1
BUSINESS INTELLIGENCE (PART 1):
DECISION SUPPORT SYSTEMS
MISM2301
Thursday May 31, 2018
Associate Teaching Professor Martin Dias
2
Agenda
• Why think about decision-making processes?
• What are decision-support systems?
• Big data business intelligence
• PepsiAmericas debrief
• Next Session Planning
3
Wide range of BI & Analytics providers
Guess who paid for this complimentary publication
of the BI / Analytics Magic Quadrant?
4
Revisiting “decision making”
de·ci·sion (dĭ-sĭzh'ən) n.
1.The passing of judgment on an
issue under consideration
2.The act of reaching a
conclusion or making up one's
mind
5
Deciding = work
“The first rule for anyone hoping for
improved decision-making is to view
decisions as work.”
(Thomas Davenport, HBR, 2013)
6
Decision making process (routine)
Input
Process
Goal
Was
goal
met?
Output
Feedback
Decision
Adapted by Martin Dias from original by Mike Zack
7
Decision making process (problem-oriented)
What is the problem?
What are solution alternatives?
What is best solution?
Performance feedback
(adapted from Simon, 1960, as cited in Laudon and Laudon 2011)
8
Decision Making Process
• Intelligence:
• Specify desired situation (goal/hypothesis)
• Observe current situation (feedback)
• Describe gap between current and desired situation (control)
• Design:
• Identify alternative actions to close gap
• Determine expected result/payoff of each alternative
• With some probability of success (just like other risk analyses)
• Choice:
• Choose best alternative (weighted matrix)
• Based upon highest expected “payoff” / return-on-investment
• Implementation:
• People, process, and technology considerations
Feedback & Learning cycle: repeat a step as needed
Adapted by Martin Dias from original by Mike Zack
9
Levels
Making
Levelsof
of Decision
Information
Processing
Learning at scale - capture & disseminate
Learning/Strategic
Input
Process
Output
Innovate
Control/Tactical
Input
Process
Output
Informate
Transaction/Operational
Input
Process
Output
Automate
Adapted by Martin Dias from Original by Mike Zack
10
Decisions support by Organizational Level
Structure
Low
Not by accident
High
11
Types of Business Decisions - by Level
• Strategic - Affect the long-term direction of the firm
• Unstructured problems and solutions (“wicked problems”)
• Exploring and exploiting value in opportunities and threats
• Involve external sources of data and multiple decision-makers
• Are we doing the right things?
• Tactical - Medium-term outcomes about implementation of strategy
• Either structured or unstructured
• Discovering methods to optimize resource utilization / value
• Are we doing things right?
• Operational – Daily, routine, administrative in nature
• Generally structured/highly structured
• Identifying best means of delivering value (best depends on value discipline)
• Is this transaction an exception and what do I if so?
Adapted by Martin Dias based upon originals by © 2009 Luth Computer Specialists, Inc.
12
Types of Business Decisions – by structure
Structured
Unstructured
• DM process can
be specified and
codified
• DM process can’t be
easily specified or
codified
• Tend to be
repetitive/routine
• Often novel or ad
hoc
What can we digitize for unstructured decision-making?
13
Decisions vary in structure
Structured
Semi-structured
Unstructured
Insights
• human interaction
to specify issue
before solving
Algorithms
• Set formula
• Set procedure
Heuristics
• Rules of thumb
• If/then conditions
• Automation target
• Semi-automation target • Not automated
• DSS can compute the • DSS can infer (not
compute a solution)
optimal solution
• E.g., delivery route,
recommended
product
Operational DSS
• E.g., Shift scheduling
Expert System
• DSS can support
collaboration
• E.g., NPD
Group-DSS
Adapted by Martin Dias from originals by Michael Zack
14
Decision Support Systems
Information
What to do?
When to? Where to?
How to?
Decision Support System
Source: Michael Zack
15
DSS and Value Disciplines
Customer
Intimacy
Competitive advantage zone
Survival zone
Improved Decision Making zone
3
Value
Product
Leadership
Operational
Excellence
(Treacy & Wiersema 1993)
16
Decisions Support Systems
Why would you pay
more for machine/deep
learning capabilities?
(Laudon & Laudon 2011)
Using DSS
• What-if
Analysis
• Sensitivity
Analysis
• Goal-Seeking
Analysis
• Optimization
Analysis
Source: Michael Zack
OLAP and Data mining
MIS
19
Online Analytical Processing (OLAP)
• Multidimensional: multi-faceted statistical and
mathematical analysis using many variables
(attributes)
• Created especially to handle this efficiently:
• Real-time analysis & Output : immediate vs batch
• Computation: compute the model-driven combinations and
store as an N-dimensional “cube”
• Customers with incomes greater than $100,000 and who have a
mortgage with us also tend to have car loans with us but not mutual
funds.”
• Visualization: User can select /view output as a “cube”
• Pivot tables in Excel and Access
• Outputs as 3-D figures that can be “rotated”
• Interactive: user changes the variables, filters, and sorting
Adapted by Martin Dias from original by Mike Zack
Online Analytical Processing (OLAP)
OLAP Server
Internal External
Business Data
Databases sources
User Device
OLAP
Database
Software for
visualization and
interaction
IT components:
…hardware
…software
…data
…networking
Pre-determined
models
• Database
Indexing
• Pre-computed
data
combinations
Middleware for
data aggregation
• Data retrieved
from business
databases
• Staged in OLAP
multidimensional
database for
retrieval by client
Adapted by Martin Dias from original by Mike Zack
OLAP Examples – looking at trends from
different perspectives (time, product, location)
Pivot
(Pandre, A 2011)
22
DSS Output: Digital Dashboard example
*Image from Dotnetcharting
23
Data Mining – looking for success factors
Interesting (data-driven)
relationships - e.g.,
• Market basket by
demographics
• Loan failure indicators
• Fraud identification
Social Network
analysis
Classic
Data Cube
Computed
combinations
Data mining software
Source: Michael Zack
24
Data Mining process – looking for success factors
Identifying
important
factors/predictors
Interesting (data-driven)
relationships
Data mining software
Image source: http://www.slidegeeks.com/
25
Data mining output
Source: http://www.decision-making-confidence.com/decision-trees-examples.html
26
Big Data Business Intelligence – 3Vs +1
Source: IBM (2013)
27
Big Data Business Intelligence – 3Vs +1
Source: IBM (2013)
28
Big Data Business Intelligence – 3Vs +1
Source: IBM (2013)
29
Big Data Business Intelligence – 3Vs +1
What’s left?
Source: IBM (2013)
…this?
30
Big Data Business Intelligence – 3Vs +1
Source: SAP & IDC (2013)
…this?
31
Big Data Business Intelligence
If this prediction bears out, what does it mean for your information systems
investments?
IDC predicts that “decision and automation solutions, utilizing a mix of
cognitive computing, rules management, analytics, biometrics, rich
media recognition software and commercialized high-performance
computing infrastructure”
If this prediction bears out, what does it mean for your career?
“companies will begin to replace or significantly impact knowledge worker
roles.” …companies will focus more on the optimal mix between human and
machine capability and judgment.” If you take humans too much out of the
equation, their decision making will atrophy, warned IIA, asking “If you don’t
have experts, who will train the next generation of [machine learning]
software?”
Source: Press, Gil, "$16.1 Billion Big Data Market: 2014 Predictions From IDC And IIA," Forbes Magazine,
December 12, 2013.
32
case discussion
Describe the legacy IT architecture at PepsiAmericas before
its Customer Optimization initiative.
2. Describe the information problems that drove PepsiAmericas
to adopt a more aggressive attitude towards the capture,
use, and dissemination of information.
3. Describe what investments in information systems were
made to address these information needs (excluding DSS).
4. Explain the role of decision support systems (DSS) in
enabling the business success of PAS in terms of customer
intimacy and operational excellence. Your answer should
include treatment of both structured and unstructured
decisions. Describe one additional way that a DSS could
potentially enhance information processing within a firm.
1.
33
DSS: Business Considerations
When does an investment in DSS make sense?
Integration
Business
Problems
Drives
Drives
Alignment
Alignment
Information
Problems
Enables
Enables
Innovation
MIS capabilities
People
Process
Technology
34
case discussion
Describe the legacy IT architecture at PepsiAmericas before
its Customer Optimization initiative.
2. Describe the information problems that drove PepsiAmericas
to adopt a more aggressive attitude towards the capture,
use, and dissemination of information.
3. Describe what investments in information systems were
made to address these information needs (excluding DSS).
4. Explain the role of decision support systems (DSS) in
enabling the business success of PAS in terms of customer
intimacy and operational excellence. Your answer should
include treatment of both structured and unstructured
decisions. Describe one additional way that a DSS could
potentially enhance information processing within a firm.
1.
35
Pepsi invested significantly to optimize their customer
interactions and value
Figure 1 – IT Architecture PepsiAmericas
Source: Beath & Ross (2010)
36
Takeaways for today?
• What are different types of decisions and
how can DSS be used to improve decisionrelated outcomes?
• What are the different technology
components for DSS?
• What are business considerations for
implementing DSS?
37
Next session planning
• Due:
• Textbook reading: Belanger et al (2016) Chapter 14
• Diagnostic quiz
• Submit CDM-Smith case
Session 15: PepsiAmericas Case Study Questions
1. What drove PepsiAmericas to adopt a more aggressive attitude towards the utilization
of transaction data to run the business?
a. .
b. .
c. .
d. .
e. add more liens as needed…..
2. What investments in information management capabilities were taken in this regard?
Why were they obtained and how did they actually contribute?
Information System
Investment
Information Management
Need
Contribution of the System
in Operation
3. Establish for yourself a picture of PepsiAmericas before and after the decision to exploit
the use of information to inform business processes across the enterprise.
PepsiAmericas Before Its Investment in BI
Operations:
• .
• .
• .
Management and Control:
• .
• .
• .
Planning, Corporate Learning, and
Innovation:
• .
• .
• .
prepared by rmk 102518
PepsiAmericas After Its Investment in BI
Operations:
• .
• .
• .
Management and Control:
• .
• .
• .
Planning, Corporate Learning, and
Innovation:
• .
• .
• .
Page 1
PepsiAmericas Case Study
Case description:
In 2009 PepsiAmericas (PAS) was the world’s second largest manufacturer and distributor of
Pepsi beverages, operating in the U.S. (69% of sales), central and Eastern Europe (26% of Sales)
and the Caribbean (5% of sales). Net sales in 2008 totaled nearly $5 billion or 20% of PepsiCo’s
total US beverage sales. A recession hit the U.S. economy, but PepsiAmericas was also faced with
two more important long-term challenges: (1) a declining U.S. market for carbonated soft drinks,
and (2) increasingly powerful retailers who were squeezing PAS profit margins. In addition,
PepsiAmericas product line had moved from 35-40 products in the mid-1990’s to nearly 400
products by 2009.
These developments forced PepsiAmericas to embrace a completely new operating model. In
the past, distribution was handled by the local delivery person, who “owned” a particular route
of retail customer stores. The delivery person would load his/her truck in anticipation of what
was needed at each of his/her assigned locations. Over time, the delivery person knew what to
expect and could pretty much address customer needs on a day-to-day basis. However, as Pepsi
moved from 35 to 400 products and as the packaging for these products became less uniform, it
proved difficult to know about, let alone carry in inventory, the right mix of products in the truck.
Furthermore, chains like Wal-Mart, CVS, and Mobil Gas Stations, preferred highly centralized
procurement arrangements and national contracts. Pepsi was therefore obliged to create a
three-tier distribution platform that would address the needs of (1) national, (2) regional, and (3)
local customers.
In response to these pressures and challenges, PepsiAmericas invested heavily in supply-chain
management (SCM) and enterprise resource planning (ERP) systems. With these systems the
firm integrated its core business processes (i.e. procurement, manufacturing, selling, and
warehousing and distribution) and automated data capture at every key step along its value
chain. To PepsiAmericas, one of the biggest benefits of its ERPs was the collection and
measurement of business process outcomes for better performance assessment, forecasting,
and risk mitigation. The company used these rich data resources and related process knowledge
to negotiate better contracts for raw materials, lower supply chain operating costs, more
accurately monitor consumer demand, and ultimately strike more profitable deals with its large
retail customers. In effect PepsiAmericas employed customer data as a competitive asset,
collecting vast amounts of data as part of daily operations (transacting) and then employing that
data for management and control as well as for innovation in product development and customer
service.
This transformation process was dubbed the “Customer Optimization to the 3 rd power – Planning
+ Selling + Delivery” program and was intended to reduce inventory management issues, increase
productivity across PepsiAmericas’ production platforms, and improve overall customer service.
For example, national customers, like Wal-Mart, fed point-of sales data directly into PAS’s SCM
system, informing the detailed product mix and quantities going from PAS to particular sales
location. And at the other end of the spectrum, those PAS employees serving the small local
prepared for MISM 2301 by RMK, rev’d by MD Fall 2014
Page 1
PepsiAmericas Case Study
shops had access to detailed historical sales data to forecast the requirements for local stores
and to provide the right mix of products day in and day out. The ability to consistently adjust
prices was also significantly improved. Prior to having integrated systems and aligned operations,
PAS customers within the same ZIP code would have different prices depending upon who they
dealt with in PAS and the pricing data available to that sales person. In addition, productivity for
PAS customer service representatives rose noticeably.
Figure 1 – IT Architecture PepsiAmericas (Customer Optimization to the 3rd Power)
Source: Beath & Ross (2010)
Continuous data feeds from PAS SCM and ERP systems to the firm’s decision support systems
provided PepsiAmericas executives with ready access to real-time data to fine tune business
processes and to promptly address performance and customer servicing issues. This same
approach drove decisions concerning the acquisition of both additional production capabilities
and new lines of products. It also contributed to the continuous improvement of ongoing firm
business processes and services, and the shift to a more data-driven decision making culture
across the organization.
PepsiAmericas leadership employed their data assets to build
competitive knowledge in three areas that were critical to their long-term success, namely:
customer alignment and relationship management, supply-chain process improvement, and
enabling more dynamic pricing across the company. PAS continues to mine data across the
enterprise as a means to measure business results and to inform best practices.
prepared for MISM 2301 by RMK, rev’d by MD Fall 2014
Page 2
PepsiAmericas Case Study
Source:
Beath, Cynthia M. and Jeanne W. Ross, “PepsiAmericas: Building an Information Savvy Company”, Feb. 2010, MIT Sloan School
of Management Journal, accessed online http://dspace.mit.edu/bitstream/handle/1721.1/68550/Pepsi%20BeathRoss2.pdf
Case Questions:
For the below questions think first about the appropriate representation for your answer (bullet
list, multi-part bullet list, table) and then the appropriate framework to ensure both accuracy and
completeness.
1. Describe the information problems that drove PepsiAmericas to adopt a more
aggressive attitude towards the capture, use, and dissemination of information.
2. Explain the role of decision support systems (DSS) in enabling the business success of
PAS in terms of customer intimacy and operational excellence. Your answer should
include treatment of both structured and unstructured decisions. Describe one
additional way that a DSS could potentially enhance information processing within at
PAS.
prepared for MISM 2301 by RMK, rev’d by MD Fall 2014
Page 3
Purchase answer to see full
attachment