Business Analytics og Optimization, BI på superoktan (IBM Global Business Services)

Business Analytics & Optimization: BI på superoktan

Ivar JuulPartner, IBM GBS, Nordic BAO Leader

© 2010 IBM Corporation

Agenda

� What is BAO

� Having data simply is not enough

� Advanced Analytics have many forms– Checkout Analytics Surveillance Engine– Fraud and Abuse Management System

� Develop a business case for Advanced Analytics

� What problems are our customers facing

� Consolidation of the data warehousing platform


How IBM’s Business Analytics & Optimization practice helps clients

BAO Strategy

Business Intelligence & Performance Management

Advanced Analytics and Optimization

Enterprise Information

Management

Enterprise Content

Management

• Identify and prioritize opportunities for improvement

• Change business processes and operations to exploit analytics

• Implement management systems to maintain control and achieve goals

• Report outcomes of business processes and programs

• Automate management dashboards and scorecards

• Create planning, budgeting, & forecasting tools

• Apply advanced statistical and regression analysis upon historical data for predictive decision-making

• Integrate optimization algorithms and technology into operations

• Ensure robust and trusted data is available when needed and is easy to consume

• Provide a consolidated and efficient information platform to support optimization initiatives

• Manage document & records, including archives

• Manage structured and unstructured content

• Manage digital assets & rights

• Provide efficiency and transparency to complicated workflows


When having data simply is not enough:A Post September 11th

Relationship Analysis of the Hijackers

© 2010 IBM CorporationTransforming the Business of Government9

2 Known Terrorists in US

WATCH LIST: CIA/INS/FBI POSSIBLE TERRORISTS IN THE US:

� On or before August 23rd, 2001, Nawaq Alhamzi and Khalid Al-Midhar added to INS watch list

MAKE PLANE RESERVATIONS USING SAME NAMES:

� On or about August 25, 2001, Khalid Al-Midhar purchases cash ticket for American Airlines flight #77 scheduled for September 11, 2001

� On or about August 27, 2001, Nawaq Alhamzi books a flight on American Airlines flight #77 scheduled for September 11, 2001


Two of the Nineteen Terrorists Known

United Airlines Flight 175

Target: World Trade CenterSouth Tower

Marwan Al-ShehhiAhmed Alghamdi Fayez Ahmed Mohand Alshehri Hamza Alghamdi

Hani HanjourSalem Al-HazmiNawaq Alhamzi Majed MoqedKhalid Al-Midhar

American Airlines Flight 77

Target:

Pentagon

Mohamed Atta Satam Al SuqamiAbdulaziz Alomari Waleed AlshehriWail Alshehri


Target: World Trade CenterNorth Tower


Target: Unknown

Saeed Alghamdi Ahmed AlnamiAhmed Al Haznawi Ziad Jarrah


Address Connections

RESERVATIONS MADE WITH ADDRESS #1 AND ADDRESS #2

� On or about August 25, 2001, Khalid Al-Midhar makes a reservation on American Airlines flight #77 scheduled for September 11, 2001 using Common Address #1

� On or about August 27, 2001, Nawaq Alhamzi books flight on American Airlines flight #77 scheduled for September 11, 2001 using Common Address #2

� ADDRESSES ARE USED BY THREE (3) ADDITIONAL PASSENGERS

� Mohamed Atta has reservation on American Airlines flight #11 scheduled for September 11, 2001 using Common Address #1 as a contact address

� Marwan al-Shehhi has reservation on United Airlines flight #175 scheduled for September 11, 2001 using Common Address #1 as a contact address

� Salem Alhamzi has reservation on American Airlines flight #77 scheduled for September 11, 2001 using Common Address #2 as a contact address


Five of the Nineteen Terrorists Linked






Target:

Pentagon





Target: Unknown



Phone Number Connections

ONE (1) ALERTED PASSENGER MAKES RESERVATION USING COMMON

TELEPHONE NUMBER

� On or about August 28, 2001, Mohamed Atta uses Florida Telephone #1 as a contact number when making reservations on American Airlines flight #11 scheduled for September 11, 2001

NUMBER IS USED BY FIVE (5) ADDITIONAL PASSENGERS

� On or about August 26, 2001, Waleed Alshehri and Wail Alshehri make reservations on American Airlines flight #77 scheduled for September 11, 2001 using Florida Telephone #1 as a contact number

� On or about August 27, 2001, reservations for electronic, one-way tickets were made for Fayez Ahmedand Mohand Alshehri for United Airlines flight #175 using Florida Telephone #1 as a contact number

� On or about August 28, 2001, Abdulaziz Alomari reserves a seat on American Airlines flight #11 using Florida Telephone #1 as a contact number


Ten of the Nineteen Terrorists Linked






Target:

Pentagon





Target: Unknown



Frequent Flyer Connections

ONE (1) ALERTED PASSENGER MAKES RESERVATION USING A

FREQUENT FLYER NUMBER

� On or about August 25, 2001, Khalid Al-Midhar makes a reservation on American Airlines flight #77 scheduled for September 11, 2001 using Frequent Flyer #1

FREQUENT FLYER NUMBER IS USED BY ONE (1) ADDITIONAL

PASSENGER

� On or about August 25, 2001, Majed Moqed makes a reservation on American Airlines flight #77 scheduled for September 11, 2001 using Frequent Flyer #1


Eleven of the Nineteen Terrorists Linked






Target:

Pentagon





Target: Unknown



Remaining ConnectionsPUBLIC RECORDS

� Alerted subjects Nawaq Alhamzi and Khalid Al-Midhar lived with Hani Hanjour

� Alerted subject Wail Ashehri was roommates and shares PO Box with Satan Al Suqami

WATCH LIST: INS ILLEGAL/EXPIRED VISAS

� On or about August 29, 2001, Ahmed Alghamdi reserves an electronic one-way ticket on United Airlines flight #175 scheduled for September 11, 2001

FIVE (5) ADDITIONAL PASSENGERS

� Alerted subject Ahmed Alghamdi and Hamza Alghamdi both use same address on their airline reservations

� Alerted subject Hamza Alghamdi has/does live with Saeed Alghamdi, Ahmed Alhaznawi, and Ahmed Alnami

� Alerted subject Ahmed Alhaznawi has/does live with Ziad Jarrah


All Nineteen Terrorists Linked






Target:

Pentagon





Target: Unknown



Nawaq AlhamziKhalid Al-Midhar Ahmed Alghamdi

Marwan Al-Shehhi Hani HanjourSalem Al-HazmiMajed Moqed Mohamed Atta Hamza Alghamdi

Fayez Ahmed Mohand Alshehri Abdulaziz AlomariWaleed Alshehri Wail Alshehri Saeed Alghamdi Ahmed AlnamiAhmed Alhaznawi

Satam Al Suqami Ziad Jarrah

Only Three Degrees of Separation Links All Nineteen 9/11 Hijackers


The world is becoming smaller and flatter,but also needs to be smarter.


Agenda

� What is BAO







All Scan Avoidance Transactions Displayed

Video Frame linked to each item


Checkout Analytics Surveillance Engine

� POS Data, Video, Analytics – Bringing it all together

– Detect scan actions visually

– Match up visuals scans with barcode reads

– Identify visual scans which have no barcode read = “Noscan”

– Insert Noscans into data mining database

� Trial results

– 15 cases of passing

– 10 cases of Void or Over charging error

– 2 operational issues (1 potential staff

discount fraud)

– 3 cashiers with more than 1 type of error

� Simple ROI

– £75 in 3 hours across 10 lanes– £75 * 3 hour loss total store = £225 – Assume 12 hours of operation = £900 per

store per day– 364 day year = £327,600


Agenda

� What is BAO







ProviderProvider

ProfilesProfiles

DownstreamDownstream

BusinessBusiness

ProcessesProcesses

Profile Analysis

•Provider Rosters•Profile Values & Scores•Profile & Claims Reports

InvestigatorsInvestigators

FAMS SystemFAMS System

Analysis Results

FAMS Solution

FAMSFAMSProfile ScoringProfile Scoring

Profile AnalystsProfile Analysts

FAMS SystemFAMS System

TransactionTransactionHistoryHistory

Fraud and Abuse Management System

Why Base FAMS?Why Base FAMS?

� Detection of anomalous behavior

� Developed with innovative industry leaders

� Proven technology

� Significant levels of success achieved by clients

� Designed for ease of use for broad range of users

� Powerful Spectrum of Integrated Analytical and Reporting Components

FAMS is a full function retrospective analysis solution


Agenda

� What is BAO







Predict Behavior & PreferencesIBM SPSS Modeler for Modeling Future Behavior

� Maximize Analyst Output– Easy to learn, no programming approach to data

mining

� Create Practical Solutions Faster– Automatically create accurate, deployable predictive

models– Choose the best solution with multi- model evaluation

� Open & Efficient Architecture– Data mining within standard databases– Multithreading, clustering and use of embedded

algorithms

Building a business case:• Is there a pattern in the data and is it usable ?

• Before investing in SPSS we can analyse the data for patterns• Develop a business case

• Fast way of getting business insight


Agenda

� What is BAO







What pain points do our customers have

� Missing loading predictability� Long response times� Too many errors promoted to production� Funding only for per project, not for architecture� Lack of confidence in realizing benefits from BI

investments� Inconsistent measurement/data definition� No mechanism to correct errors in source

systems� Too many “standard” BI tools � Many uncoordinated and redundant data

mart/data warehouse efforts� Data strategy and data management initiatives

not linked to business priorities and needs� No common framework to address

integration/communication needs across organization boundaries

� No roadmap to identify integration points across organizations, project types, or logically sequence re-engineering and/or transformation initiatives in the company

� No center of competency to execute/support BI initiatives

� Confusion over lack of common language around needs, solutions, and approaches

� No BI technology standards� Inability to recognize and manage cross initiative

dependencies� Inability to leverage past investments � Inability to anticipate BI needs� Multiple versions of the truth� Lack of improvement in cycle time to meet new

BI needs� Not effective leverage skills and resources

across BI initiatives


Loading issues

� Many data warehouse installations have problems with

– Loading predictability– Keeping the batch window open

� No procedures for follow up on errors in the source systems effecting the confidence

� Loading is a project related issue

Records per Day

0

100.000.000

200.000.000

300.000.000

400.000.000

500.000.000

600.000.000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Date

No o

f R

eco

rds

2005-01

2006-01

2007-01

Number of loaded COPA records (2006)

0

100000

200000

300000

400000

500000

600000

700000

800000

1 2 3 4 5 6 7 8 9 10 11 12

Month

Reco

rds High

Low

Average

High in month 6 & 11 has been removed

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 2 3 4 5 6 7 8 9 10

Month

Red

Amber

Green

Records loaded by day in January (3 years)

Variations in number of loaded records Loading predictability per month


Agenda

� What is BAO







Annual cost saving of around $100 million in an American Bank

BI AnalyticsSources

Supply side Consumption side

Data Repository

Reference Data Views

CUSTOMER

PRODUCT

INDUSTRY

ASSOCIATE

ORGANIZATION

Reference Data Views

CUSTOMER

PRODUCT

INDUSTRY

ASSOCIATE

ORGANIZATION

Semantic Layer

Data Governance / Data Quality

Metadata

Security and Data Privacy

Systems Management & Administration

Hardware & Software Platforms

$70MM - $85MM $45MM - $55MM$30MM -$35MM

Landing Zone

DLZ

Hard Exceptions

ProcessMonitor & Control

Logging

Private Staging Area

Translated / Re-formatted Data

Original Data

Soft Exceptions

Certified Business

Data

Data Integration

$12MM - $25MM9

11

45

10

7

6

6

8

3

1

24

1

Supply:1. Data integration hub2. Consolidate warehouse model3. Consolidate external data4. Sunset existing IM environment5. Design new IM data integration COE

Consumption:6. Design semantic layer7. Migrate SAS functions to the

database8. Rationalize BI software9. Design new IM org model –

reporting tools (BI COE)

Cross:10. Design

governance Frameworks

11. Metadata layer

Tra

ns

form

ati

on

In

itia

tive

s

Estimated Direct Annual Cost SavingsLegend:


Cost Take Out

� Tangible cost reductions– Disk requirements– MIPS consumption and CPUs requirements– License cost– Operations cost

• Reduction of operations cost in monitoring and managing loading

• Reduction of hardware requirements

� Intangible cost reductions– Development time and resource requirements

• Easier access to data via the semantic layer• Documented meta data• Increased reuse

– Maintenance costs– Higher loading predictability– Fewer errors due to better governance

The projects had a positive cash flow after 9 and 12 months and IBM worked

on a risk sharing contract with IBM Global finance involved

Core

InitiativesDescription

ETL Consolidate the ETL and batch

processes via a methodical

engineering approach

6% $7.750.000 22% $15.575.000

Data Quality Formalize a data quality effort to

minimize the risk of errors9% $11.900.000 4% $2.775.000

Master and

reference data

Enable a metadata strategy that

includes a repository, user

interface and creation of

metadata artefacts

9% $11.700.000 5% $3.760.000

Data model Rationalize the current data

objects to a single logical

conformed data model

3% $3.300.000 12% $8.775.000

Data marts Sunset data marts in the analytic

layer36% $46.400.000 22% $15.225.000

Reporting/anal

ytics tools

Consolidate analytics tools and

rationalize external data8% $10.800.000 5% $3.775.000

SAS Re-platform, consolidate and

governor the SAS environment6% $7.600.000 0% $0

Analytics Deploy a consolidated analytics

layer to remove the complexity

and reduce redundancies15% $18.900.000 20% $14.445.000

Governance Initiate a governance program to

manage changes, data

definitions and processes

associated with the analytics

solution

9% $11.700.000 9% $6.375.000

100% $130.050.000 100% $70.705.000

American Bank 1 American Bank 2

Total


Example: Business case for re-implementation of DW

� The client is running 6 data warehouses across the Scandinavian countries

� Business case for data warehouse transformation project

– One unified data warehouse based on IIW– Migration of all historical data

� Business case– Cost is implementation cost excl. internal

hours– Benefits does only include tangible benefits

like• Reduction of MIPS and CPU capacity• Disk capacity• Reduction of license cost• Reduction of operation costs• Easy of solvency II implementation• Reduction of man power due to easier

access to data and economy of scale benefits

– The business case does not include• Better reporting• Single point of truth

-60

-40

-20

0

20

40

60

80

100

120

140

Year 1 Year 2 Year 3 Year 4 Year 5 Year 6

© 2010 IBM Corporation35

© Copyright IBM Corporation 2010 All rights reserved. The information contained in these materials is provided for informational purposes only, and is provided AS IS without warrantyof any kind, express or implied. IBM shall not be responsible for any damages arising out of the use of, or otherwise related to, these materials. Nothing contained in these materials is intended to, nor shall have the effect of, creating any warranties or representations from IBM or its suppliers or licensors, or altering the terms and conditions of the applicable license agreement governing the use of IBM software. References in these materials to IBM products, programs, or services do not imply that they will be available in all countries in which IBM operates. Product release dates and/or capabilities referenced in these materials may change at any time at IBM’s sole discretion based on market opportunities or other factors, and are not intended to be a commitment to future product or feature availability in any way. IBM, the IBM logo, Cognos, the Cognos logo, and other IBM products and services are trademarks of the International Business Machines Corporation, in the United States, other countries or both. Other company, product, or service names may be trademarks or service marks of others.

Technology

Business Analytics og Optimization, BI på superoktan (IBM Global Business Services)