moving Average
January 10
February 12
March 13
April 16
May 19
June 23
July 26
 Weights Applied Period
SCA
3 Last month
Weighted Moving Average
2 Two months ago
1 Three months ago
6 Sum of weights

Actual 3-Month Weighted

Month Shed Sales Moving Average
January 10
February 12
March 13
April 16 [(3 x 13) + (2 x 12) + (10)]/6 = 121/6
May 19 [(3 x 16) + (2 x 13) + (12)]/6 = 141/3
June 23 [(3 x 19) + (2 x 16) + (13)]/6 = 17
July 26 [(3 x 23) + (2 x 19) + (16)]/6 = 201/2
 Potential Problems With Moving Average
SCA

• Increasing n smooths the forecast but makes it less sensitive to

changes
• Do not forecast trends well
• Require extensive historical data
Moving Average And Weighted Moving Average
SCA

Weighted
30 – moving
average
25 –
Sales demand

20 – Actual
sales
15 –
Moving
10 – average

5 –
| | | | | | | | | | | |
J F M A M J J A S O N D
Graph of Moving Average
SCA

• Weighted moving averages reacts more quickly to demand changes

• Even in periods of down turn this more closely tracks the demand
Exponential Smoothing
SCA

• Form of weighted moving average

 Weights decline exponentially
 Most recent data weighted most
• Requires smoothing constant (alpha)
 Ranges from 0 to 1
 Subjectively chosen
• Involves little record keeping of past data
Exponential Smoothing
SCA

New forecast = last period’s forecast

+ a (last period’s actual demand
– last period’s forecast)

Ft = Ft – 1 + a(At – 1 - Ft – 1)

where Ft = new forecast

Ft – 1 = previous forecast
a = smoothing (or weighting)
constant (1  a  0)
 SCA

Exponential Smoothing Example

Predicted demand = 142 Ford Mustangs

Actual demand = 153
Smoothing constant a = .20
 SCA

Exponential Smoothing Example

Predicted demand = 142 Ford Mustangs

Actual demand = 153
Smoothing constant a = .20

New forecast = 142 + .2(153 – 142)

= 142 + 2.2
= 144.2 ≈ 144 cars
Smoothing Constants
SCA

• For business applications range from .05 to .50

• Can be changed to give more weight to recent data (α is high) or
past data (α is low)
• α reaches 1 then, this is same as naïve method
•
Ft = At – 1
Forecasting Performance- How good is the forecast?
SCA

• Mean Forecast Error (MFE or Bias):

 Measures average deviation of forecast from actuals.

• Mean Absolute Deviation (MAD):

 Measures average absolute deviation of forecast from actuals.

• Mean Absolute Percentage Error (MAPE):

 Measures absolute error as a percentage of the forecast.

• Standard Squared Error (MSE):

 Measures variance of forecast error
Forecasting Performance Measures
SCA

n
1
MFE   ( Dt  Ft )
n t 1
1 n
MAD   Dt  Ft
n t 1
100 n Dt  Ft
MAPE  
n t 1 Dt
1 n
MSE   ( Dt  Ft ) 2

n t 1
Mean Forecast Error (MFE or Bias)
SCA

1 n
MFE   ( Dt  Ft )
n t 1

• Want MFE to be as close to zero as possible -- minimum bias

• A large positive (negative) MFE means that the forecast is
undershooting (overshooting) the actual observations
• Note that zero MFE does not imply that forecasts are perfect (no
error) -- only that mean is “on target”
• Also called forecast BIAS
Mean Absolute Deviation (MAD)
SCA

1 n
MAD   Dt  Ft
n t 1

• Measures absolute error

• Positive and negative errors thus do not cancel out (as with MFE)
• Want MAD to be as small as possible
• No way to know if MAD error is large or small in relation to the actual
data
Mean Absolute Percentage Error (MAPE)
SCA

100 Dt  Ftn
MAPE  
n t 1 Dt

• Same as MAD, except ...

• Measures deviation as a percentage of actual data
Mean Squared Error (MSE)
SCA

1 n
MSE   ( Dt  Ft ) 2
n t 1
• Measures squared forecast error -- error variance
• Recognizes that large errors are disproportionately more “expensive”
than small errors
• But is not as easily interpreted as MAD, MAPE
Comparison of Forecast Error
SCA

• During past 8 quarters a port unloaded a lot of grain

• OM wants to test exponential smoothing
• Guesses the forecast of grain unloaded in first quarter was 175 tons
• Two values α = .1 and .5 are used
Comparison of Forecast Error
SCA

Rounded Absolute Rounded Absolute

Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1 180 175 175
2 168
3 159
4 175
5 190
6 205
7 180
8 182
Comparison of Forecast Error
SCA

Rounded Absolute Rounded Absolute

Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1 180 175 5 175 5
2 168 176 8 178 10
3 159 175 16 173 14
4 175 173 2 166 9
5 190 173 17 170 20
6 205 175 30 180 25
7 180 178 2 193 13
8 182 178 4 186 4
84 100
 SCA

Comparison of Forecast Error

∑ |deviations|
Rounded Absolute Rounded Absolute
MADActual
= Forecast Deviation Forecast Deviation
Tonage n
with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1
For a180
= .10 175 5 175 5
2 168 = 84/8 = 10.50
176 8 178 10
3 159 175 16 173 14
4 For a175
= .50 173 2 166 9
5 190 173 17 170 20
6 205 = 100/8
175 = 12.5030 180 25
7 180 178 2 193 13
8 182 178 4 186 4
84 100
 SCA

Comparison of Forecast Error

∑ (forecast errors)2
Rounded Absolute Rounded Absolute
MSE = Actual Forecast Deviation Forecast Deviation
Tonage
n
with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1
For a180
= .10 175 5 175 5
2 = 1,558/8
168 = 194.758
176 178 10
3 159 175 16 173 14
4 For a175
= .50 173 2 166 9
5 190 173 17 170 20
6 = 1,612/8175=
205 201.5030 180 25
7 180 178 2 193 13
8 182 178 4 186 4
84 100
MAD 10.50 12.50
 SCA

n
Comparison of Forecast Error
100 ∑ |deviationi|/actuali
MAPE = i =Rounded
1 Absolute Rounded Absolute
Actual Forecast Deviation Forecast Deviation
Tonage with n for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1
a = .10 175
For 180 5 175 5
2 168 = 45.62/8
176 = 5.70%
8 178 10
3 159 175 16 173 14
4 a=
For 175 .50 173 2 166 9
5 190 173 17 170 20
6 205 = 54.8/8
175 = 6.85%
30 180 25
7 180 178 2 193 13
8 182 178 4 186 4
84 100
MAD 10.50 12.50
MSE 194.75 201.50
Comparison of Forecast Error
SCA

Rounded Absolute Rounded Absolute

Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1 180 175 5 175 5
2 168 176 8 178 10
3 159 175 16 173 14
4 175 173 2 166 9
5 190 173 17 170 20
6 205 175 30 180 25
7 180 178 2 193 13
8 182 178 4 186 4
84 100
MAD 10.50 12.50
MSE 194.75 201.50
MAPE 5.70% 6.85%
Comparison of Forecast Error
SCA

• Most computerized software automatically finds smoothing constant with lowest

error
• Drawback of using MSE is it tends to accentuate large deviations due to
squaring
• One problem with MAD and MSE is their values depend on magnitude of the
item being forecast
• MAPE avoids this problem and easiest to interpret