Académique Documents
Professionnel Documents
Culture Documents
ISMT 101
Business implications
• Why do people Mgmt
make bad decisions?
Organizational use
• How do people Major IS (1) (1) Digital EB
make decisions? Economy (7)
• How can IT help Decision
- DSS Making (8)
- Data warehouse
- Data mining
HW SW Network Internet DB
(2) (3) (5) (6) (4)
Decision Making Today
• Biz decisions are increasingly difficult to make
– Globalization
• Face very diversified environments (political, economic,
cultural…)
– The speed of commerce
• Need to respond to the changing environment quickly
– The increasing number of business choices available
• E.g., B2B data exchange platform:
paper-based systems private EDI industry-wide EDI
national EDI Web-based EDI Internet
Decision Making Today
Rich Data …
– The use of multiple channels to communicate
with consumers
(e.g., email, call center, field interview, mail…)
– The explosion of computers’ processing
power and data storage capabilities
(Recall Moore’s law)
Decision Making Today
• In 80’s, Apple was targeting the giant IBM as the major rival
• Microsoft was a 100-employee small company
• Application SW companies develop SW based on Microsoft’s OS
• Microsoft was one of the major reasons for Apple’s failure in the
PC market.
Smart people, bad decision
• Mini Case: Iridium
Iridium
• Global telephone service via low-earth-orbiting
(LEO) satellites
• First envisioned in 1985 by a Motorola engineer
– After his wife complained she couldn’t reach clients via
her cell phone.
• Idea
– 66 LEO satellites cover the entire area of the earth
• Total project cost: $5 billion
• Strong management team
– CEO Ed Staiano led Motorola
– His division at Motorola accounted for 40% of total
sales of $27 billion
Up and Down
• The peak in 1998
– Star at wall street, stock price tripled in one year
– Over 1,000 patents
• The Drop in 1999
– Customers
• April: 10,000, only 5% of projected number
– CEO resigned
– Filed for bankruptcy in Aug
– One of the 20 largest bankruptcies in US history
Mistakes Smart People Make
• 11 years passed through the development period,
the world has changed
2. Intelligence
3. Design
4. Choice
Intelligence Phase
• To detect a problem by scanning the
environment.
• To determine if decision-maker can solve
the problem.
– Within scope of influence
• E.g., THE 1997-98 ASIAN FINANCIAL CRISIS vs.
Your supplier raises price
• To gather more information to achieve
deeper understanding of the problem.
Design Phase
• To develop a model of the problem
– Determine type of model
– Verify model
• To develop and analyze potential solutions
Choice Phase
• Select the solution and implement
– Do more detailed analysis of the selected
solutions
– Verify initial conditions to ensure none of the
problem’s major factors has changed
• Recall the Iridium case
– Analyze proposed solution against real-world
constraints
A mini case
• Promoting m-commerce
(m-commerce – using of mobile devices
to exchange information)
• Intelligence
– Problem: m-commerce is developing slow
• Design
– Model: communication reach, richness, and costs
• Choice
– The government’s choice: adopt 3G
– The vendors’ choice: wait, or do m-commerce with lower
requirement on richness (e.g., securities, sell tickets…)
How IT help decision making
• Decision support systems (DSS) are one tool
– Definition: A computer-based system that
supports and improves human decision making
– Function: Help process huge amounts of data,
and analyze complex problems
• Data warehouse – store large amounts of historical
data
• Data mining – extract patterns from data
Decision Support Systems
• Major components
– Data management system
• DBMS is useful
– Model management system
• Typically mathematical in nature
– User interface
• Data visualization is the key
Data Visualization Examples
Data Visualization Examples
Data Visualization Examples
Data Mgmt for Decision Making
• Data warehouse — Database that is
– Subject-oriented – data organized around subjects
• DB in student account office and registrar’s office is
function-oriented
• In a data warehouse, data are organized around
students
– Integrated – contains ALL data about the subject
• Integrate data of different formats
(e.g., 2004-03-23, 03-23-2004, 04/Mar/23…)
• Integrate different database models
(e.g., relational, network, table)
Data Mgmt for Decision Making
• Data warehouse — Database that is
– Time-variant – database contains a time component,
that is, historical data
• Transactional databases are accurate at a given time
• Data warehouse contains data over multiple time periods
• e.g. a student data warehouse would contain data on what
students were registered in which classes for every term
covered by the data warehouse
– NOT a transactional database
• E.g., milk price in a transactional database is always the
newest (i.e., only one price at a time)
Modeling Tools and Techniques
• Simulation
– Computerized model of the problem
– Used to examine proposed solutions and their impacts
• Sensitivity analysis
– Determine how changes in one part of the model
influence other parts of the model
• What-if analysis
– Manipulate independent variables to see what would
happen in the dependent variable
• Goal-seeking analysis
– Work backward from desired outcome (dependent var)
Sensitivity analysis
Data Mining
• Process of applying analytical and statistical
methods to data to find patterns
Examples of patterns
– Credit card issuers:
• Given a database of 10M names, which persons are
the least likely to default on their credit cards?
• Which types of transactions are likely to be fraud?
– Retailers:
• Which of my customers are likely to be the most loyal,
and which are most likely to switch to competitors?
Data mining mini case
• Price and product differentiation
– Which market segment is cash cow?
– E.g., airline tickets – economy, business and first class
– 80/20 rule (80% of profits come from 20% of customers)
(at Harrah’s 82/26)
Applications
Any area where large amounts of historic data, if
understood better, can help shape future decisions.
DSS