Operations Management

Sessions 11-12

Indian Institute of Management

Lucknow
 We’ll Discuss…
 Forecasts and their need

 Different forecasting approaches and methods

 Basic elements of forecasting: the choice of planning

horizon, different types of data patterns, and
calculating forecasting errors.

 Popular Methods of forecasting

 Managerial Implications

11/23/2021 OM 2
 What is Forecasting?
• Process of predicting a future event
(The art and science of predicting future events)
• Underlying basis of
all business decisions
• Production
??
• Sourcing
• Inventory
• Personnel
• Facilities
Businesses use forecasts to deal with the uncertainty of the future.
11/23/2021 OM 3
 Types of Forecasts
 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 or new product/

service
11/23/2021 OM 4
 Forecasting Time Horizons
 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 operations planning, budgeting

 Long-range forecast
+
◼ 3 years

◼ New product planning, facility location,

research and development
11/23/2021 OM 5
 Distinguishing Differences
 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
11/23/2021 OM 6
 Characteristics of Forecasts
 Forecasts are seldom perfect.

 Long-term forecasts are less accurate than short-term

forecasts (forecast horizon is important)

 Aggregate forecasts are more accurate than

disaggregate forecasts

 Most techniques assume an underlying system stability

 Should include expected value and measure of error.

11/23/2021 OM 7
 Influence of Product Life Cycle
Introduction – Growth – Maturity – Decline

Introduction and growth require longer forecasts

than maturity and decline

As product passes through life cycle, forecasts are

useful in projecting
Staffing levels

Inventory levels

Capacity
11/23/2021 OM 8
 Product Life Cycle
Introduction Growth Maturity Decline

Product design and Forecasting critical Standardization Little product

development differentiation
OM Strategy/Issues

Product and process Less rapid product

critical reliability changes – more Cost
Frequent product Competitive product minor changes minimization
and process design improvements and Optimum capacity Overcapacity in
changes options the industry
Increasing stability
Short production Increase capacity of process Prune line to
runs eliminate items
Shift toward product Long production
High production focus runs not returning
costs good margin
Enhance distribution Product
Limited models improvement and Reduce capacity
Attention to quality cost cutting

11/23/2021 OM 9
 Seven Steps in Forecasting
 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
11/23/2021 OM 10
 Forecasting Methods
 Qualitative: primarily subjective; rely on
judgment and opinion
◼ Used when situation is vague and little data
exist
⚫ New products
⚫ New technology
◼ Involves intuition, experience
⚫ e.g., forecasting sales through m-commerce
11/23/2021 OM 11
 Jury of Executive Opinion
• Involves small group of high-level managers

• Group estimates demand by working together

• Combines managerial experience with statistical

models

• Relatively quick

• ‘Group-think’
disadvantage
11/23/2021 OM 12
 Sales Force Composite

• Each salesperson projects his or her sales

• Combined at district and national levels

• Sales reps know customers’ wants

• Tends to be overly optimistic

11/23/2021 OM 13
 Delphi Method
Iterative group process,
continues until
consensus is reached

3 types of participants
Decision makers

Staff Respondents
(People who can make valuable judgments)

Respondents
11/23/2021 OM 14
 Consumer Market Survey

• Ask customers about purchasing plans

• What consumers say, and what they

actually do are often different

• Sometimes difficult to answer

11/23/2021 OM 15
 …Forecasting Methods
 Quantitative: used when environment is ‘stable’
and historical data exist and is amenable to
mathematical analysis
◼ Existing products

◼ Current technology

 Involves mathematical models

◼ e.g., forecasting sales of CFL Lamps
11/23/2021 OM 16
 Quantitative Forecasting Methods
 Causal: use the relationship between demand
and some other factor to develop forecast
◼ Linear
◼ Non-Linear
 Time Series: use historical demand only
◼ Static
◼ Adaptive
 Simulation
◼ Imitate consumer choices that give rise to demand
◼ Can combine time series and causal methods
11/23/2021 OM 17
 Overview of Quantitative Approaches

1. Naive approach

2. Moving averages Time-Series

Models
3. Exponential smoothing

4. Trend projection
Associative
5. Linear regression Model

11/23/2021 OM 18
 Components of Demand
 Average demand for a period of time

 Trend

 Seasonal element

 Cyclical elements

 Random variation

 Autocorrelation
11/23/2021 OM 20
 …Components of Demand

Trend
component
Seasonal peaks
Demand for product or service

Actual
demand

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

11/23/2021 OM 21
 Components of an Observed Demand
Observed demand (O) =
Systematic component (S) + Random component (R)
Level (current deseasonalized demand)

Trend (growth or decline in demand)

Seasonality (predictable seasonal fluctuation)

• Systematic component: Expected value of demand

• Random component: The part of the forecast that deviates
from the systematic component
• Forecast error: difference between forecast and actual demand
11/23/2021 OM 22
 Trend Component

Persistent, overall upward or downward

pattern
Changes due to population, technology, age,
culture, etc.
Typically several years duration

11/23/2021 OM 23
 Seasonal Component
 Regular pattern of up and down fluctuations
 Due to weather, customs, etc.
 Occurs within a single year

Number of
Period Length Seasons
Week Day 7
Month Week 4-4.5
Month Day 28-31
Year Quarter 4
Year Month 12
Year Week 52

11/23/2021 OM 24
 Cyclical Component

 Repeating up and down movements

 Affected by business cycle, political, and
economic factors
 Multiple years duration
 Often causal or
associative
relationships

0 5 10 15 20

11/23/2021 OM 25
 Random Component
 Erratic, unsystematic, ‘residual’
fluctuations
 Due to random variation or unforeseen
events
 Short duration and
non-repeating

M T W T F

11/23/2021 OM 26
 Time Series
A time series is a set of observations
measured at successive points in time or
over successive periods of time. A time
series pattern may have one or more of
the following five characteristics:
✓ Trend
✓ Seasonal
✓ Cyclical
✓ Random Variation
✓ Irregular (one time) Variation
11/23/2021 OM 27
 Time Series Forecasting
Quarter Demand Dt
II, 2017 8000
III, 2017 13000
How to
IV, 2017 23000
I, 2018 34000 Forecast
II, 2018 10000 demand
III, 2018 18000 for the
IV, 2018 23000
I, 2019 38000 next four
II, 2019 12000 quarters?
III, 2019 13000
IV, 2019 32000
I, 2020 41000
11/23/2021 OM 28
 …Time Series Forecasting

50,000
40,000
30,000
20,000
10,000
0

11/23/2021 OM 29
 Time Series

11/23/2021 OM 30
 Time Series Forecasting Methods
 Goal is to predict systematic component of
demand
◼ Multiplicative: (level)(trend)(seasonal factor)

◼ Additive: level + trend + seasonal factor

◼ Mixed: (level + trend)(seasonal factor)

 Static methods

 Adaptive methods
11/23/2021 OM 31
 …Time Series Forecasting Methods
 Static : assume a mixed static model

 Adaptive: estimates of level, trend, and seasonality

are adjusted after each demand observation
◼ Moving average

◼ Simple exponential smoothing

◼ Holt’s model (with trend)

◼ Winter’s model (with trend and seasonality)

11/23/2021 OM 32
 Moving Average
 Used when demand has no observable trend or seasonality
 Systematic component of demand = level
 The level in period t is the average demand over the last N
periods (the N-period moving average)
 Current forecast for all future periods is the same and is based
on the current estimate of the level
Lt = (Dt + Dt-1 + … + Dt-N+1) / N
Ft+1 = Lt and Ft+n = Lt
After observing the demand for period t+1, revise the
estimates as follows:
Lt+1 = (Dt+1 + Dt + … + Dt-N+2) / N
Ft+2 = Lt+1

11/23/2021 OM 33
 Moving Average Example
From the example shown earlier: 50,000
40,000
30,000
20,000
10,000
0

At the end of period 4, what is the forecast demand for periods 5

through 8 using a 4-period moving average?
L4 = (D4+D3+D2+D1)/4 = (34000+23000+13000+8000)/4 = 19500
F5 = 19500 = F6 = F7 = F8
Observe demand in period 5 to be D5 = 10000
Forecast error in period 5, E5 = F5 - D5 = 19500 - 10000 = 9500
Revise estimate of level in period 5:
L5 = (D5+D4+D3+D2)/4 = (10000+34000+23000+13000)/4 = 20000
F6 = L5 = 20000
11/23/2021 OM 34
 Problem 4.2

Week Demand
1 7 Starting in week 4 and going to
2 9 year 12, forecast demand
3 5 using a 3-week moving
4 9 average; plot the graph.
5 13 Starting in week 4 and going to
6 8 year 12, forecast demand
7 12
8 13
using a 3-week moving
9 9 average; with weights of
10 11 0.1,0.3 and 0.6 plot the graph.
11 7

11/23/2021 OM 35
Moving Average &Weighted Moving Averages

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
Figure 4.2
 Potential Problems With Moving Average
 Increasing n smooths the forecast but
makes it less sensitive to changes
 Do not forecast trends well
 Require extensive historical data
∑ (weight for period n)
Weighted = x (demand in period n)
moving average ∑ weights
Used when trend is present
Older data usually less important

11/23/2021 OM 37
 Weighted Moving Averages: Another Example
Weights Applied Period
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 [(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

11/23/2021 OM 38
 Exponential Smoothing – Problem 4.4

A cheque processing centre uses exponential

smoothing to forecast the number of incoming
cheques each month. The number of cheques
received in June was 40 million, while the forecast
was 42 million. A smoothing forecast of 0.2 is used.
(a) What is the forecast for July?
(b) If the centre received 45 million cheques in July,
what would be the forecast for August?
(c) Why might be this an inappropriate forecasting
method for this situation?

11/23/2021 OM 39
 Exponential Smoothing – Problem 4.4
(a) What is the forecast for July?
FJuly= FJune + 0.2 Forecasting Error
FJuly= 42 + 0.2 (40-42) = 41.6
If the centre received 45 million cheques in July,
what would be the forecast for August?
FAugust= FJuly + 0.2 Forecasting Error
FAugust= 41.6 + 0.2 (45-41.6) = 42.3
Why might be this an inappropriate forecasting
method for this situation?
The banking industry may have seasonality in its
processing requirements
11/23/2021 OM 40
 Exponential Smoothing– Problem 4.43
Emergency calls for Women helpline for past 24 weeks
are shown in the table below.
Week 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Actual
Value 50 35 25 40 45 35 20 30 35 20 15 40 55 35 25 55 55 40 35 60 75 50 40 65

Assuming initial forecast of 50 calls for the first week,

and using α=0.2, compute the exponentially smoothed
forecasts. What is the forecast for week 25?
Reforecast each period using α=0.6.
Actual calls during week 25 were 85. Which smoothing
constant provides a superior forecast?
11/23/2021 OM 41
 Trend Projections
Fitting a trend line to historical data points to
project into the medium to long-range
Linear trends can be found using the least squares
technique
^
y = a + bx
where y^ = computed value of the variable to be
predicted (dependent variable)
a = y-axis intercept
b = slope of the regression line
x = the independent variable

11/23/2021 OM 42
 Least Squares Method
Values of Dependent Variable

Actual observation Deviation7

(y value)

Deviation5 Deviation6

Deviation3 Least squares

method minimizes
Deviation4 the sum of the
squared errors
(deviations)
Deviation1
(error) Deviation2 ^ a + bx
Trend line, y =

Time period Figure 4.4

 Least Squares Method

Equations to calculate the regression variables

^
y = a + bx

Sxy - nxy
b=
Sx2 - nx2

a = y - bx
 Least Squares - Example
Time Electrical Power
Year Period (x) Demand x2 xy
2001 1 74 1 74
2002 2 79 4 158
2003 3 80 9 240
2004 4 90 16 360
2005 5 105 25 525
2005 6 142 36 852
2007 7 122 49 854
∑x = 28 ∑y = 692 ∑x2 = 140 ∑xy = 3,063
x=4 y = 98.86

∑xy - nxy 3,063 - (7)(4)(98.86)

b= = = 10.54
∑x - nx
2 2 140 - (7)(4 2)

a = y - bx = 98.86 - 10.54(4) = 56.70

11/23/2021 OM 45
 Least Squares - Example
Time Electrical Power
Year Period (x) Demand x2 xy
2009 1 74 1 74
2010 2 79 4 158
2011 3 80 9 240
2012 4 90 16 360
2013 5 105 25 525
2014 6 142 36 852
2015 7 122 49 854
Sx = 28 Sy = 692 Sx2 = 140 Sxy = 3,063
x=4 y = 98.86
Sxy - nxy 3,063 - (7)(4)(98.86)
b= = = 10.54
Sx2 - nx2 140 - (7)(42)
a = y - bx = 98.86 - 10.54(4) = 56.70
Trendline: y^ = 56.70 + 10.54x
11/23/2021 OM 46
 Least Squares - Example

Trend line: y^ = 56.70 + 10.54x

160 –
150 –
140 –
–
Power demand

130
120 –
110 –
100 –
90 –
80 –
70 –
60 –
50 –
| | | | | | | | |
2001 2002 2003 2004 2005 2006 2007 2008 2009
Year

11/23/2021 OM 47
 Multiple Regression Analysis
If more than one independent variable is to be used in
the model, linear regression can be extended to
multiple regression to accommodate several
independent variables
y^ = a + b x + b x …
1 1 2 2

Computationally, this is generally complex

and generally done using software

11/23/2021 OM 48
 Seasonality Index– Problem 4.27
Attendance (in thousands) for the newest Disneyland is
given below. Compute seasonal indices.
Quarter Year 1 Year 2 Year 3
Winter 73 65 89
Spring 104 82 146
Summer 168 124 205
Fall 74 52 98
Average
Average Quarterly Seasonal
Quarter 2003 2004 2005 Demand Demand Index Seasonal Indices
0.709
Winter 73 65 89 75.67 106.67 0.709 1.038
Spring 104 82 146 110.67 106.67 1.037
1.553
Summer 168 124 205 165.67 106.67 1.553
Fall 74 52 98 74.67 106.67 0.700
0.700
11/23/2021 OM 49
 Measures of Forecast Error
 Forecast error = Et = Ft - Dt

 Mean squared error (MSE)

MSEn = (Sum(t=1 to n)[Et2])/n

 Absolute deviation = At = |Et|

 Mean absolute deviation (MAD)

MADn = (Sum(t=1 to n)[At])/n

s = 1.25MAD
11/23/2021 OM 50
The MAD Statistic to Determine Forecasting Error

n
1 MAD  0.8 standard deviation
A
t=1
t - Ft
1 standard deviation  1.25 MAD
MAD =
n

 The ideal MAD is zero which would mean

there is no forecasting error

 The larger the MAD, the less the accurate the

resulting model

11/23/2021 OM 51
 …Measures of Forecast Error
 Mean absolute percentage error (MAPE)
MAPEn = (Sum(t=1 to n)[|Et/ Dt|100])/n
 Bias
 Shows whether the forecast consistently under- or
overestimates demand; should fluctuate around 0
biasn = Sum(t=1 to n)[Et]
 Tracking signal
TSt = bias / MADt
Should be within the range of +6; Otherwise, possibly use a
new forecasting method
11/23/2021 OM 52
 MAPE
n

å100 Actual i
- Forecast i / Actuali
MAPE = i=1
n
ACTUAL
TONNAGE FORECAST FOR ABSOLUTE PERCENT ERROR
QUARTER UNLOADED a = .10 100(|ERROR|/ACTUAL)
1 180 175.00 100(5/180) = 2.78%
2 168 175.50 100(7.5/168) = 4.46%
3 159 174.75 100(15.75/159) = 9.90%
4 175 173.18 100(1.82/175) = 1.05%
5 190 173.36 100(16.64/190) = 8.76%
6 205 175.02 100(29.98/205) = 14.62%
7 180 178.02 100(1.98/180) = 1.10%
8 182 178.22 100(3.78/182) = 2.08%
Sum of % errors = 44.75%

MAPE =
å absolute percent error 44.75%
= = 5.59%
n 8
 Monitoring and Controlling Forecasts

Tracking Bias
signal =
MAD

∑(actual demand in
period i -
forecast demand
Tracking in period i)
=
signal (∑|actual - forecast|/n)

11/23/2021 OM 54
 Tracking Signal

Signal exceeding limit

Tracking signal
Upper control limit
+

0 MADs Acceptable
range

–
Lower control limit

Time

11/23/2021 OM 55
 Tracking Signals
PERCENTAGE OF THE AREA OF THE NORMAL PROBABILITY DISTRIBUTION
WITHIN THE CONTROL LIMITS OF THE TRACKING SIGNAL

Control Limit Spread Equivalent Percentage of Area

(number of MAD) Number of s 2 within Control Limits

± 1.0 ± 0.80 57.62

± 1.5 ± 1.20 76.98
± 2.0 ± 1.60 89.04
± 2.5 ± 2.00 95.44
± 3.0 ± 2.40 98.36
± 3.5 ± 2.80 99.48
± 4.0 ± 3.20 99.86

11/23/2021 OM 56
 Tracking Signal Example
Cumulative
Tracking Absolute Absolute
Signal Forecast
Actual Forecast Forecast
Qtr(Bias/MAD)
Demand Demand Error RSFE Error Error MAD
-10/10 = -1
1
-15/7.5 90
= -2 100 -10 -10 10 10 10.0
2 0/10 =950 100 -5 -15 5 15 7.5
3-10/10115
= -1 100 +15 0 15 30 10.0
4
+5/11 100
= +0.5 110 -10 -10 10 40 10.0
5
+35/14.2125
= +2.5110 +15 +5 15 55 11.0
6 140 110 +30 +35 30 85 14.2

The variation of the tracking signal between -2.0

and +2.5 is within acceptable limits

11/23/2021 OM 57
 Basic Formula for Adaptive Forecasting
Ft+1 = (Lt + lT)St+1 = forecast for period t+l in period t
Lt = Estimate of level at the end of period t
Tt = Estimate of trend at the end of period t
St = Estimate of seasonal factor for period t
Ft = Forecast of demand for period t (made period t-1 or
earlier)
Dt = Actual demand observed in period t
Et = Forecast error in period t
At = Absolute deviation for period t = |Et|
MAD = Mean Absolute Deviation = average value of At
11/23/2021 OM 58
 General Steps in Adaptive Forecasting
1. Initialize: Compute initial estimates of level (L0),
trend (T0), and seasonal factors (S1,…,Sp). This is
done as in static forecasting.
2. Forecast: Forecast demand for period t+1 using the
general equation
3. Estimate error: Compute error Et+1 = Ft+1- Dt+1
4. Modify estimates: Modify the estimates of level
(Lt+1), trend (Tt+1), and seasonal factor (St+p+1), given
the error Et+1 in the forecast
5. Repeat steps 2, 3, and 4 for each subsequent period

11/23/2021 OM 59
 Web-Based Forecasting: CPFR
 Collaborative Planning, Forecasting, and
Replenishment (CPFR) is a Web-based tool used to
coordinate demand forecasting, production and
purchase planning, and inventory replenishment
among supply chain partners.
 Used to integrate the multi-tier or n-Tier supply
chain, including manufacturers, distributors and
retailers.
 CPFR’s objective is to exchange selected internal
information to provide for a reliable, longer term
future views of demand in the supply chain.
 It uses a cyclic and iterative approach to derive
consensus forecasts.
11/23/2021 OM 60
 What is CPFR?
• CPFR is a process used to achieve better
supply chain coordination.
• It is practiced by SC partners for demand
creation and demand fulfillment activities.
• CPFR uses EDI, Internet and other ICT
tools to share information about
o Promotions
o Forecasts
o Item data
o Orders
11/23/2021 OM 61
CPFR as a set of Process and Technology Models

11/23/2021 OM 62
 CPFR Business Practices

ANALYSIS PLANNING /
STRATEGY

DEMAND &
EXECUTION SUPPLY
MANAGEMENT

11/23/2021 OM 63
 Forecasting in Practice

 Managers use a variety of judgmental and quantitative

forecasting techniques.

 Statistical methods alone cannot account for such

factors as sales promotions, competitive strategies,
unusual economic disturbances, new products, large
one time orders, natural disasters or labour
complications.

11/23/2021 OM 64
 Popular Methods of Forecasting
Moving
Averages/
Exponential Boosted

Intelligence
Algorithms
Smoothing LSTM Decision

Artificial
Trees
Seq2Seq
Support
ARIMA/ ARFIMA LSTM
Vector
ARCH Regression
Bayesian and Random Lasso
Approach variants Forest Regression GRU Dilated
CNN

1950s early 1970s late 1970s 1980s 1995 1996 1997 1999 2014 2016 2020

Statistical Classic Machine Learning Deep Learning Surge of AI

Methods OM
methods

11/23/2021 65
 Open Source Forecasting Packages

LinkedIn Open-Sources ‘Greykite’

11/23/2021 OM 66
 Managerial Implications
 Understand the purpose, time horizon, and level of
aggregation in developing a practical forecast .
 Be sure to distinguish between demand and sales
 Different forecasting methods require different levels of
technical ability and understanding of mathematical
principles and assumptions.
 Collaborate in building forecasts
 The value of data depends on where you are in the
supply chain

11/23/2021 OM 67
 Basic Approach to Demand Forecasting
 Understand the objectives of forecasting
 Identify major factors that influence the forecast
 Integrate demand planning and forecasting
 Understand and identify customer segments
 Determine the appropriate forecasting technique
 Establish performance and error measures for
the forecast
11/23/2021 OM 68
 Forecasting in the Service Sector

 Presents unusual challenges

◼ Needs differ greatly as function of

industry and product

◼ Special need for short term records

◼ Holidays and other calendar events

◼ Unusual events

11/23/2021 OM 69