MIS 315

Data Analysis and Decision Modeling

A decision is a Solution is a means of
choice from solving a problem or
available dealing with a difficult
alternatives. situation.
Typical Business Decision Aspects
Several, possibly contradictory objectives
Many alternatives
Unevaluated alternatives
Results can occur in the future
Attitudes towards risk
Too much information- gathering information takes time and expense
“What-if” analysis, Scenarios
Trial-and-error experimentation may result in a loss
Experimentation with the real system - only once
Changes in the environment can occur continuously
A decision is a Modeling is a process A mathematical model is
choice from where you abstract the a quantitative
available essence of a real representation, or
alternatives. problem into a model. idealization, of a real
problem
 “The muscles of mathematics are connected
to the bones of experimental science by the
tendons of mathematical modeling.”

Glenn Ledder
The Modeling Approach to Decision Making
Anyone can use models to make decisions.

Types of models:
Mental (arranging furniture, semester schedule)
Visual (road maps)
Physical/Scale (aerodynamics, buildings)
Mathematical (what we’ll be studying)
Who Uses the Modeling Most? Consulting Services?
Best Consulting Companies
to Work For
Skill Set for Consulting

Skills Required How are they How to

tested? prepare?

PROBLEM SOLVING Case Interviews, Case

a. Analytical abilities Guesstimates, discussion
b. Business acumen Response to hints in groups
c. Ability to listen and learn interviews Case books

PERSONAL IMPACT Confidence in

a. Presence an interview Mock
b. Communication skills Is it enjoyable to talk Interviews
c. Teamwork abilities to you?

LEADERSHIP AND DRIVE

Your CV Cannot be
a. Integrity and Maturity
Highlights prepared
b. Track record
Let’s make a quick test
On LMS
McKinsey Problem Solving Test
Why are we here ?
We will understand and experience as much as possible
What is decision modeling?
What types of business problems may need/benefit from
quantitative modeling?
What types of using quantitative methods can be used:
Optimization modeling: LP, IP, Non-linear etc.
Simulation modeling
Risk modeling
In this course
We will concentrate on
quantitative models
solution techniques
many successful and realistic management science applications
Excel spreadsheets to solve problems
employed by external and internal consultants.,

In other words, study practical management science..

 What is Management Science? (MS)

Getting people together to Organized body of

accomplish desired goals knowledge that is derived
and objectives using from such observations and
that can be verified or
available resources tested by further
efficiently and effectively. investigation

A field of study that can get help from mathematics, computers,

and statistics to solve business problems.
Also known as:
Operations Research
Decision Science
Quantitative Methods
Managerial Decision Making
What about “Business Analytics”?
Analytics is the extensive use of
▪ data,
▪ statistical and quantitative analysis,
▪ explanatory and predictive models, and
▪ fact-based management to drive decisions and actions.
So, Management Science with an additional “big data” portion
Why do businesses need MS to make a decision?

Decision making:
Art vs.
Science?
Increasing complexity of organizational environment
(Market/Supplier/Competitor/Customer/Technology/....).
Numerous factors must be considered simultaneously due to
their interactions (Marketing/Finance/R&D/.......).
Explosive information processing and analysis requirements
before decision-making.
Number of alternatives and outcomes to be evaluated.
Need to make sensitivity (what if ) analyses
Increase in Organizational, Business and
Industrial Process’ Complexity
When you complete this course,
You should think logically.
Your quantitative skills will be sharpened immensely.
You will be a proficient spreadsheet user.
You will develop your intuition, and see that intuition alone
sometimes fails.
Flow of the course
PART I – Basics of Modeling in Decision Making
PART II – Optimal Decision Making Under Certainty
PART III – Decision Making Under Uncertainty
PART IV – Modeling Preferences
Course Syllabus
Instructor: Masoud Shahmanzari
Office : AB2 350 or ZOOM
Phone : +90 (216) 564 9438
E-mail : [email protected]
Syllabus
Learning resources:
Required Textbook: Practical Management Science, 6th edition,
by Wayne Winston and Chris Albright, 2019. ISBN:
9781337406659.(E-Text)
Articles:
A reading list will be provided.
Syllabus
Hardware & Software: Effective learning is best supported by having students
actively engaged in the learning process.
The Add-In Solver for Microsoft Excel for Windows will be used extensively; to
see if your copy of Excel has Solver, pull down the “Tools” header on the top of
your screen and check if there is a command “Solver”. If not, you should invoke
“Add-Ins”, and check if Solver is among the available add-ins. If not, you should
reinstall Excel fully.
Other Excel add-ins will be obtained from the textbook support material.
Article Presentations
Successful management science applications are reported in
the INFORMS Journal on Applied Analytics (earlier called
Interfaces) journal.
The applications are in a wide range of industries, they are
global, and they use a variety of management science
techniques and have generated proven benefits, often even
hundreds of millions of dollars.
Each group (of 2) will choose an article from the reading list
and then present a summary of it during class time that will
be spread out over the semester.
It will be around 15 minutes (excluding Q&As). By the
beginning of the semester, each group will choose an article
from a list that will be posted on LMS.
For example:
Amaral, Jason and Dorothea Kuettner, “Analyzing Supply Chains at HP Using
Spreadsheet Models,” Interfaces, 28(4) (July-August 2008), pp. 228-240.
Computer Exercises: You are expected to attend class sessions,
arrive on time, and stay in the room for the entire session.

Homework & Quizzes: Some problems will be recommended

after completion of each subject. There will be no make-up for
quizzes missed. When computing the average grade of quizzes,
the one with the lowest grade will be dropped.
What is Quantitative Decision Modeling?
Analyses

Quant.Analysis
Logic
Historic Data
Marketing Research
Problem Scientific Analysis Decision
Modeling

? Qual. Analysis
Weather
State and federal
legislation
New technological
breakthroughs
Election outcome
Mathematical model
A mathematical model is a quantitative representation, or
idealization, of a real problem.
It can be phrased in terms of
mathematical expressions (equations and inequalities)
or a series of interrelated cells in a spreadsheet.
Provides several advantages:
Enables managers to understand the problem better;
Helps to define the scope of the problem, the possible solutions,
and the data requirements
Mathematical Models Characterized by Risk
Deterministic models - we know all values used in the model with certainty (Parts I
and II of the course)
Probabilistic models - we know the probability that parameters in the model will take
on a specific value (Parts III and IV of the course)
Model types

Descriptive models: models that simply describe a situation.

Optimization models: models that suggest a desirable course of
action.
Example
Capital Budgeting Problem on LMS
Descriptive model (1 of 3)
A company faces capital budgeting decisions, and there are
seven potential investments. Each has an investment cost
and a corresponding stream of cash flows (including the
investment cost) summarized by a net present value (NPV).
The problem is on LMS called “Capital_Budgeting”
Please work in groups to answer
So, which investments should be taken?
Please submit your worked file.
Descriptive model (2 of 3)
A descriptive model can take at least two forms.
One form is to show all of the elements of the problem in a diagram, as below.
This method helps the company to visualize the problem and to better understand
how the elements of the problem are related.
Descriptive model (3 of 3)
Although the diagram helps the company visualize the problem, it does not
provide any numeric information.
This can be accomplished with the second descriptive form of the model as
below
Optimization model (1 of 2)
The company uses Solver, an add-in to Excel, to quickly search
through all potential sets of decisions and find the one that
maximizes total NPV while staying within the budget.
After creating the descriptive model, they invoke Solver,
where the company can specify the objective cell, the range of
decision variable cells, and any constraints. Then Solver finds
the optimal solution, usually in a matter of seconds.
Optimization model (2 of 2)
Optimal solutions are sometimes surprising.
You might have expected the company to choose the investments with the largest
ROIs, at least until the budget runs out.
However, this is not what the optimal solution prescribes.
For example, investment 4 has the largest ROI, but it is not chosen, and investment 5,
which is chosen, has a lower ROI than three of the investments that are not chosen.
Incorporating uncertainty
This takes us to simulation models. Simulation lets the
company see how an output such as total NPV varies, for any
given set of decisions, as uncertain inputs vary over their
ranges of possible values.
This is discussed in PART III of the course.
The seven-step modeling process

The figure below illustrates the seven-step process.

The model as a beginning, not an end
A completed model is really a starting point.
After you have a working model of the problem, you can – and
you should – use it as a tool for gaining insights.
For most models, many what-if questions can be asked.
If the model has been developed correctly, it should be capable
of answering such what-if questions fairly easily.
Spreadsheet Modeling
Basic spreadsheet modeling: Concepts and best practices (1 of 3)

Inputs have given

fixed values, at least
for the purposes of
the model.
Outputs are the
ultimate values of
interest; determined
by the inputs and the
decision variables. Decision variables
are those a
decision maker
controls

The Spreadsheet modeling is the process of entering the

inputs and decision variables into a spreadsheet and then
relating them appropriately, by means of formulas, to
obtain the outputs.
Concepts and best practices
(2 of 3)

After the outputs are obtained,

Can perform a sensitivity analysis to see how one or more outputs
change as selected inputs or decision variables change.
Can find the values of the decision variable(s) that minimize or
maximize a particular output, possibly subject to certain
constraints.
Can create charts that show graphically how certain parameters of
the model are related.
Concepts and best practices
(3 of 3)

Features that improve readability include:

A clear, logical layout to the overall model
Separation of different parts of a model, possibly across multiple worksheets
Clear headings for different sections of the model and for all inputs, decision
variables, and outputs
Use of range names
Use of boldface, italics, larger font size, coloring, indentation, and other
formatting features
Use of cell comments
Use of text boxes for assumptions and explanations
EXAMPLE SPREADSHEET MODELS
Breakeven analysis
What is break-even?
The break-even point is the point at which total cost and total revenue are equal,
i.e. "even". There is no net loss or gain.
How do you get?
FC+V*Q=R*Q
Solve for Q
Do you really need a spreadsheet?
For what-if-analysis
For graph
Breakeven analysis at Quality Sweaters
The Quality Sweaters Company sells hand-knitted sweaters. They
are planning to print a catalog of its products and undertake a direct
mail campaign. The cost of printing the catalog is $20 000 plus $0.1
per catalog.
The cost of mailing is $0.15. In addition, direct reply envelopes
will be included at a cost of $0.2 for each mail. The average order size
is $40 and the variable cost per order is 80% of value.
They plan to mail 100 000 catalogs.
Please answer:
How does a change in the response rate affect profit?
For what response rate does the company break-even?
Let’s use the file “Quality Sweaters-Breakeven” on LMS and work
in groups.
Using Goal Seek
How does a change in the response rate affect profit?
Let’s use a “Data table” under What-if-Analysis.
From the Data Table in the example, we see that profit goes
from negative to positive when the response rate is somewhere
between 5% and 6%.
To find the exact breakeven point, use Excel’s Goal Seek tool. Goal
Seek is used to solve a single equation with a single unknown.
Select Goal Seek from the What-If Analysis dropdown in the Data
ribbon and fill in the resulting dialog box as shown below.
Breakeven analysis at Quality Sweaters
Now, suppose that the response rate is fixed at “8%”
So, what should be the number to be mailed to maximize profit?
The more you mail, the more profit you get..

What about the break-even response rate? How does it change as

you send more mails?
As you send more mails, the break-even response rate needed
decreases.
Ordering with quantity discounts and demand uncertainty
Sam’s Bookstore needs to a new book Management Science. The cost of the book
will depend on the order quantity such as if the quantity is less than 1000, it is $24,
$23 if less than 2000, etc.
The demand can be anywhere from 500 to 4500.
How much of a product to order when customer demand for the product is
uncertain?
Let’s use the file “Quantity Discounts” on LMS and work in groups.
Ordering with quantity discounts and demand uncertainty
Use Vlookup to use the right price for any order quantity selected:

Use Two-way table under What-if-analysis to get profit under each demand-order
quantity pair
Then, you can get expected profits with Sumproduct to compare
Take-Aways
Reviewed a capital budgeting model to provide a concrete understanding of the book
concepts.
Explored a seven-step model-building process from problem definition to final
implementation.
Discussed why the study of management science is a valuable experience, even if
you do not intend to pursue a professional career in this field.
Learned how to design spreadsheet models for readability, and explored some of Excel’s
powerful tools, particularly data tables.
Summary of key management science terms
Summary of key Excel terms
(1 of 2)
Summary of key Excel terms
(2 of 2)