Contenu connexe Similaire à Discover the value in IBM Business Analytics (20) Plus de Daryl Pereira (20) Discover the value in IBM Business Analytics1. IBM Business Analytics and Optimization
Discovering the Value of
Business Analytics
Lennart Frantzell alf@us.ibm.com Giuseppe Accardo gaccard@us.ibm.com, Chris Heckhart checkart@us.ibm.com
IBM San Mateo Innovation Center, San Mateo, California
2011/05/6
© 2009 IBM Corporation
2. Event Agenda
IEEE Accessing the Future Conference, Boston, July 2009.
Discover the Value of IBM Business Analytics
Wednesday, May 25, 2011
TIME TOPIC
10:00 a.m. Registration
10:15 a.m. Introduction, an overview of Business Analytics
Lennart Frantzell, San Mateo IBM Innovation Center
11:30 a.m. IBM Business Analytics product portfolio introduction
Giuseppe Accardo, San Mateo IBM Innovation Center
12:00 Lunch, networking
1:00 p.m. IBM Business Analytics product portfolio introduction, cont
Giuseppe Accardo, San Mateo IBM Innovation Center
1:30 p.m. Prescriptive analytics in the real world with ILOG
Jeremy Bloom, IBM
2:00 p.m. Build Online Revenue
Gene Hoffman, Vindicia
2:30 p.m. Where do we go from here?
Lennart Frantzell, San Mateo IBM Innovation Center
© 2009 IBM Corporation
3. IEEE Accessing the Future Conference, Boston, July 2009.
The sea change, from analog to digital data
• Historical change, from analog to digital data
• Today, mankind generates staggering amounts of digital data
• How do we search vast amounts of digital data?
• How do we make sense of all this data?
• Mobile computing, social networks and Cloud Computing make
business analytics accessible everywhere
• Mankind entering era of informed decision making
© 2009 IBM Corporation
4. 1 billion transistors
IEEE Accessing the Future Conference, Boston, July 2009.
for each person
on earth.
1 trillion things
connected to
the net.
THINK
By 2010,
30 billion RFID tags,
embedded into
our world.
© © 2009 IBM Corporation 4
ILO
5. IEEE Accessing the Future Conference, Boston, July 2009.
Artificial Intelligence and Analytics
• AI
– Inference engines, Expert Systems, Rete
Algorithm, Prolog, Neural Networks,
Rules-based systems Analytics
– Complex algorithms and very little data
– Answers intertwined with algorithms
– Snakebites in Australia Patient HIV treatment
– Airline scheduling
• Analytics
– Staggering amounts of data
– Separation of algorithms and data • Match incoming patient against patients
– Successful HIV treatment in Ethiopia, who have been successfully treated for HIV,
match patients against data. • Select that treatment
© 2009 IBM Corporation
6. IEEE Accessing the Future Conference, Boston, July 2009.
Searching large amounts of data
The canonical example application of MapReduce is a process to count the appearances
of each different word in a set of documents:
void map(String name, String document):
// name: document name
// document: document contents
for each word w in document:
EmitIntermediate(w, "1");
void reduce(String word, Iterator partialCounts):
// word: a word
// partialCounts: a list of aggregated partial counts
int result = 0;
for each pc in partialCounts:
result += ParseInt(pc);
Emit(AsString(result));
© 2009 IBM Corporation
7. IBM’s Grand Challenges: Deep Blue
IEEE Accessing the Future Conference, Boston, July 2009.
1997
IBM’s chess-playing computer. Each chip was equipped with a million transistors, which evaluated 2 million positions
Each second.
In Deep Blue, some 256 chips were teamed together under the overall control of a general-purpose IBM SP2®,
a parallel computer consisting of, in this case, 32 processor nodes.
The parallelism derived from these 32 processors and 256 chess accelerator chips is what mades Deep Blue
the most powerful chess computer in the world.
It was capable of looking at an average of 100 million positions per second.
© 2009 IBM Corporation
8. IBM’s Grand Challenges: Deep Blue, Blue Gene
IEEE Accessing the Future Conference, Boston, July 2009.
1997
2005
Blue Gene
Blue Gene is an IBM Research project dedicated to exploring the frontiers in supercomputing: in computer architecture,
in the software required to program and control massively parallel systems, and in the use of computation to advance
our understanding of important biological processes such as protein folding.
© 2009 IBM Corporation
9. IBM’s Historical Grand Challenges: Deep Blue,
IEEE Accessing the Future Conference, Boston, July 2009.
Blue Gene and Watson
1997
2005
2011 © 2009 IBM Corporation
10. IEEE Accessing the Future Conference, Boston, July 2009.
Watson and structured versus unstructured data
The canonical example application of MapReduce is a process to count the appearances
of each different word in a set of documents:
void map(String name, String document):
// name: document name
// document: document contents
for each word w in document:
EmitIntermediate(w, "1");
void reduce(String word, Iterator partialCounts):
// word: a word
// partialCounts: a list of aggregated partial counts
int result = 0;
for each pc in partialCounts:
result += ParseInt(pc);
Emit(AsString(result));
© 2009 IBM Corporation
11. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Watson and IBM’s DeepQA Technology
• Watson runs IBM’s DeepQA technology, developed using Apache UIMA, a framework implementation
• of the Unstructured Information Management Architecture.
• UIMA was designed to support interoperability and scale-out of text
and multimodal analysis applications.
• The Watson database includes Wikipedia and other sources
• Powered by IBM POWER7 processor technology, Watson is
an example of the complex analytics workloads that are
becoming increasingly common in business
• Watson also uses Apache Lucene, Indri, SPARQL and the Jena Toolkit
• Watson’s DeepQA UIMA annotators were deployed as mappers
in the Hadoop map-reduce framework,
which distributed them across processors in the cluster.
IBM DeepQA
Apache UIMA
The Regular Expression Annotator (RegexAnnotator) Database/Wikipedia
is an Apache UIMA analysis engine that detects entities
like email addresses, URLs, phone numbers, zip codes or IBM Power 7 Hardware
any other entity based on regular expressions and concepts.
July 20, 2009 © 2009 IBM Corporation
12. IEEE Accessing the Future Conference, Boston, July 2009.
Top right: World of Warcraft Bottom Right: Wicked Left: Frank Baum
© 2009 IBM Corporation
13. IEEE Accessing the Future Conference, Boston, July 2009.
•
July 20, 2009 IBM Confidential © 2009 IBM Corporation
14. IEEE Accessing the Future Conference, Boston, July 2009.
Watson in Healthcare
Natural Language Processing in Healthcare
• As Electronic Healthcare Records systems are adopted by
Government mandate, physician notes are digitized in a computer
readable format…, the Mayo Clinic and IBM have already
announced a partnership to open source much of the UIMA
annotators Mayo developed to mine its own medical records.
• Mining patient reported data is another interesting area. Patient communities such as PatientsLikeMe
and Association of Cancer Online Resources.
• In 1999 by BMJ (British Medical Journal) a team of researchers observed 103 physicians over one
work day. Those physicians asked 1,101 clinical questions during the day. The majority of those
questions (64 percent) were never answered.
http://www.ibm.com/developerworks/industry/library/ind-watson/index.html
July 20, 2009 IBM Confidential © 2009 IBM Corporation
15. IEEE Accessing the Future Conference, Boston, July 2009.
July 20, 2009 IBM Confidential © 2009 IBM Corporation
16. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Smarter Planet
•
© 2009 IBM Corporation
17. IEEE Accessing the Future Conference, Boston, July 2009.
Business Analytics in Action
HIV treatment in Ethiopia
Sequoia Hospital in Silicon Valley
© 2009 IBM Corporation
18. IEEE Accessing the Future Conference, Boston, July 2009.
EuResist, HIV Treatment in Ethiopia
Doctors in Ethiopia can instantly
compare this blood sample to over
41,000 HIV treatment histories.
EuResist is helping doctors predict
patient response with over 78%
accuracy – outperforming 9 out of 10
human experts.
The tool is built on an IBM analytics solution that integrates a variety of disparate
databases onto a flexible IBM DB2® platform to process complex metadata more
effectively than anything else on the market.
Link: http://www.euresist.org/
© 2009 IBM Corporation
19. IEEE Accessing the Future Conference, Boston, July 2009.
EUResist Demo
© 2009 IBM Corporation
21. IEEE Accessing the Future Conference, Boston, July 2009.
How do we use Business Analytics?
Reference Implementation
© 2009 IBM Corporation
22. IEEE Accessing the Future Conference, Boston, July 2009.
Optimization and Analytics, an Overview
What’s the best that can happen Stochastic
including the effects of variability? Optimization
Prescriptive
What’s the best that can happen ? Optimization
Competitive Advantage
What will happen next ? Predictive Modeling
What if these trends continue? Forecasting Predictive
Statistical
Analysis What could happen…. ?
Alerts What actions are needed?
Query/Drill Down What exactly is the problem?
Descriptive
Ad Hoc Reports How many, how often, where?
Std Reports What happened?
Degree of Complexity Based on: Competing on Analytics, Davenport and Harris, 2007
© 2009 IBM Corporation
23. Architecture pattern: Service Orientation architecture and analytics
IEEE Accessing the Future Conference, Boston, July 2009.
Analysis
Dashboarding
Optimization
Analysis
Predictive
Statistical
Datamining
Analysis
Analysis
Analytics
ETL (Extract Transform Load)
ETL Data
Data sources: Warehouse
Patient data, e-meters, Cycle initiation
data streams, Build reference data
Extract (from sources)
unstructured data Validate
Transform (clean, apply business rules, check for data integrity… )
Stage (load into staging tables, if used)
Audit reports (for example, on compliance with business rules. )
Publish (to target tables)
Archive
Clean up © 2009 IBM Corporation
24. Cross River, HIF Reference Implementation at Global
IEEE Accessing the Future Conference, Boston, July 2009.
Solutions Center
• Premier test harness, product showcase, and center of excellence at the IBM Global Solutions
Center (Dallas)
• Proof point for integrating all essential components for an enterprise class health analytics platform
(integration, analytics, presentation layer)
Data Source Layer
(Clinical, Financial, (Operational, Administrative)) …
© 2009 IBM Corporation
25. Cross River, HIF Reference Implementation at Global
IEEE Accessing the Future Conference, Boston, July 2009.
Solutions Center
• Premier test harness, product showcase, and center of excellence at the IBM Global Solutions
Center (Dallas)
• Proof point for integrating all essential components for an enterprise class health analytics platform
(integration, analytics, presentation layer)
Integration Layer
InfoShpere WSTX Adapters Rational Data
Information Analyzer, DataStage, QualityStage, Service Director (HL7) (e.g. Cache) Architect
Data Source Layer
(Clinical, Financial, (Operational, Administrative)) …
© 2009 IBM Corporation
26. Cross River, HIF Reference Implementation at Global
IEEE Accessing the Future Conference, Boston, July 2009.
Solutions Center
• Premier test harness, product showcase, and center of excellence at the IBM Global Solutions
Center (Dallas)
• Proof point for integrating all essential components for an enterprise class health analytics platform
(integration, analytics, presentation layer)
Data Layer Data InfoShpere InfoShpere WH InfoSphere Meta InfoSphere
BCU
Models Warehouse Cubing Services Data Management Business Glossary
Integration Layer
InfoShpere WSTX Adapters Rational Data
Information Analyzer, DataStage, QualityStage, Service Director (HL7) (e.g. Cache) Architect
Data Source Layer
(Clinical, Financial, (Operational, Administrative)) …
© 2009 IBM Corporation
BCU: Balanced Configuration Unit
27. Cross River, HIF Reference Implementation at Global
IEEE Accessing the Future Conference, Boston, July 2009.
Solutions Center
• Premier test harness, product showcase, and center of excellence at the IBM Global Solutions
Center (Dallas)
• Proof point for integrating all essential components for an enterprise class health analytics platform
(integration, analytics, presentation layer)
Analytic Layer Cognos Performance InfoShpere Structured and
Cognos BI
Management Unstructured Data Mining
Data Layer Data InfoShpere InfoShpere WH InforSphere Meta InfoSphere
BCU
Models Warehouse Cubing Services Data Management Business Glossary
Integration Layer
InfoShpere WSTX Adapters Rational Data
Information Analyzer, DataStage, QualityStage, Service Director (HL7) (e.g. Cache) Architect
Data Source Layer
(Clinical, Financial, (Operational, Administrative)) …
© 2009 IBM Corporation
28. Cross River, HIF Reference Implementation at Global
IEEE Accessing the Future Conference, Boston, July 2009.
Solutions Center
• Premier test harness, product showcase, and center of excellence at the IBM Global Solutions
Center (Dallas)
• Proof point for integrating all essential components for an enterprise class health analytics platform
(integration, analytics, presentation layer)
Presentation Layer
Clinicians Researchers Patients Administrators
Chronic Disease Track, analyze, Planning and Ad Hoc
Cognos Cohort Analysis
Management report events Forecasting Analysis
WebSphere Portal Server
Analytic Layer Cognos Performance InfoShpere Structured and
Cognos BI
Management Unstructured Data Mining
Data Layer Data InfoShpere InfoShpere WH InforSphere Meta InfoSphere
BCU
Models Warehouse Cubing Services Data Management Business Glossary
Integration Layer
InfoShpere WSTX Adapters Rational Data
Information Analyzer, DataStage, QualityStage, Service Director (HL7) (e.g. Cache) Architect
Data Source Layer
(Clinical, Financial, (Operational, Administrative)) …
© 2009 IBM Corporation
30. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Discovering the Value of Business Analytics
IBM Product Portfolio
© 2009 IBM Corporation
31. IEEE Accessing the Future Conference, Boston, July 2009.
Technology Evolution of BI & Analytics (Blog: Wayne Eckerson - BeyeNetwork)
Sub-market Segments:
BI Tools
Data Integration tools
DB Management Systems
Hardware Platform
Reporting languages
(Focus and Ramis)
© 2009 IBM Corporation
32. IEEE Accessing the Future Conference, Boston, July 2009.
The integrated platform
The integrated platform
32 © 2009 IBM Corporation
33. IEEE Accessing the Future Conference, Boston, July 2009.
The integrated platform
The integrated platform
33 © 2009 IBM Corporation
34. IEEE Accessing the Future Conference, Boston, July 2009.
Actionable Optimization & Analytics
What should we do, given the
What-if Analysis alternatives and real-time changes?
Prescriptive
Mathematical Optimization How can we achieve the best outcome?
Foresight
Monte Carlo simulation What could happen …?
Predictive
Predictive modeling What will happen next if ?
Forecasting What if these trends continue?
Competitive Advantage
Alerts What actions are needed?
Query/drill down What exactly is the problem?
Insight
Descriptive
Ad hoc reporting How many, how often, where?
Standard Reporting What happened or is happening?
Degree of Complexity
Based on: Competing on Analytics, Davenport and Harris, 2007
34
© 2009 IBM Corporation
35. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Portfolio – Key Products
… What happened ?
35 © 2009 IBM Corporation
36. IEEE Accessing the Future Conference, Boston, July 2009.
The integrated platform
The integrated platform
36 © 2009 IBM Corporation
37. IEEE Accessing the Future Conference, Boston, July 2009.
Business Intelligence & Performance Management
Answer three important questions that drive better performance
Finance
Sales Operations
How are we doing?
Scorecards and Dashboards
Marketing What should we be doing?
Planning, Forecasting and Budgeting
Why?
Reporting & Analytics
Customer Human
Service Resources
IT/Systems
37 © 2009 IBM Corporation
38. IEEE Accessing the Future Conference, Boston, July 2009.
10 Capabilities
Querying and Reporting
Querying and Reporting
Analysis & Planning
Analysis & Planning
Dashboarding
Dashboarding
Scorecarding
Scorecarding
© 2009 IBM Corporation
39. IEEE Accessing the Future Conference, Boston, July 2009.
10 Capabilities
Real time monitoring
Real time monitoring
Statistics
Statistics
Extending BI
Extending BI
Collaborative BI
Collaborative BI
© 2009 IBM Corporation
40. IEEE Accessing the Future Conference, Boston, July 2009.
10 Querying and Reporting
•Design and build – Create
report templates to include
standard report objects, queries,
and layouts.
•Analyze and share – View,
interact with and analyze the
result set, and share the results
generate a unique perspective
around information.
•Assemble and format widgets
from BI, TM1, Real-Time
Monitoring, Metric Studio,
PowerPlay, RSS and HTML
elements etc and put them in a
single report
© 2009 IBM Corporation
41. IEEE Accessing the Future Conference, Boston, July 2009.
10 Querying and Reporting
•Relational databases
from IBM, Oracle,
Microsoft, Teradata, and
Sybase, various sources •Satellite data sources,
•Content management accessible via ODBC and including Microsoft Excel files,
data, including IBM dimensionally aware Microsoft PowerPoint® files,
FileNet®, EMC sources like SAP BW. Microsoft Access® files, flat
Documentum, OpenSoft files and more.
and others
•Mainframe sources, •Modern data sources,
including VSAM, IMS, Supported such as XML, LDAP and
IDMS, COBOL® WSDL
data
copybooks and others
sources
•Enterprise data •Widely deployed ERP
warehouses and marts, systems, including
•All widely used OLAP mySAP (R/3), PeopleSoft
with both 3NF and star sources, including IBM
schemas. Enterprise, JD Edwards
DB2 OLAP Server, IBM EnterpriseOne, Oracle
Cognos PowerCube, eBusiness Suite and
Microsoft Analysis Siebel CRM.
Services, Oracle 10G and
Oracle EssbaseOLAP.
© 2009 IBM Corporation
42. IEEE Accessing the Future Conference, Boston, July 2009.
10 Dashboarding
Louis Barton, a Frost Bank IT executive,
dashboards add value by “reducing the
cycle time it takes to analyze information
[key performance metrics], You can
make a decision sooner. That means
people are more productive.”
© 2009 IBM Corporation
43. IEEE Accessing the Future Conference, Boston, July 2009.
TM1 Planning Software
With Cognos Planning you can access current actual data to assess fiscal performance, and proceed from what-
is to evaluate the what-if scenarios critical to forecasting future performance.
• Rapid development.
• Sophisticated modeling. PLANS & FORECASTS
• Flexibility.
• Finance friendliness.
• Less time on process
Power of “sandboxing”:
DEMO
DEMO
Video
Video
Link: http://forms.cognos.com/?elqPURLPage=2293&offid=od_tm1
© 2009 IBM Corporation
44. IEEE Accessing the Future Conference, Boston, July 2009.
10 Scorecarding -
Communicate strategy - Understand key relationships
- Build metrics and scorecards based on reliable
information
It allows executives and business managers to instantly
visualize how the business is performing against key
performance indicators.
At the operational level, departments and employees
Strategy Map with associated metrics can use scorecards to monitor their performance
against targets set for specific projects and activities.
Metrics grouped by owner Cause and effect diagram Advance initiative tracking
© 2009 IBM Corporation
45. IEEE Accessing the Future Conference, Boston, July 2009.
10 Extending Business Intelligence
Provide actionable intelligence to users, no matter their
location or their connectivity.
Business users, from executives to mobile field workers, can
know and understand the health of the business at all
times, and have the tools to take action on what they see.
Reduce the burden on IT to redevelop reports for various
devices.
Take full advantage of the mobile network infrastructure —
an excellent opportunity for low-cost BI deployment.
© 2009 IBM Corporation
46. IEEE Accessing the Future Conference, Boston, July 2009.
10 Collaborative Business Intelligence
With integrated Lotus Connections, users can:
• Link directly from Lotus Connections to a Cognos Business Insight dashboard
• Use single sign-on for both Business Insight dashboards and Lotus Connections
• Add other individuals to an Activity at any point in the decision-making process
• Search for Activities directly from the Business Insight window
• Send email notifications directly from the Activity
© 2009 IBM Corporation
47. IBM Cognos Express – Solution for the Mid Market
IEEE Accessing the Future Conference, Boston, July 2009.
Features of IBM Cognos Express Features of IBM Cognos Express Features of IBM Cognos Express
Reporter Advisor Xcelerator
•Complex reporting tool designed for •Create multidimensional view of your •Delivers the powerful and fast in-memory
business business based from your relational data multidimensional database
•Reports against a single common data with a few clicks while employing the •Create scenarios, versions, variance and
source will harmonize your business powerful and fast in-memory what-if analysis against live data directly in
•Self-service flexible reports to meet the multidimensional database Excel
needs of different users, including financial, •Get maximum information from your data •Build and edit your plans real-time with
production, operational, transactional, using drill-down and drill-up capability in write-back capability
managed or ad hoc reports combinations with lucid graphical outputs •Use worksheets - employ your strong
•No matter if relational or multidimensional •Conformable self-service ad-hoc analysis knowledge of Excel and extend it with
OLAP data are used for reports according to your needs without waiting for powerful Cognos Express functionality like
•Ergonomic Web interface IT department implementation multidimensional data functions
•Drag&Drop style of work •Step into the world of what-if analysis and •Web interface available for easy data
•Publish reports to web portal, HTML, PDF planning with the write-back and data contribution and work with excel
or Excel files spread features worksheets without having Excel installed
•Interactive dashboard for quick orientation •Ergonomic Web interface on your machine
and decision making across the whole •Employ the power of dashboards and •A single common base for metadata and
company interactive reports data, business rules and calculations,
•Integration with other modules, a single •Integration with other modules, a single which harmonizes the view of your
platform for BI and planning platform for BI and planning business
47 © 2009 IBM Corporation
48. Case Study #1 - BMR tones up its sales performance with advanced
IEEE Accessing the Future Conference, Boston, July 2009.
analytics
Business need:
BMR was in the process of replacing its core ERP solution, and saw this as an opportunity to
enhance its business analytics capabilities to deliver improved sales performance
management. As a mid-sized business, BMR wanted to find an affordable solution that would
offer enterprise-class functionality.
Solution:
ProStrategy Colman, an IBM Business Partner, helped BMR become the first company in
Europe to implement IBM Cognos Express – an all-in-one business intelligence and
planning solution designed for mid-sized companies. The solution is integrated with the
company’s new Microsoft Dynamics NAV ERP system, and also draws data from sales
channels such as eBay, BMR’s Slendertone website and retail customer databases.
Benefits:
Provides real-time analysis of sales performance, helping sales teams and managers
work more productively. Reduces time spent on collecting and checking data by more than
30 percent, allowing users to focus on actual analysis. Eliminates data silos and provides a
‘single version of the truth’ with accurate, up-to-date information.
http://www-01.ibm.com/software/success/cssdb.nsf/CS/STRD-8CEE4P?OpenDocument&Site=default&cty=en_us
© 2009 IBM Corporation
49. Case Study #2 - Mercury Medical a healthcare manufacturer improves
IEEE Accessing the Future Conference, Boston, July 2009.
reporting and analysis with IBM Cognos Express
Business need:
IBM Software Valuenet Reselling Partner, BlueNET Technologies introduced Mercury Medical
to Cognos Express through the 30-day product trial that allowed BlueNET to create a
custom report and analysis demo tailed to Mercury’s specific data and user needs.
Solution:
Cognos Express met Mercury Medical’s recovery time objectives, giving the company
confidence in its decision.
Benefits:
Mercury’s users can now create the most critical reports that they were previously relying on
a Legacy Reporting Platform to produce. These include sales commissions, weekly sales, a
rolling 12-month sales report, and an inventory summary report.
http://www-01.ibm.com/software/success/cssdb.nsf/CS/SANS-8DBM29?OpenDocument&Site=default&cty=en_us
© 2009 IBM Corporation
50. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Portfolio – Key Products
… What could happen ?
50 © 2009 IBM Corporation
51. Statistical Package for the Social Sciences (SPSS) V.19
IEEE Accessing the Future Conference, Boston, July 2009.
The integrated platform
The integrated platform
51 © 2009 IBM Corporation
52. IEEE Accessing the Future Conference, Boston, July 2009.
Imagine you could gain new insights to….
…predict …apply social …adjust credit …determine
regions where relationships of lines as discount levels for
doctors customers to transactions are select people at
prescribe high prevent churn? occurring to time of sale
volume of account for risk instead of
medication? fluctuations? offering to all?
Pharma Telco Call Loan Officer Retail Sales
Sales Center Rep Associate
Manager
52 © 2009 IBM Corporation
53. IEEE Accessing the Future Conference, Boston, July 2009.
SPSS Enables New Solution Value for IBM Cognos
Customers
How are Why are we What should
we doing? on/off track? we be doing?
Addition of KPPs
Addition of KPPs Broad distribution of
Broad distribution of Time series
Time series
(Key Performance
(Key Performance statistical results
statistical results forecasting
forecasting
Predictors)
Predictors)
New customer
New customer Predictive analytics for
Predictive analytics for
insight through
insight through deeper understanding of
deeper understanding of
Data Collection
Data Collection the data
the data
53 © 2009 IBM Corporation
54. IEEE Accessing the Future Conference, Boston, July 2009.
DEMO
DEMO
Video
Video
Traditional decision processes evolved
Traditional Approach Breakaway
Sense and Respond Predict and act
Back Office Point of impact
Skilled analytics experts Everyone
Instinct and Intuition Realtime fact driven
Automated Optimized
© 2009 IBM Corporation
55. IEEE Accessing the Future Conference, Boston, July 2009.
IBM SPSS Statistics is a comprehensive, easy-to-use set of data and
predictive analytics tools for business users, analysts and statistical
programmers.
SPSS Statistics Family
Linear models – make your analysis more accurate and reach more dependable conclusions
Nonlinear models – have the ability to apply more sophisticated models to your data
IBM SPSS Statistics Standard
Customized tables – quickly slice and dice your data using pivot tables
Data preparation – Prevent outliers from skewing analyses and results
Decision trees – Better identify groups, discover relationships between groups and
predict future events
IBM SPSS Statistics Professional
Forecasting – Deliver information in ways that your organization’s decision makers
can understand and use
Structural equation modeling - you can quickly create models to test hypotheses
Bootstrapping - Estimate the standard errors and confidence intervals of parameters
IBM SPSS Statistics Premium Direct marketing and product decision making procedures - Develop a marketing
strategy
High-end charts and graphs - Extend the capabilities of templates or create your own
Provide more flexible pricing and licensing options
Easily extend usage throughout the university
Foster a permanent link between academic and corporate institutions
IBM SPSS for Education Recognize IBM SPSS software users for their contributions to their respective industries
Support more effective teaching with IBM SPSS software
Ensure that students will be sought by employers
© 2009 IBM Corporation
56. IEEE Accessing the Future Conference, Boston, July 2009.
Product Family
Data Collection Modeller Deployment
Survey and market IBM® SPSS® Modeler is Drive results-oriented
researchers worldwide a powerful, versatile decisions by building
use this rich suite of data mining workbench analytics into your
products to achieve that helps you build operations. Integrate the
deeper understanding accurate predictive analytics that predict
of people’s attitudes, models quickly and outcomes. Automate
preferences and intuitively, without processes to deliver
behavior. programming insight at the point of
impact.
-Authoring
-Interviewing
-Reporting
-Management
© 2009 IBM Corporation
57. IEEE Accessing the Future Conference, Boston, July 2009.
Data Collection tools
Author Desktop Paper/Scan
Author Professional Phone Interviews
Remote Administration
Author Server
Survey Reporter Desktop
Base Professional
Survey Reporter Developer Kit
Data Entry Station
Survey Reporter Professional
Data Model
Survey Reporter Server
Dialer
Survey Tabulation
Interviewer Translation Utility
Interviewer Server Administrator Web Interviews
© 2009 IBM Corporation
58. Modeller
IEEE Accessing the Future Conference, Boston, July 2009.
IBM SPSS Modeler includes advanced, interactive visualization for models that use single technique, or ensemble
models that combine techniques making modeling results easy to understand and communicate.
© 2009 IBM Corporation
59. IEEE Accessing the Future Conference, Boston, July 2009.
Integration with Cognos 10
© 2009 IBM Corporation
60. IEEE Accessing the Future Conference, Boston, July 2009.
Deployment
© 2009 IBM Corporation
61. IEEE Accessing the Future Conference, Boston, July 2009.
Case study: Predictive Analytics on Human Capital Management
Problem:
Optimize recruitment effort for a given position (Corporate job, Military school, etc ..).
The volume of potential recruits or the intricacies of a specific job requirement can overwhelm
the efforts of even the best individual recruiter.
Solution:
Build a predictive performance model.
Apply the experience and intuition of expert recruiters in creating a model that helps an
organization to prioritize and target the individuals most qualified for a specific position.
Example:
One of the branches of the U.S. military is responsible for getting more than 100,000 new
recruits every year under contract. Approximately 600,000 leads that must then be prioritized
and sent to individual recruiters.
Baseline:
Predicting the success of a potential employee or recruit in a given work environment is
difficult, there are numerous variables that affect a successful outcome for that person’s
career. (Examples: changes in management, co-workers, and mission goals …. )
Reference Link:
http://forms.cognos.com/?elqPURLPage=4206&offid=sb_spssrc_human_capital_mgmnt_imw14291&mc=-web_ibm_spss_stat_products
© 2009 IBM Corporation
62. IEEE Accessing the Future Conference, Boston, July 2009.
Case study: Predictive Analytics on Human Capital Management (cont.)
Performance Prediction with SPSS Modeling:
- Collect Data (predictors)
- Data cleansing
- Eliminate Variables with low variance
- Eliminate var. with too many missing values
- Screen, rank and select predictor variables
- Rank the importance of each variable
Employee opinions and outlooks can be an IBM® SPSS® Modeler can consolidate data visually from multiple sources,
important predictor of performance. such as demographics data and attitudinal data.
Text Analytics and Text Mining with SPSS:
Example: Analysis of open ended questions to model
employee satisfaction
Provides a technical foundation for extracting usable
knowledge from unstructured text data through
identification of core concepts and sentiments. Text
analytics allows users to understand the
relationships between concepts and the sentiment
around concepts, and ultimately create a structure
for unstructured text data that can be integrated with A view into text analytics within IBM® SPSS® Modeler Premium. On the left is
analytics. a list of extracted categories and on the right is a visual representation of the
linkages between concepts and sentiments (sentiment analysis).
© 2009 IBM Corporation
63. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Portfolio – Key Products
… What’s the best that can happen ?
63 © 2009 IBM Corporation
65. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Discovering the Value of
Business Analytics
Where Do We Go From Here?
Lennart Frantzell alf@us.ibm.com Giuseppe Accardo gaccard@us.ibm.com, Chris Heckart checkart@us.ibm.com
IBM San Mateo Innovation Center, San Mateo, California 2011/04/27
© 2009 IBM Corporation
66. IEEE Accessing the Future Conference, Boston, July 2009.
Step 1) Read up on the IBM Products
http://www.redbooks.ibm.com/redbooks/pdfs/sg247912.pdf
http://www.redbooks.ibm.com/redpapers/pdfs/redp4710.pdf
http://www.redbooks.ibm.com/redbooks/pdfs/sg247881.pdf
© 2009 IBM Corporation
67. IEEE Accessing the Future Conference, Boston, July 2009.
Step 2) Install IBM Cognos Express
http://www.ibm.com/developerworks/downloads/im/cognosexpress/
© 2009 IBM Corporation
68. IEEE Accessing the Future Conference, Boston, July 2009.
Step 3) Join IBM PartnerWorld or IBM Academic Initiative
http://www.ibm.com/partnerworld
https://www.ibm.com/developerworks/university/academicinitiative
/
© 2009 IBM Corporation
69. IEEE Accessing the Future Conference, Boston, July 2009.
Step 4) Follow-on Business Analytics education at the San
Mateo Innovation Center
•Netezza Bootcamp (6/21-6/24)
•Cognos seminar
•SPSS seminar
•ILOG seminar
/
© 2009 IBM Corporation
70. IEEE Accessing the Future Conference, Boston, July 2009.
Step 5) Join IBM Social Networks, read Business Analytics
Blogs and the San Mateo IBM Innovation Center blog
https://www.ibm.com/developerworks/mydeveloperworks/groups/service/forum/topicThread?
topicUuid=45358eb2-315a-43e3-8e5f-5e94fd60009a#fullpageWidgetId=Members
https://www.ibm.com/developerworks/mydeveloperworks/blogs/business-analytics/?lang=en
https://www.ibm.com/developerworks/mydeveloperworks/blogs/iic-san-mateo/?lang=en
http://www-935.ibm.com/services/us/gbs/bao/
/
© 2009 IBM Corporation
71. IEEE Accessing the Future Conference, Boston, July 2009.
Reference Links:
Cognos:
http://www.reporters.cz/en/index.php?option=com_content&task=view&id=123&Itemid=168
SPSS:
http://www-01.ibm.com/software/analytics/spss/downloads/
http://www-01.ibm.com/software/analytics/spss/products/modeler/
http://www-01.ibm.com/software/analytics/spss/products/modeler/professional.html
http://support.spss.com/ProductsExt/Data%20Collection/ProductMatrix.html
iLOG:
http://www-01.ibm.com/software/websphere/ilog/
http://www-01.ibm.com/software/solutions/soa/newsletter/nov10/brms.html
Blog
http://www.b-eye-network.com/blogs/eckerson/archives/business_analyt/
March 24, 2011 © 2009 IBM Corporation
72. IEEE Accessing the Future Conference, Boston, July 2009.
BACKUP
SLIDES
© 2009 IBM Corporation
73. IEEE Accessing the Future Conference, Boston, July 2009.
Watson, the hardware
• Each of Watson’s 90 clustered IBM Power 750 servers features 32 POWER7 cores
running at 3.55 GHz.
• Running the Linux®operating system, the servers are housed in 10 racks along with
associated I/O nodes and communications hubs.
• The system has a combined total of 16 Terabytes of memory and can operate at over 80
Teraflops (trillions of operations per second).
• POWER7 also features 500 gigabytes of on-chip communications bandwidth, contributing
to exceptional efficiency of both memory and processor utilization. And since each server
packs 32 high performance POWER7 cores with up to 512 GB of memory, the Power 750
makes an ideal platform for Watson’s processor and memory-hungry Java processes.
• Designing Watson on commercially available Power 750 servers was a deliberate choice
to ensure more rapid adoption of optimized systems in industries such as healthcare and
financial services.
• That goal was a fundamental difference between Watson and Deep Blue, which was a
highly customized supercomputer. Deep Blue was based on an earlier generation of
Power processor technology, featuring a.But in addition to the regular POWER2
processors, Deep Blue’sperformance was enhanced with 480 special purpose chess
processor chips.
July 20, 2009 IBM Confidential © 2009 IBM Corporation
74. IEEE Accessing the Future Conference, Boston, July 2009.
Inside Watson
• Watson uses UIMA-AS to scaleout across 2,880 POWER7 cores in a cluster of 90 IBM
Power®750 servers.
• UIMA_AS manages all of the inter-process communication using the open JMS standard.
• The UIMA-AS deployment on POWER7 enabled Watson to deliver answers in one to six
seconds.
• Watson has roughly 200 million pages of natural language content (equivalent to reading 1
million books).
• Watson uses the Apache Hadoop framework to facilitate preprocessing the large volume
of data in order to create in-memory datasets used at runtime.
• Watson’s DeepQA UIMA annotators were deployed as mappers in the Hadoop map-
reduce framework, which distributed them across processors in the cluster.
The Regular Expression Annotator (RegexAnnotator)
is an Apache UIMA analysis engine that detects entities
like email addresses, URLs, phone numbers, zip codes or
any other entity based on regular expressions and concepts.
July 20, 2009 IBM Confidential © 2009 IBM Corporation
75. IEEE Accessing the Future Conference, Boston, July 2009.
Madrid First Responders Demo
© 2009 IBM Corporation
76. IEEE Accessing the Future Conference, Boston, July 2009.
Madrid’s emergency first responders
You're invited to take a ride with Madrid’s emergency first responders as they rush to the scenes
of three separate incidents.
In the wake of the 2004 Madrid bombings, the city implemented a business process
management solution from IBM to integrate the disparate applications, data and processes of
its various emergency departments.
IBM helped the city reduce emergency response times by 25%.
the ride.
http://www-03.ibm.com/innovation/us/leadership/response/index.html © 2009 IBM Corporation
77. IBM Watson and Healthcare. How natural language and semantic
IEEE Accessing the Future Conference, Boston, July 2009.
search could revolutionize clinical decision support
According to an observational study published in 1999 by BMJ (British Medical Journal)
a team of researchers observed 103 physicians over one work day. Those physicians
asked 1,101 clinical questions during the day. The majority of those questions (64 percent)
were never answered. And, among questions that did get answered, the physicians spent
less than two minutes looking for answers. Only two questions out of the 1,101 triggered
a literature search by the physicians attempting to answer them. Hence, providing quick
answers to clinical questions could have major impact in improving the quality of
healthcare. Enter Watson.
To see the kinds of questions Watson can answer, check out the two example questions Dr.
David Ferrucci showed to German Chancellor Merkel and Turkish PM Erdogan at the CeBIT
2011 Opening Ceremony..
Question: Streptococci cause this childhood "fever" characterized by a bright red rash and
high temperature.
Answer: 98% Scarlet fever, 15% Rheumatic fever, 8% Strep throat
Question: This disease can cause uveitis in a patient with family history of arthritis
presenting circular rash, fever, and headache.
Answer: 76% Lyme Disease, 1% Behcet's Disease, 1% Sarcoidosis
http://www.ibm.com/developerworks/industry/library/ind-watson/index.html © 2009 IBM Corporation
78. IEEE Accessing the Future Conference, Boston, July 2009.
IBM SPSS Statistics is a comprehensive, easy-to-use set of data and
predictive analytics tools for business users, analysts and statistical
programmers.
SPSS Statistics Family
Linear models – make your analysis more accurate and reach more dependable conclusions
Nonlinear models – have the ability to apply more sophisticated models to your data
IBM SPSS Statistics Standard
Customized tables – quickly slice and dice your data using pivot tables
Linear models Nonlinear models
• General linear models (GLM) • Multinomial logistic regression (MLR)
• Generalized linear mixed models (GLMM) • Binary logistic regression
• Hierarchical linear models (HLM) • Nonlinear regression (NLR) and constrained
• Generalized linear models (GENLIN) nonlinear regression (CNLR)
• Generalized estimating equations (GEE) • Probit analysis
Customized tables
IBM SPSS Statistics Standard enables you to quickly “slice and
dice” your data. Then you can create customized tables to help
you better understand your data and easily report your results.
© 2009 IBM Corporation
79. IEEE Accessing the Future Conference, Boston, July 2009.
IBM SPSS Statistics is a comprehensive, easy-to-use set of data and
predictive analytics tools for business users, analysts and statistical
programmers.
SPSS Statistics Family
Data preparation – Prevent outliers from skewing analyses and results
Decision trees – Better identify groups, discover relationships between groups and
predict future events
IBM SPSS Statistics Professional
Forecasting – Deliver information in ways that your organization’s decision makers
can understand and use
Data preparation Decision trees
IBM SPSS Statistics Professional helps you streamline the Create classification and decision trees to help you better
data preparation stage of the analytical process – saving identify groups, discover relationships between groups and
time and ensuring greater accuracy. Perform data checks predict future events. Decision trees present categorical
based on each variable’s measure level, quickly find results in an intuitive manner, allowing you to explore
multivariate outliers by searching for unusual cases based results and visually determine how your model flows, and
upon deviations from similar cases and preprocess data then clearly explain categorical results to non-technical
prior to model building with an optimal binning procedure. audiences. You can also find specific subgroups and
relationships that you might not uncover using more
traditional statistics.
Forecasting
Predict trends and develop forecasts quickly and easily
with advanced statistical techniques to work with time-
series data. Regardless of your level of experience, you
can analyze historical data, predict trends faster and
deliver information in ways that your organization’s
decision makers can understand and use. © 2009 IBM Corporation
80. IEEE Accessing the Future Conference, Boston, July 2009.
IBM SPSS Statistics is a comprehensive, easy-to-use set of data and
predictive analytics tools for business users, analysts and statistical
programmers.
SPSS Statistics Family
Structural equation modeling - you can quickly create models to test hypotheses
Bootstrapping - Estimate the standard errors and confidence intervals of parameters
IBM SPSS Statistics Premium Direct marketing and product decision making procedures - Develop a marketing
strategy
High-end charts and graphs - Extend the capabilities of templates or create your own
Structural equation modeling Bootstrapping
Structural equation modeling (SEM) can help you gain provides an efficient way to ensure that your models are
additional insight into causal models and explore the stable and reliable. It estimates the sampling distribution of
interaction effects and pathways between variables. SEM an estimator by re-sampling with replacement from the
lets you more rigorously test whether your data supports original sample. With bootstrapping, you can reliably
your hypothesis. You create more precise models than if you estimate the standard errors and confidence intervals of a
used standard multivariate statistics or multiple regression population parameter, including the mean, median,
models alone. proportion, odds ratio, correlation coefficient, regression
coefficient and numerous others.
Direct marketing and product decision-making procedures
Quickly perform various kinds of analyses, including recency, frequency and monetary value (RFM)
analysis, cluster analysis and prospect profiling. Increase your understanding of consumer preferences to
more effectively design, price and market successful products – maximizing campaign effectiveness and
return on investment.
© 2009 IBM Corporation
81. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization Portfolio
IBM acquisition landscape
81 © 2009 IBM Corporation
83. Business Analytics - acquisition landscape
IEEE Accessing the Future Conference, Boston, July 2009.
Coremetrics, is a leader in Web analytics software. Coremetrics, based in San Mateo,
CA, will expand IBM's business analytics capabilities by enabling organizations to use
cloud computing services to develop faster, more targeted marketing campaigns.
Unica is an enterprise and cloud-based marketing software solutions that help businesses
streamline and automate marketing processes, and understand and predict customer preferences.
Through Unica, IBM will enable its clients to develop more relevant and targeted communications
while minimizing marketing expenditures.
OpenPages, a leading provider of software that helps companies more easily identify
and manage risk and compliance activities across the enterprise through a single
management system.
Clarity Systems delivers financial governance software that enables organizations to automate the
process of collecting, preparing, certifying and controlling financial statements for electronic filing, in
support of mandates by the SEC and other financial regulatory agencies.
Netezza data warehouse appliances bring analytics directly into the hands of business
users within every department of an organization such as sales, marketing, product
development and human resources. Netezza appliances makes the technology ideal for
the needs of high-performance analytics, requiring minimal administration and IT skills,
and enables clients to run complex data queries within days of deploying the solution.
Initiate's software helps healthcare clients work more intelligently and efficiently with timely
access to patient and clinical data. By adding Initiate's software to its software portfolio, IBM
will be better equipped to help clients draw on data from hospitals, doctors' offices and payers
to create a single, trusted shareable view of millions individual patient records.
Guardium, a market leader in real-time enterprise database monitoring and
protection. Guardium's technology helps clients safeguard data, monitor database
activity and reduce operational costs by automating regulatory compliance tasks.
© 2009 IBM Corporation
84. IEEE Accessing the Future Conference, Boston, July 2009.
IBM Business Analytics and Optimization
Portfolio – Key Products
… What’s the best that can happen ?
84 © 2009 IBM Corporation
85. IEEE Accessing the Future Conference, Boston, July 2009.
Where it fits
85
85 © 2009 IBM Corporation
86. IEEE Accessing the Future Conference, Boston, July 2009.
a recognized industry leader in Business Rule Management Systems (BRMS),
visualization components, optimization and supply chain solutions enrich IBM
software portfolio and fortify IBM's Smarter Planet initiative.
WebSphere ILOG BRMS Optimization and
Analytical Decision
Support Solutions
WebSphere ILOG BRMS Family
WebSphere ILOG JRULES
CPLEX Optimization Studio
WebSphere ILOG LogicNet Plus XE
Visualization
ability for non-technical business create the best possible plans,
users to be directly involved in Elixir Enterprise
explore alternatives,
business rules management, understand trade-offs, and
enabling flexible decision JView Enterprise respond to changes in business
automation. environment
industry’s most comprehensive set of
graphics products for creating highly
graphical, interactive displays.
© 2009 IBM Corporation
87. IEEE Accessing the Future Conference, Boston, July 2009.
What is a Business Rules Management System BRMS?
A business rule management system
(BRMS) enables organizational policies
to be defined, deployed, monitored and
maintained separately from core
application code. By externalizing
business rules and providing tools to
manage them, a BRMS allows business
experts to define and maintain the
decisions that guide systems behavior,
reducing the amount of time and effort
required to update production systems,
and increasing the organization’s
ability to respond to changes in the
business environment.
© 2009 IBM Corporation
88. IEEE Accessing the Future Conference, Boston, July 2009.
Why Business Event Processing (BEP) matters?
Business Event Processing describes a wide range of ways that enterprises
approach events, simple or complex. But in all cases, information about the event
needs to be quickly disseminated to others affected by the event for both
awareness and to take appropriate action.
DEMO
DEMO
Video
Video
© 2009 IBM Corporation
89. IEEE Accessing the Future Conference, Boston, July 2009.
Visualization
Diagrams
Platforms: Gantt Charts
Java
Maps
.Net Business DashBoard
Adobe Flex Charts
User Interfaces
C++
© 2009 IBM Corporation
90. IEEE Accessing the Future Conference, Boston, July 2009.
What is ILOG Optimization ?
A software based solution that enables enterprises to create the best possible plans,
explore alternatives, understand tradeoffs and respond to changes in the business
environment
IBM ILOG optimization maximizes resource efficiency
• By helping companies make Decisions
• To reach a Goal
• While observing Requirements
• Determined by Analyzing Data
Using powerful, robust, scalable and diversified optimization software and services
Requirements
Requirements
Decisions
Decisions
Bus. Rules
Bus. Rules Plans – alternatives - tradeoffs
Plans – alternatives - tradeoffs
Goals
Goals
Data
Data
© 2009 IBM Corporation
91. IEEE Accessing the Future Conference, Boston, July 2009.
What optimization can do?
Optimization helps businesses make complex decisions and trade-offs about
limited resources
• Discover previously unknown options or approaches
• Automatically evaluate millions of choices
• Automate and streamline decisions
• Compliance with business policies and regulations
• Free up planners and operations managers so that they can leverage their
expertise across a wider set of challenge
• Explore more scenarios and alternatives
• Understand trade-offs and sensitivities to various changes
• Gain insights into input data
• View results in new ways
© 2009 IBM Corporation
92. IEEE Accessing the Future Conference, Boston, July 2009.
Optimization based problems
They exist in all industries…
© 2009 IBM Corporation
93. IEEE Accessing the Future Conference, Boston, July 2009.
Optimization based problems
… and are critical for the companies !
© 2009 IBM Corporation
94. IEEE Accessing the Future Conference, Boston, July 2009.
Success Story – Unit Commitment at REE
Business Problem – Use exact mathematical methods to replace the
approximate, heuristic methods Red Eléctrica de España, in charge of
managing the Spanish national power grid, had been using for the last 20 years
The methodology applied until now was an
interactive methodology, which did not
guarantee an optimum solution. There were
many difficulties in the smaller systems and it
was hard to find the most viable solution.
Thanks to the new methodology, we have
resolved this type of problem.
- Mr. Mustafa Pezic, REE Project Director
© 2009 IBM Corporation
95. IEEE Accessing the Future Conference, Boston, July 2009.
Benefits
• The implementation of the ILOG based solution has provided great operational
advantages to company’s managers and engineers
– “The new tool allows us to simplify all maintenance tasks and any changes made to the model, which
in our particular case, are very frequent.”
– “From a user viewpoint, it has brought greater trust in the solution and a significant reduction in
planning time required by users. In parallel with this, from a development and maintenance viewpoint,
there has been a significant reduction in associated costs, as well as in the duration of the
processes.”
• The bottom line:
– REE reduced production costs by between €50,000 and €100,000 per day.
– REE has reduced its carbon emissions by approximately 100,000 tons of CO2 annually.
© 2009 IBM Corporation
96. Saving $140,000 Per Day:
How Companies are Achieving
Breakthrough Improvements in Bottom-
Line Performance Using Optimization
Dr. Jeremy Bloom
Product Marketing Manager, ILOG
Optimization
May, 2010
97. The Story In Brief
Better decisions faster
• IBM ILOG Optimization Products are Helping Many
Businesses Run More Efficiently
• IBM ILOG Optimization Uses Sophisticated Technology to
Solve Hard Business Problems
• IBM ILOG Optimization Products and Services Can Help Your
Business Run More Efficiently
• IBM ILOG Optimization Can Generate Hard Benefits to Your
Bottom Line
2
98. What Can Optimization Do?
increased productivity at Europe’s most efficient car production
Automobile Manufacturer
facility by 30%
• South American country’s
two largest forest-products reduced their truck fleets by 30% and saved $20 million annually
companies
• Major Electronics
cut wafer-processing cycle time in half, to just 30 days
Manufacturer
responded to unexpected delays with efficient crew rescheduling,
International airline
saving $40 million in one year
cut package delivery costs by $87 million over 2 years and reduced
Package delivery company
its aircraft fleet by 10%
Television network increased annual advertising revenue by $50 million
Investment firm cut transaction costs by $100 million
Consumer packaged goods dramatically increased the direct loading of trucks off its packaging
manufacturer lines
3
99. What Can Optimization Do?
• Whether the problem is large or
small, straightforward or
complex,
optimization supports effective
decision-making across a wide
range of issues.
• Firms in many industries use
optimization software to solve
business problems ranging from
long-term planning to real-time
scheduling and rescheduling.
4
101. Benefits of Optimization
• Calculable ROIs, with paybacks within months, sometimes even
weeks
– Capital expense avoidance or deferral
– Operating expense reductions
– Total revenue, revenue mix, and margin improvements
• Improved customer satisfaction
– Provide better and more customized customer service
• Improved employee satisfaction
– Satisfy schedule preferences while improving productivity
– Better planning and scheduling processes
6
102. Sophisticated Optimization Technology Solves Hard
Business Problems
• IBM ILOG Optimization helps businesses maximize resource
efficiency
– by helping companies make Choices
– to reach Targets
– while observing Limits
– driven by analyzing Data
• Using powerful, robust, scalable, and diversified optimization
technology and services
– Optimization has most value when there are many choices with
complex relationships that force trade-offs
7
104. Case Study
Cash Management:
Restocking Automatic Teller
Machines
9
105. Restocking Automatic Teller Machines
The Customer
• Provides financial electronic commerce services and
products to financial institutions worldwide
• Provides systems processing more than two-thirds of 14
billion annual automated clearing house transactions in the
US
• Provides reconciliation, financial messaging, workflow and
compliance products and services to more than 600 banks
and businesses
• Its clients manage more than 2.6 million portfolios totaling
about US $1.8 trillion in assets
10
106. Restocking Automatic Teller Machines
The Business Problem
Schedule restocking taking into account customer withdrawal
habits and government cash management regulations
• Too much cash some times – carrying costs
• Too little cash at other times – angry customers
• Forecast errors – volatility
• Data errors – static, dirty, missing, wrong!
11
107. Restocking Automatic Teller Machines
Vaults as Distribution Centers
• Services: counting, verifying, sorting, packaging, shipping
• Federal Reserve Regulations
– Cross-shipping penalties
– Custodial Inventory: De Minimis Exemptions, Fitness Issues, etc.
• Banks Organize Vaults Geographically by FRB zone
– 33 Zones in US
– From 2 to 12 Vaults per Zone
• High Service Levels
– Due to nature of product (cash) and customer (ATM’s and bank
branches)
– Substantial business case for optimization solution
12
108. possible Day 1 Day 2 Day 3 Day 4
solution
+10 -10 +40
10 0
v1 v1 v1 v1
40
20
+10 -50 +20 FED
10 0 10 DEPOSITS
FED
ORDERS
10
v2 v2 v2 v2
20 10
+10 +10 0
10 -10
v3 v3 v3 v3
Note: Uses 4 trucks
13
13
109. Restocking Automatic Teller Machines
Business Case Synopsis: Top-10 Bank Client
• Daily Retail Cash Dispensed
– $ 200 million (+20,000 retail outlets - Branches & ATM’s)
• Total Cash in System (before optimization)
– $ 7 billion
• Optimization Development Goals
– No change of current replenishment schedules
– Reduce cash inventory levels (i.e. carrying costs)
– Reduce replenishment costs (i.e. deliveries)
– Reduce cross-shipping costs (penalties at Fed)
– Improve reporting capability (information)
– “Piggybacking” fixed-charge denomination shipments
– Must solve overnight for implementation next day 14
110. Restocking Automatic Teller Machines
The Bottom Line: Results After 6 Months
• 58 Vault Pilot
• Reduced cash inventories by 35%*
• Reduced replenishment costs by 55%
• Cross-shipping fees decreased about 63%
• CPLEX runtimes within overnight window
• Project rated “Highly Successful” by client’s internal Six Sigma
Unit
• Rolled-out to entire enterprise in 2008
* Attributable to the optimization model and other factors including better forecasting, better
operations, better people, and better measurement.
15
111. Restocking Automatic Teller Machines
What the Customer Says:
• “Our OPL model solved by CPLEX has proven to be a
powerful platform from which advanced uses of MILP can be
studied, showcased, and advanced. Several successful
efforts have been accomplished thus far with respect to
speed improvements, always the challenge for us.”
• “We like IBM ILOG’s people, and the reason we like them is
we could call people up and talk to intelligent, well-versed,
experienced people who either could answer our questions
directly or could point us to a resource that could answer our
questions.”
16
112. Case Study
Transportation
Scheduling:
Train Timetabling
17
113. Train Timetabling
The Customer
• Netherlands Railways
• Operates the busiest national railway network in Europe
• Manages more than 4,800 trains per day
• Has 2,100 km of track and 279 stations
• Between 1970 and 2006, traffic has nearly doubled
from 8 billion passenger km in to 15.8 billion
• During the same period, freight transport increased by
285 percent
In 2006,
• 9 million different passengers traveled 15.8 billion
passenger km
• Operating revenues of €1.5 billion and operating
income of €200 million 18