Course Objectives

1
QMETH 520 (Fall), 530 579 (Spring) MBA
Orientation, April 20, 2006 1230 to 120 PM in
BLM 307
2
Statement from Two Pioneers of Statistics for
Business
3
Two Methods of Forecasting
Model Based
Which is More Accurate?
Data
Intuition
4
Paul Meehls Study (1954)- A Legacy in Cognitive
Psychology
L et me emphasize the brute fact that we have
here, depending upon ones standards for
admission as relevant, from 16 to 20 studies
involving a comparison of expert and statistical
methods, in all but one of which the predictions
made statistically were either approximately
equal or superior to those made by an expert.
Clinical vs. Statistical Prediction, p.
119. (expert clinical, statistical actuarial)
5

• Limited Human Data Processing Capacity
• - Problems of Expert Judgment

6
Data AttributesProblematic to Intuition

• Randomness

2. Complex Series
3. Multiple Cues
7
Randomness - Counterintuitive
Human Intuition Is To Look for Regularity
Human Data Processing
8
2. Complex Series

9
3. Multiple Sources of Information
10
Objectives of 520, 530, 579

• 520 Statistical Modeling for Combining
  Multiple Cues for Forecasting - SPSS
• 530 Statistical Modeling of Complex Series
  for Forecasting Eviews
• 579 Practicum on Marketing Mix Modeling
  Using Data From Hudson River Group (230 weeks
  for 90 DMA)

11
520 Data An Example
12
Variety of DynamicsProblem that 530 focuses on
Example 1
Trend Seasonality Cycle Unequal variance
13
Example 2
Cycle
14
Example 3
Unequal Variance
15
Marketing Mix Modeling

• Marketing Mix Modeling is a multivariate
  statistical technique used to isolate the
  individual impact of simultaneously occurring
  marketing and other business drivers.

Television
Newspaper
Radio
Magazine
Outdoor
Direct Mail
Inserts
Clients Measure of Success
Events
Public Relations
Competitive Openings
External/Econometric Drivers
Category Trends
Competitive Advertising
Seasonality
16
Two Drivers for Enhancing Managers Data Analysis
Capability

• Standard Statistical Models
• Regression and related models
• Dedicated Efficient Software
• Eviews, SPSS, and others

17
Email from a Graduate
Having said that, most of the folks who have
graduated from the MBA program have the same
comment - we wish we could take more stats
classes in the school. Maybe you should suggest
adding more applicable stats in the curriculum )
18
A Memo From a Recent Graduate
I assume that most MBA students do not ultimately
wish to work as data analysts, but the underlying
skills are universally applicable and quite
valuable, regardless of their ultimate goals.
In multi-location and online marketing
organizations, the need is most common. These
companies often gather a wide variety of data on
the performance of their products, channels, and
customers. Whether their focus is on customer
segmentation, productivity of marketing
campaigns, or inventory planning, data skills are
in high demand.