Sunteți pe pagina 1din 42

Spreadsheet Modeling and

Decision Analysis
DSO 547
Prof Dasgupta
Lecture 1

Session Outline
Course Overview
Introduction to Modeling and Problem Solving
Spreadsheets and Decision Analysis

Course Structure
Mix of Lecture, working sessions and exercises, exams, and course
project.

5 Cases (team work)


Midterm
Final
Team Project
TOTAL

25%
25%
25%
25%
100%

Group Project
25% of course grade. Grade division is as follows:

30% for rigor of model and insights


30% for quantitative analysis
20% for quality of problem solving
20% for in-class presentation

Rigor in quantitative problem solving and


spreadsheet-enabled analysis is the focus of
assessment
I will pay close attention to informal peer
feedback on your participation

Introduction
We face numerous decisions in life &
business.
We can use computers to analyze the
potential outcomes of decision
alternatives.
Spreadsheets are the tool of choice for
todays managers.

What is Business Analytics?


A field of study that uses computers,
statistics, and mathematics to solve
business problems.
Also known as:
Operations research
Management Science
Decision science

Home Runs
in Business Analytics
Proctor & Gamble
Developed multi-echelon inventory planning
tool for safety stock optimization
Lowers inventory while maintain customer
service
Benefits:
Reduced inventory investments by $1.5 billion

Home Runs
in Business Analytics
New Brunswick (CA) Dept of Transportation
Developed linear programming-based strategic
planning tool
Includes long-term objectives and operational
constraints on costs, timings, asset life cycle
Benefits:
Estimated $72 million in annual savings from a $2 million
investment

Home Runs
in Business Analytics

Industrial and Commercial Bank of China


Worlds largest publicly traded bank
16,000 branch locations
Worked with IBM to develop a tool to predict
where new branches should be opened
Implemented in 40 cities throughout China
Benefits:
Estimated $1 billion in new deposits in typical major
cities

Home Runs in
Business Analytics
Midwest Independent Transmission System
Operator (MISO)
Manages power generation in 13 U.S. midwest states
Uses optimization model to determine when various
plants should be running
Other models predict energy output & prices
Benefits:
Improved plant efficiency & grid reliability
Savings of ~$3 billion from 2007-2010

10

What is a Computer Model?


A set of mathematical relationships and
logical assumptions implemented in a
computer as an abstract representation of a
real-world object of phenomenon.
Spreadsheets provide the most convenient
way for business people to build computer
models.

11

The Modeling Approach


to Decision Making
Everyone uses models to make decisions.
Types of models:
Mental (arranging furniture)
Visual (blueprints, road maps)
Physical/Scale (aerodynamics, buildings)
Mathematical (what well be studying)

12

Characteristics of Models
Models are usually simplified versions of the
things they represent
A valid model accurately represents the
relevant characteristics of the object or
decision being studied

13

Benefits of Modeling
Economy - It is often less costly to analyze
decision problems using models.
Timeliness - Models often deliver needed
information more quickly than their real-world
counterparts.
Feasibility - Models can be used to do things that
would be impossible.
Models give us insight & understanding that
improves decision making.

14

Example of a Mathematical Model


Profit = Revenue - Expenses
or
Profit = f(Revenue, Expenses)
or
Y = f(X1, X2)

15

A Generic Mathematical Model


Y=

f (X1, X2, , Xn)

Where:

Y = dependent variable
(aka bottom-line performance measure)
Xi = independent variables (inputs having an impact on Y)
f (.) = function defining the relationship between the Xi & Y

The Problem-Solving Framework for


Leveraging Business Opportunities

Identify
Problem
Probortunity

Formulate &
Implement
Model

Analyze
Model

unsatisfactory
results

Test
Results

Implement
Solution

The Psychology of Decision Making


Models can be used for structurable
aspects of decision problems.
Other aspects cannot be structured easily,
requiring intuition and judgment.
Caution: Human judgment and intuition is
not always rational!

18

Anchoring Effects
Arise when trivial factors influence initial
thinking about a problem.
Decision-makers usually under-adjust from
their initial anchor.
Example:
What is 1x2x3x4x5x6x7x8 ?
What is 8x7x6x5x4x3x2x1 ?

19

Framing Effects
Refers to how decision-makers view a
problem from a win-loss perspective.
The way a problem is framed often
influences choices in irrational ways
Suppose youve been given $1000 and must
choose between:
A. Receive $500 more immediately
B. Flip a coin and receive $1000 more if heads
occurs or $0 more if tails occurs

20

Framing Effects (Example)


Now suppose youve been given $2000 and
must choose between:
A. Give back $500 immediately
B. Flip a coin and give back $0 if heads occurs or
give back $1000 if tails occurs

21

A Decision Tree for Both Examples

Payoffs
$1,500

Alternative A
Initial state

Heads (50%)
Alternative B
(Flip coin)

Tails (50%)

$2,000
$1,000

Good Decisions vs. Good Outcomes

Good decisions do not always lead to good outcomes...

A structured, modeling approach to decision making


helps us make good decisions, but cant guarantee good
outcomes.

23

Decisions & Outcomes


Outcome Quality
Decision
Quality

Good

Bad

Good

Deserved Success

Bad Luck

Bad

Dumb Luck

Poetic Justice

24

Designing Spreadsheet-Based Business


Models

Base Case modeling

4 - 25

Introduction: base case models


Deterministic base case model:
Each set of inputs yields a single set of outputs (no multiple solns)

Uncontrollable Inputs
(Random events,
Observations, etc.)
Controllable Inputs
(Decisions)

Base Case
Model

Outputs

In practice we can always build a model if we know the structure


of the environment dummy numbers can be used for specific
inputs, and then updated later
26

26

Example model:

Income Statement
Pro Forma

Revenue

Influence
Chart:

Price

Var Cost

Total Cost
Profit

Unit Cost

Fixed Cost

27

27

Why build spreadsheet models?


What-if analysis / Decision-making
Threshold analysis (e.g., breakeven sales quantity)
Create potential scenarios (e.g., upside, downside, base case)
Calculation of the answer (e.g., target profit, capital requirements,
production schedule amounts)

Sensitivity analysis
Understand Model Drivers Understand the dynamics of your business
system
Understand linearity of response (more on this later)
Know which aspects of a business environment deserve the more
attention

28

28

Sensitivity analysis
Sensitivity analysis is any technique used to determine the impact of an independent variable
(i.e., input) on a dependent variable (i.e., output or calculation).
Method:
Increase/(decrease) a cell input by a finite amount (i.e., 10 units) or percentage (i.e., +10%),
and measure the change in various output/calculation cells
Purpose:
Identify model drivers, which are the inputs of greatest importance

29
Image source: http://www.epixanalytics.com/modelassist

29

A lifecycle of quantitative analysis

Discovery &
Problem
Definition

Wheres the
problem?

Solution

The math
problem

Analysis

Action

So what

Now what?

Remember to
iterate!

30

30

Proper Modeling Etiquette


Structure
Whenever possible,
work down the spreadsheet
include consecutive time periods in consecutive columns
dont leave gaps or subtotals b/w columns, only in the right-most column

Headings and Titles


1st row: Title
2nd row: Description
3rd row: Name/date/update/version number
Columns and rows appropriately labeled, including units, where appropriate
Cell fill using light pastel colors

Calculations
Input cells in blue font;
Calculation cells in black font;
Alert/Check cells in red font
Irregular cells and calculations to other sheets in green font;
Pay mind to *units* label appropriately whenever possible

31

31

Proper Modeling Etiquette, continued


Notes and comments
Use on-page comments, wherever possible; reserve floating cell comments when
spreadsheet is cramped.
Include a Notes section at the end worksheets/workbooks (data sources,
references, assumptions, etc.)
When in doubt, provide an explicit notes. The best models can be understand by a
3rd party and/or layperson without explanation from the modeler.

Formatting/presentation
Column headings in bold, cell color light pastel fill, if at all
Colors and borders used sparingly
Format to print properly

32

32

Error checking must happen before analysis

Perform checks in order of


increasing tedium

Test extremes by changing inputs and predicting your model response. Use
Zero values
Breakeven points
Huge positive numbers, Huge negative numbers,
Tiny non-zero numbers
Calculate reality check that your model outputs approximate expected
outcomes
Display and check formulas
Ctrl+` shows all formulas
F2 displays contents of a selected cell
Trace dependents/precedents
Get fresh eyes ask a friend
Create check cells aka error traps
Example: create a cell that yields an error value if sum of column subtotals
does not equal sum of row subtotals
Example: Balance sheet cell check
Finally hand check every single cell yes, really!

33

33

Designing Spreadsheet-Based
Business Models

Excel Basics

4 - 34

Tips on using advanced functions


Functions always calculate according to order of operations (PEMDAS)
Avoid embedding formulas more than several layers deep
Ex: =If(C2>A2,Yes,IF(ISERROR(AND(

Formula auditing
Trace dependents, precedents
Check each argument, range, quotes around each text string, check
embedded functions independently

Be careful with running operations/formulas using logicals TRUE and


FALSE as arguments

35

35

Common Function Error Outputs

#Name? formula cant identify inputted text

#Num! formula or function contains invalid numeric values

Wrong data type supplied for a function that requires a numeric argument
Iterative function cant find result (e.g., IRR or RATE)
Resulting number too large or too small to be output in Excel

#N/A a value not available to a function or formula

Cell range called which doesnt exist (or misspelled)


Function call misspelled or nonexistent
Text string entered without quotes
Colon (:) omitted/deleted in a range reference

VLOOKUP/HLOOKUP/MATCH/LOOKUP functions that cannot find an appropriate match


One or more arguments omitted from a formula
Array formula references do not contain the same number of rows and columns
Data missing and #N/A or NA() entered in its place

#Null! incorrect range reference

Ex: =SUM(A1:B3 C1:D5)

36

36

Common Function Error Outputs (contd)

#Value! formula computes on mismatched data types


=32+dog or referenced cells containing those

values

#Div/0! (1) formula computes divide by zero or by a cell that contains no data,
or (2) formula computes a legal operation on a #Div/0! outputting cell

#Ref! formula cell reference not valid


Ex: deleted the cell/row/column referenced
Dynamic link reference broken

##### column too narrow to display contents

37

37

VLOOKUP Example

=VLOOKUP( Lookup_value, Table_array, Col_index_no, FALSE)


What we are
looking for

Where
to look

What to
return

FALSE means look


for exact match only

Things to note:
Every lookup updates every time sheet is altered or recalculated
Be careful with absolute/relative references on Lookup_value and Table_array
Impossible to match to anything but the first column in Table_array
Return field must be to the right of the lookup field
True/false argument should be set to FALSE to find exact match
Error values denote no found match
How can we embed the formula ISERROR to remove error values?

38

38

Pivot Table Pop Quiz!!!

Pivot this data by hand


Group by color and pass
Find sum of score in each

A few additional notes

Index

Colorsubtotal,
Score
group,
total

Pass

Red

Red

Blue

Yellow

Yellow

Yellow

Link multiple pivot tables from the same data


Note advanced display, sorting, calculated fields

39

39

Create a Two-Variable Data Table

Step 1:
Build the
benchmark model

Step 2:
Create the
periphery of the
data table

Link to the output cell

40

40

Create a Two-Variable Data Table (Cont.)

Step 3:
Select the range, click Data Data Tools What-if Analysis Data Table

Step 4:
Link to the two
input cells

41

Summary Time

Who will take the lead on


summarizing the key learning today?

S-ar putea să vă placă și