project Facebook
Ronaldo777
ch9.pptSCM 304 Principles of Supply Chain Management
Copyright © 2019, 2016, 2013 Pearson Education, Inc. All Rights Reserved
1
Forecasting
Chapter 9
1
You will learn
Discuss the importance of forecasting
Select the most appropriate type of forecasting approach
Develop causal forecasting models using linear regression
Apply a variety of time series forecasting models, including moving average and exponential smoothing
Introduction (1 of 3)
You want to throw a party next weekend!
How much snacks?
How about planning for another party next year?!
3
Source: mememaker.net
Source: indiamart.com
Our life is full of uncertainties and so is the business!
How does a business know about the
demand for snacks for next 3 month?
Changes in behavior of consumer over time
All the supply chain operations depend on demand!
4
Introduction (2 of 3)
Source: indiamart.com
Decisions that need forecasts in the business:
Which markets to pursue?
What products to produce?
How many people to hire?
How many units to purchase?
How many units to produce?
5
Introduction (3 of 3)
Source: indiamart.com
Source: nohat.cc
Forecast – An estimate of the future level of some variables:
Demand
Supply
Price
Source: esportsgroup.net
Forecasting (1 of 2)
In a supply chain, forecasting is the first step to determine:
Long-term capacity needs
Yearly business plans
Shorter-term operations and supply chain activities
Example: Health center
Physical size of a new hospital
Number of doctors and nurses needed
Amount of supplies needed
Source: patientengagementhit.com
Source: nationalpost.com
Forecasting (2 of 2)
Forecast Types (1 of 3)
Most common forecast types in organizations:
Demand
Overall market demand vs. firm-level demand
Worldwide demand for new hybrid vehicles: 8 million by 2019.
Honda must decide the percentage of this to capture
Combines with warranty repairs and spare parts
8
Forecast Types (2 of 3)
Most common forecast types in organizations:
Supply
As important as demand forecasts!
Interruption in supply breaks the flow of goods and services to the final customer
Number of current producers and suppliers
Projected aggregate supply levels
Technological and political trends
9
Forecast Types (3 of 3)
Most common forecast types in organizations:
Price
Businesses need to forecast prices for key materials and services they purchase
Forward buying in expectation of price increase
Buy larger quantities and store in inventory
In expectation of price decrease
Buy more frequently in smaller quantities
E.g., How jet fuel prices can affect a wide range of decisions for airlines
10
Laws of Forecasting (1 of 4)
Laws of Forecasting to avoid the misapplication or misinterpretation of forecast results:
Law 1: Forecasts Are Almost Always Wrong (But They Are Still Useful).
Too many factors involved
Goal - get close estimates
Source: slate.com
Laws of Forecasting (2 of 4)
Law 2: Forecasts for the Near Term Tend To Be More Accurate.
Forecasting for the next month vs. for 10 or 20 years from now
Economic, political and technological changes
Source: indiamart.com
Laws of Forecasting (3 of 4)
Law 3: Forecasts for Groups of Products or Services Tend to Be More Accurate.
Demand for dark green cars vs. all cars
Color fashion may affect demand for green cars
Impact of color fashion disappears in overall demand – balanced out
Laws of Forecasting (4 of 4)
Law 4: Forecasts Are No Substitute For Calculated Values.
Don’t use forecasting if a better approaches to determining the variable is available!
E.g., weekly demand forecast of a rubber plant
Plan: How many products to be made in the coming week
Exact calculation is possible for how much raw rubber would be needed
How Samsung does demand forecasting
https://www.youtube.com/watch?v=Bsbp44DIC6U
15
Questions:
Inaccurate forecasts result in excess or shortage of inventory. why?
What factors affect demand forecasting?
16
Selecting a Forecasting Method (1 of 2)
Qualitative forecasting techniques – Forecasting techniques based on intuition or informed opinion.
Used when data are scarce, not available, or irrelevant.
Quantitative forecasting models – Forecasting models that use measurable or historical data to generate forecasts.
Time series and causal models
Selecting a Forecasting Method (2 of 2)
Consider two forecasting situations facing a large recording company:
Total music sales, including downloads and CDs, for the year
Music sales for a new recording artist
Question: which forecast method to use?
18
Qualitative Forecasting Methods (1 of 4)
Market surveys
Structured questionnaires submitted to potential customers
Solicit opinions about (potential) products
Effective if well-structured and represent population
Drawback: Expensive and time-consuming to perform
Source: dailydot.com
Source: evolvor.com
self-balancing electric device - hyperwalk
19
Qualitative Forecasting Methods (2 of 4)
Panel consensus forecasting
Experts come together to develop forecasts
Delphi method
Experts work individually to develop forecasts
Individual forecasts are shared among the group
Each participant is allowed to modify their forecast based on information from other experts
Drawback: time-consuming
Source: phoenix.edu
self-balancing electric device - hyperwalk
20
Qualitative Forecasting Methods (3 of 4)
Life-cycle analogy method
Used when the product or service is new
Many products and services have a fairly well-defined life cycle:
Introduction stage, growth stage, maturity stage, and decline stage
Identify the time frames and demand levels of different stages of new product or service
demand
Source: slideplayer.com
Qualitative Forecasting Methods (4 of 4)
Build-up forecasts
Experts familiar with specific market segments estimate the demand within these segments
These individual market segment forecasts are then added up to get an overall forecast
Each regional sales manager to estimate sale and then combined to estimate the overall sales
Quantitative forecasting models (1 of 2)
Causal Models:
Explores cause-and-effect relationships
Uses leading indicators to predict the future
A common tool of causal modeling is linear regression
23
| Variable | Cause of Change |
| Mortgage refinancing applications | Interest rates |
| Amount of food eaten at a party | Number and size of guests |
Linear Regression
A maker of golf shirts has been tracking the relationship between sales and advertising dollars. Use linear regression to find out what sales might be if the company invested $53,000 in advertising next year.
| Sales $ (Y) | Adv.$ (X) | |
| 1 | 130 | 48 |
| 2 | 151 | 52 |
| 3 | 150 | 50 |
| 4 | 158 | 55 |
| 5 | ? | 53 |
Sales ($)
Advertisement ($)
48
130
Sales ($)
Sales ($)
50
52
54
56
140
150
160
157
24
Quantitative forecasting models (2 of 2)
Time Series Forecasting Methods
Assumes the future will follow same patterns as the past
A quantitative forecasting model that uses a time series to develop forecasts.
Time series – A series of observations arranged in chronological order
25
Time Series Forecasting Methods (1 of 14)
Figure 9.3 Time Series of Weekly Demand at an Emergency Care Facility
Time Series Forecasting Methods (2 of 14)
Demand Patterns
Randomness – Unpredictable movement from one time period to the next.
Trend – Long-term movement up or down in a time series.
Seasonality – A repeated pattern of spikes or drops in a time series associated with certain times of the year.
Time Series Forecasting Methods (3 of 14)
Seasonality
Table 9.8 Examples of Products and Service That Experience Seasonality
| Product or Service | Peak Season(s) |
| Gasoline | Summer months, as more people are traveling |
| Caribbean cruises | Winter months |
| Cub Scout uniforms | Fall, as new scouts are joining up |
| Emergency medical care | Summer months, as more people are involved in outdoor activities |
| Fruitcake | November and December holiday season, after which no one buys it (or eats it) |
Time Series Forecasting Methods (4 of 14)
Figure 9.4 Time Series Showing Randomness, a Downward Trend, and Seasonality (Higher Demand in the Winter Months)
Time Series Forecasting Methods (5 of 14)
Last Period Model – The simplest time series model which uses demand for the current period as a forecast for the next period.
Time Series Forecasting Methods (6 of 14)
Table 9.3 Last Period Forecasting for an Emergency Care Facility
| Week | Number of Patients | Last Period Forecast |
| 1 | 84 | Blank |
| 2 | 81 | 84 |
| 3 | 89 | 81 |
| 4 | 90 | 89 |
| 5 | 99 | 90 |
| 6 | 106 | 99 |
| 7 | 127 | 106 |
| 8 | 127 |
Time Series Forecasting Methods (7 of 14)
[Table 9.3 Continued]
| Week | Number of Patients | Last Period Forecast |
| 9 | 127 | 117 |
| 10 | 103 | 127 |
| 11 | 96 | 103 |
| 12 | 96 | 96 |
| 13 | 86 | 96 |
| 14 | 101 | 86 |
| 15 | 109 | 101 |
| 16 | Blank | 109 |
Time Series Forecasting Methods (8 of 14)
Figure 9.5 Last Period Forecasting for an Emergency Care Facility
Time Series Forecasting Methods (9 of 14)
Moving Average Model – A time series forecasting model that derives a forecast by taking an average of recent demand values.
Time Series Forecasting Methods (10 of 14)
Table 9.4 Two-Period and Four-Period Moving Average Forecasts
| Week | Number of Patients | Two-Period Moving Average Forecast | Four-Period Moving Average Forecast |
| 1 | 84 | Blank | Blank |
| 2 | 81 | Blank | Blank |
| 3 | 89 | 82.5 | Blank |
| 4 | 90 | 85.0 | Blank |
| 5 | 99 | 89.5 | 86.0 |
| 6 | 106 | 94.5 | 89.8 |
| 7 | 127 | 102.5 | 96.0 |
| 8 | 117 | 116.5 | 105.5 |
| 9 | 127 | 122.0 | 112.3 |
| 10 | 103 | 122.0 | 119.3 |
Time Series Forecasting Methods (11 of 14)
[Table 9.4 Continued]
Example: Week 16 = (101 + 109)/2 = 105
| Week | Number of Patients | Two-Period Moving Average Forecast | Four-Period Moving Average Forecast |
| 11 | 96 | 115.0 | 118.5 |
| 12 | 96 | 99.5 | 110.8 |
| 13 | 86 | 96.0 | 105.5 |
| 14 | 101 | 91.0 | 95.3 |
| 15 | 109 | 93.5 | 94.8 |
| 16 | Blank | 105.0 | 98.0 |
| Average: | 100.7 | 101.0 | 102.6 |
| Minimum: | 81 | 82.5 | 86.0 |
| Maximum: | 127 | 122.0 | 119.3 |
Time Series Forecasting Methods (12 of 14)
Figure 9.6 Two-Period and Four-Period Moving Average Forecasts for an Emergency Care Facility
Time Series Forecasting Methods (13 of 14)
Weighted Moving Average Model – A form of the moving average model that allows the actual weights applied to past observations to differ.
Example 9.1 – Flavio’s Pizza
Flavio’s Pizza has recorded the following demand history for each Friday night for the past five weeks.
Develop forecasts for week 6 using a two-period moving average and a 3-period weighted moving average using the following demands and weights (0.4, 0.35, and 0.25, starting with the most recent observation).
| Week | Demand |
| 1 | 62 |
| 2 | 45 |
| 3 | 55 |
| 4 | 73 |
| 5 | 60 |
The two-period moving average forecast would be:
The three-period weighted moving average forecast would be:
Time Series Forecasting Methods (14 of 14)
Exponential Smoothing Model – A special form of the moving average model in which the forecast for the next period is calculated as the weighted average of the current period’s actual value and forecast.
Factors for Selecting a Forecasting Model
The amount and type of available data
Degree of accuracy required
Length of forecast horizon
Presence of data patterns
41
Selecting a Forecasting Method
Figure 9.2 Selecting a Forecasting Method
Forecasting Software (1 of 2)
Computer-based forecasting packages are used to develop, evaluate and change forecasting models as needed.
With enough demand history, a package could quickly evaluate alternative forecasting methods for each item and select the model that best fits the past data.
Forecasting Software (2 of 2)
Spreadsheets
Microsoft Excel, Quattro Pro, Lotus 1-2-3
Limited statistical analysis of forecast data
Statistical packages
SPSS, SAS, NCSS, Minitab
Forecasting plus statistical and graphics
Specialty forecasting packages
Forecast Master, Forecast Pro, Autobox, SCA
44
1
tt
FD
+
=
1
where:
forecastforthenextperiod,1
anddemandfor thecurrentperiod,
t
t
Ft
Dt
+
=+
=
1
1
1
n
ti
i
t
D
F
n
+-
=
+
=
å
1
1
where:
forecastfortimeperiod1
actualdemandforperiod1
numberofmostrecentdemandobservationusedt
odeveloptheforecast
t
ti
Ft
Dti
n
+
+-
=+
=+-
=
111
1
n
ttiti
i
FWD
++-+-
=
=
å
1
1
1
where:
weightassignedtothe demandinperiod1
1
ti
n
ti
i
Wti
W
+-
+-
=
=+=
=
å
6
(6073)/266.5pizzas
F
=+=
6
0.4600.35730.255563.3pizzas
F
=*+*+*=
