Post on 27-Dec-2015
TM 745 Forecasting for Business & Technology
Paula Jensen
South Dakota School of Mines and Technology, Rapid City
1st Session 1/11/2012: Chapter 1 Introduction to Business Forecasting
Agenda Class Overview/Syllabus highlights Assignment Chapter 1 by Guest Lecturer Dr.
Stuart Kellogg
Business Forecasting 6th Edition J. Holton Wilson & Barry KeatingMcGraw-Hill
Instructor Information
Instructor Paula Jensen
Office Location IE/CM 320
Office Hours CM 320 M,W 2:00-3:00 pmIER T,TH, F 11:00-11:50 AME-mail for an appointment outside of office hours.
Office Phone 605-394-1770
E-mail paula.jensen@sdsmt.eduWebsite pjensen.sdsmt.edu
Course Materials Powerpoints & Class Information
Website: pjensen.sdsmt.edu via the ENGM 745
Engineering Notebook – 9-3/4" x 7-1/2", 5x5 quad-ruled, 80-100 pp. (approx.)
Engineering/Scientific calculator Book: Business Forecasting 6th Edition J.
Holton Wilson & Barry KeatingMcGraw-Hill
One case from Harvard Business Review
Prerequisites
1) Probability and Statistics2) Understanding of
Excel/Spreadsheet software.3) It is expected that students will be
able to access and download internet files.
Course Objective to educate prospective managers about
the philosophies and tools of sound forecasting principles
to provide technical managers with a theoretical basis for statistical forecasting
to provide technical managers with the fundamentals methods available for technological and qualitative forecasts
Evaluation Procedures
60% - 2 Exams20% - 1 Project20% - Interaction
A 90-100B 80-89C 70-79D 60-69F < 60
Exams Students signed up for the on-campus
section are required to take the test at the given time.
Make-up Exams available for University-Approved reasons.
All exams are open engineering notebook, and use of a scientific calculator is encouraged.
Distance Students need proctors- See Syllabus for further details
Project & Interaction Grades
Project Criteria to be discussed through Class
Interaction Assignments will include discussions, quizzes, and other assignments
Email Policy:
If you are writing about issues relating to the class, make sure the subject line reads ENGM 745: (subject info) so I can sort my e-mails and answer accordingly.
Please be professional in your e-mails. (no texting lingo!)
Academic Honesty
Cheating: use or attempted use of unauthorized materials, information or study aids
Tampering: altering or interfering with evaluation instruments and documents
Fabrication: falsification or invention of any information
Assisting: helping another commit an act of academic dishonesty
Plagiarism: representing the words or ideas of another as one's own
ADA
Students with special needs or requiring special accommodations should contact the instructor and/or the campus ADA coordinator, Jolie McCoy, at 394-1924 at the earliest opportunity.
First Assignment
Send me a contact info e-mail. Include all important contact information phones, e-mail, and mail addresses. Preferred mode.
Send via e-mail a Current Resume Problems 1,4, & 8 in chapter 1 – I don’t
need these sent. I will post solutions.
Introduction to Business Forecasting
Quantitative Forecasting Has Become Widely Accepted
Intuition alone no longer acceptable.
Used in Future Sales Inventory needs Personnel requirements
Judgment still is needed
Forecasting in Business Today
Two Professional Societies Accountants: costs, revenues (tax
plans) Personnel: recruitment, changes in
workforce Finance: cash flows Production: raw-material needs &
finished goods inventory Marketing: sales
Forecasting in Business Today
mid-80’s 94% large American firmsused sales forecasts
Krispy Kreme New stores model with errors of < 1%
Bell Atlantic Data warehouse (shared) of monthly
history Subjective, regression, time series, Forecasts monitored & compared
Forecasting in Business Today
Columbia Gas (natural gas company) Design Day Forecast (supply)
Gas supply, transportation capacity, storage capacity, & related
Daily Operational (demand) Regression on temperatures,
wind speed, day of the week, etc.
Forecasting in Business Today
Segix Italia (Pharmaceutical company) Marketing forecasts for seven main drugs Targets for sales representatives
Pharmaceuticals in Singapore Glaxo-Wellcome, Bayer, Pfizer,
Bristol-Myers Squibb HR, Strategic planning, sales Quantitative & judgments
Forecasting in Business Today
Fiat Auto (2 million vehicles annually) All areas use centrally prepared forecasts Use macro-economic data as inputs From totals sales to SKU’s
Douglas Aircraft Top down (miles flown in 32 areas) Bottom up (160 Airlines studied)
Forecasting in Business Today
Trans World Airlines Uses a top down (from total market)
approach for sales Regression & Trend models
Brake Parts Inc. 250,000 SKU’s Forecast system saves $6M/mo. 19 time series methods
Forecasting in the Public and Not-for-Profit Sectors
Police calls for service by cruiser district
State government Texas: Personal income, electricity sales,
employment, tax revenues California: national economic
models, state submodel, tax revenues, cash flow models
Hospitals: staff, procedures,
Collaborative Forecasting Manufacturer’s forecast > Retailers
Retailer’s extra info > Manufacturers1. Lower Inventory2. Fewer unplanned shipments or runs3. Reduced Stockouts4. Increase customer satisfaction5. Better sales promotions6. Better new product intros7. Respond to Market changes
Computer Use and Quantitative Forecasting Computer use common by mid 80’s Packages run from $100 to thousands PC systems generally have replaced
mainframes for state government work PC’s dominant at conferences Chase of Johnson & Johnson
Forecasting 80% math, 20% judgment
Subjective Forecasting Methods
Only way to forecast 40 years out Sale-Force Composite
Inform sales staff of data Bonus for beating the forecast ??
Surveys of Customers/Population Jury of Executive Opinion The Delphi Method (Experts)
New-Product Forecasting
A special consideration Surveys Test marketing ( Indy, K-zoo, not
KC) Analog Forecasts: movie toys
New Product Short Life Cycle
New Product Short Life Cycle
New Product Short Life Cycle
Product Life Cycle
Bass Model
Two Simple Naive Models (4th)
Two Simple Naive Models (4th)
Evaluating Forecasts
Evaluating Forecasts
Evaluating Forecasts
Measurement ErrorsStandard Deviation
SX nX
nt
( )
.
.
2
1
0 408
11
0193
Soda Demand (1,000,000's)
Month t At At2
(At-Abar)2
Jul 1 2.47 6.1009 0.002934Aug 2 2.31 5.3361 0.011201Sep 3 2.24 5.0176 0.030917Oct 4 2.27 5.1529 0.021267Nov 5 2.15 4.6225 0.070667Dec 6 2.34 5.4756 0.005751Jan 7 2.23 4.9729 0.034534Feb 8 2.48 6.1504 0.004117Mar 9 2.46 6.0516 0.001951Apr 10 2.58 6.6564 0.026951May 11 2.74 7.5076 0.105084Jun 12 2.72 7.3984 0.092517
Sum = 78 28.99 70.44 0.408
Avg = 6.5 2.416
St. Dev = 0.193
Measurement ErrorsStandard Deviation
SX nX
nt
2 2
2
1
70 44 12 2 416
11
0193
. ( . )
.
Soda Demand (1,000,000's)
Month t At At2
(At-Abar)2
Jul 1 2.47 6.1009 0.002934Aug 2 2.31 5.3361 0.011201Sep 3 2.24 5.0176 0.030917Oct 4 2.27 5.1529 0.021267Nov 5 2.15 4.6225 0.070667Dec 6 2.34 5.4756 0.005751Jan 7 2.23 4.9729 0.034534Feb 8 2.48 6.1504 0.004117Mar 9 2.46 6.0516 0.001951Apr 10 2.58 6.6564 0.026951May 11 2.74 7.5076 0.105084Jun 12 2.72 7.3984 0.092517
Sum = 78 28.99 70.44 0.41
Avg = 6.5 2.416
St. Dev = 0.193
Measurement ErrorsMAE
| |
| 2.47 2.416 | | 2.31 2.416 | ...
12
0.159
tX XMAE
n
Soda Demand (1,000,000's)
Month t At |At - Abar|
Jul 1 2.47 0.0542
Aug 2 2.31 0.1058
Sep 3 2.24 0.1758
Oct 4 2.27 0.1458
Nov 5 2.15 0.2658
Dec 6 2.34 0.0758
Jan 7 2.23 0.1858
Feb 8 2.48 0.0642
Mar 9 2.46 0.0442
Apr 10 2.58 0.1642
May 11 2.74 0.3242
Jun 12 2.72 0.3042
Sum = 78 28.99 1.91
Avg = 6.5 2.416 0.159
St. Dev = 0.193
Measurement ErrorsMAE
| |
| 2.47 2.416 | | 2.31 2.416 | ...
12
0.159
tX XMAE
n
Soda Demand (1,000,000's)
Month t At |At - Abar|
Jul 1 2.47 0.0542
Aug 2 2.31 0.1058
Sep 3 2.24 0.1758
Oct 4 2.27 0.1458
Nov 5 2.15 0.2658
Dec 6 2.34 0.0758
Jan 7 2.23 0.1858
Feb 8 2.48 0.0642
Mar 9 2.46 0.0442
Apr 10 2.58 0.1642
May 11 2.74 0.3242
Jun 12 2.72 0.3042
Sum = 78 28.99 1.91
Avg = 6.5 2.416 0.159
St. Dev = 0.193
In general,
0.8(.193) = 0.154
0.8MAE S
Measurement ErrorsSoda Demand (1,000,000's)
Month t At (At - Ahat)
Jul 1 2.47 0.0542
Aug 2 2.31 -0.1058
Sep 3 2.24 -0.1758
Oct 4 2.27 -0.1458
Nov 5 2.15 -0.2658
Dec 6 2.34 -0.0758
Jan 7 2.23 -0.1858
Feb 8 2.48 0.0642
Mar 9 2.46 0.0442
Apr 10 2.58 0.1642
May 11 2.74 0.3242
Jun 12 2.72 0.3042
Sum = 78 28.99 0.00
Avg = 6.5 2.416 0.000
St. Dev = 0.193
Mean Error
MEe
nt
( . . ) ( . . ) ...
.
2 47 2 416 2 31 2 416
12
0 0
Measurement ErrorsSoda Demand (1,000,000's)
Month t At Ft et |et| et2
Jul 1 2.47 2.416 0.054 0.054 0.003
Aug 2 2.31 2.416 -0.106 0.106 0.011
Sep 3 2.24 2.416 -0.176 0.176 0.031
Oct 4 2.27 2.416 -0.146 0.146 0.021
Nov 5 2.15 2.416 -0.266 0.266 0.071
Dec 6 2.34 2.416 -0.076 0.076 0.006
Jan 7 2.23 2.416 -0.186 0.186 0.035
Feb 8 2.48 2.416 0.064 0.064 0.004
Mar 9 2.46 2.416 0.044 0.044 0.002
Apr 10 2.58 2.416 0.164 0.164 0.027
May 11 2.74 2.416 0.324 0.324 0.105
Jun 12 2.72 2.416 0.304 0.304 0.093
Sum = 78 28.99 28.99 0.00 1.91 0.41
Avg = 6.5 2.416 2.416 0.000 0.159 0.034
St. Dev = 0.193
Using Multiple Forecasts
Use judgment Reference:
Combining Subjective andObjective Forecasts.
Sources of Data
Internal records Timeliness & formatting problems
Government & syndicated services (good)
Web Used by gov’t & syndicated Sites changes
Domestic Car Sales (4th ed ex.)
Domestic Car Sales (4th ed ex)
Domestic Car Sales (4th ed ex)
Forecasting FundamentalsSoda Demand (1,000,000's)
Month t At
Jul 1 2.47
Aug 2 2.31
Sep 3 2.24
Oct 4 2.27
Nov 5 2.15
Dec 6 2.34
Jan 7 2.23
Feb 8 2.48
Mar 9 2.46
Apr 10 2.58
May 11 2.74
Jun 12 2.72
Consider the followingsales data over a 12 month period.
Summary StatisticsSoda Demand (1,000,000's)
Month t At
Jul 1 2.47
Aug 2 2.31
Sep 3 2.24
Oct 4 2.27
Nov 5 2.15
Dec 6 2.34
Jan 7 2.23
Feb 8 2.48
Mar 9 2.46
Apr 10 2.58
May 11 2.74
Jun 12 2.72
XX
nt
2 47 2 31 2 72
12
2 42
. . ... .
.
Mean
Summary StatisticsMedian
Sorted Demand
t At
5 2.15
7 2.23
3 2.24
4 2.27
2 2.31
6 2.34
9 2.46
1 2.47
8 2.48
10 2.58
12 2.72
11 2.74
Xm
2 34 2 46
2
2 40
. .
.
Summary StatisticsSoda Demand (1,000,000's)
Month t At
Jul 1 2.47
Aug 2 2.31
Sep 3 2.24
Oct 4 2.27
Nov 5 2.15
Dec 6 2.34
Jan 7 2.23
Feb 8 2.48
Mar 9 2.46
Apr 10 2.58
May 11 2.74
Jun 12 2.72
Mode
No number repeats no mode
Summary StatisticsModal Range
Sorted Demand
t At
5 2.15
7 2.23
3 2.24
4 2.27
2 2.31
6 2.34
9 2.46
1 2.47
8 2.48
10 2.58
12 2.72
11 2.74
2.31 - 2.47
Summary Statistics
Modal Range
2.5 to 3.0
Soda Sales
0
10
20
30
40
0.5 1.0 1.5 2.0 2.5 3.0 More
Volume
Fre
qu
en
cy
Overview of the Text Ch 1 Intro Ch 2 Forecast Process (more Intro) Ch 3 MA & Exponential Smoothing Ch 4 Regression Ch 5 Multiple Regression Ch 6 Time-Series Decomposition Ch 7 ARIMA Box-Jenkins Ch 8 Combining Forecasts Ch 9 Forecast Implementation
Upcoming Events No Class next week Figure out what your log-in/password is
to D2l if you have not yet. It is the same as WebAdvisor - (Here is the website for D2L: https://d2l.sdbor.edu/)
Watch U-tube videos posted on Website Discussions on D2L- Ready 1/20/2012 Read Chapter 2 for Class on 1/25/2012