Swoc21 Feb08 Amig
Transcript of Swoc21 Feb08 Amig
SWOC DAMA 2008 Showcase atSWOC DAMA 2008 Showcase at American Modern Insurance American Modern Insurance
February 21, 2008
Showcase AgendaShowcase Agenda Background/Business Case
20 minutes Sandy Wagner
Data Warehouse – AIIM 20 minutes Latha Subramanian
Data Model – AIIM 20 minutes Duke Ganote
Information Management – AIIM20 minutes Dan Daly
Q& A – Duke/Sandy/Latha/Dan 20 minutes
American Modern InsuranceAmerican Modern InsuranceCompany BackgroundCompany Background
Founded in 1938 as a consumer finance company
Provider of highly focused, specialty insurance products
Positioned to grow into a multi-billion dollar organization
Entrepreneurial spirit & deep commitment of employees
Approximately 1200 employees country-wide, with 1000
employees in eastern Cincinnati area (Amelia)
American Modern InsuranceAmerican Modern InsuranceCompany BackgroundCompany Background
The organization believes that the strategic The organization believes that the strategic deployment of technology can help it achieve, and deployment of technology can help it achieve, and sustain, a competitive advantage. sustain, a competitive advantage.
As stated in its Operating Principles, “Our As stated in its Operating Principles, “Our investment in information technology is part of a investment in information technology is part of a carefully planned strategy to ensure that American carefully planned strategy to ensure that American Modern's company-wide infrastructure is among Modern's company-wide infrastructure is among the most advanced in the specialty insurance the most advanced in the specialty insurance industry.” industry.”
American Modern InsuranceAmerican Modern InsuranceInitiative BackgroundInitiative Background
In 2000, American Modern embarked upon long-range initiative, coined “modernLINK,”
Business and IT collaboration Business case and funding
Three prongs: Web-enable insurance transaction processing Replace aging legacy processing systems Develop a Knowledge Management architecture
American Modern InsuranceAmerican Modern InsuranceBusiness CaseBusiness Case
The anticipated returns of this business case were: 20% annual increases in directly-attributed new
business 37% of Policy and Partner Administration moved from
existing internal units directly to point of service 25% improvement in current Product Review and
Management cycle time 21% improvement in Product Filings cycle time 2% reduction in total loss ratio directly attributed to
modernLINK initiative
American Modern InsuranceAmerican Modern InsuranceBusiness CaseBusiness Case
These returns would yield a significant recurring annual benefit through additional premium, increased profit, and decreased expenses
Almost 50% of these benefits would be attained through better knowledge/data management, richer data segmentation, and improved data and risk selection
John Hayden, President and CEO, American Modern states: We must have accurate data about the risks we insure today if
we are to ever be successful in establishing The Right Rate for Every Risk we choose to insure in the future.
American Modern InsuranceAmerican Modern InsuranceKnowledge Management RoadmapKnowledge Management Roadmap Enterprise Data Model Operational Data Store Enterprise Data Warehouse Themed analytic data marts Enterprise reporting portal Metadata management Data Stewardship
American Modern InsuranceAmerican Modern InsuranceKnowledge Management ResultsKnowledge Management Results
Business users can: Make informed decisions Respond quickly to new business initiatives Create new opportunities
Business users are: Moving from data collectors to data consumers Asking “why” instead of “what”
American Modern InsuranceKnowledge Management Results
Retention – Joe David. In the last four years, we have leveraged the corporate reporting tools to develop a series of targeted strategies that have allowed us to improve retention by nearly eight points, which equates to annualized premium of nearly $60 million
Claims. Integration of 3rd party Claim data - Heather Bolyard. This one-month sample of data for one material has identified a potential indemnity reduction of $70,000.
Reserving – Gene Stetler. The new Loss Reserving data store from the Enterprise Data Warehouse has enabled process efficiencies, thus allowing us to predict our reserving needs with accuracy.
Product – Kevin Randall. The implementation of American Modern's data warehouse has been a significant part of the successful launch of the company's right rate for every risk initiative
American Modern Insurance2007 Awards and Recognition
In 2007, American Modern received two awards from Computerworld:
Laureate - The laureate status for the Enterprise Data Warehouse presented at the Carnegie Mellon Auditorium in Washington D.C – June 2007
BI Award - Best Practices in Business Intelligence in the category “Creating an Agile BI Infrastructure” presented in Las Vegas, NV – September 2007
Showcase AgendaShowcase Agenda Background/Business Case Background/Business Case
20 minutes Sandy Wagner20 minutes Sandy Wagner
Data Warehouse – AIIM Data Warehouse – AIIM 20 minutes Latha Subramanian20 minutes Latha Subramanian
Data Model – AIIM Data Model – AIIM 20 minutes Duke Ganote20 minutes Duke Ganote
Information Management – AIIMInformation Management – AIIM20 minutes Dan Daly20 minutes Dan Daly
Q& A – Duke/Sandy/Latha/Dan Q& A – Duke/Sandy/Latha/Dan 20 minutes20 minutes
Enterprise Data WarehouseEnterprise Data Warehouse
Create an implementation roadmapCreate an implementation roadmap Content scope – January 1998 thru presentContent scope – January 1998 thru present All products loaded over 5 yearsAll products loaded over 5 years
Implement “value” after each iterationImplement “value” after each iteration Loss Cost, Retention, Loss TrianglesLoss Cost, Retention, Loss Triangles
Establish Data Stewardship - 2004Establish Data Stewardship - 2004
Enterprise Data WarehouseEnterprise Data WarehouseThe data warehouse will support:
Loss Cost Analysis
RetentionAnalysis
modernLINK Reporting
ProfitabilityAnalysis
DataWarehouse
UnderwritingAnalysis
ProductPricingAnalysis
FinancialAnalysis
Data Warehouse ValueData Warehouse Value
MH
Loss Cost
SB
Loss Cost
MC
Loss CostRetention
UVRC Pricing / GLMLoss
TrianglesmodernLINK
MH PIFmLINK
vs. Legacy
Retro
StudiesMapping
Renewal
Reporting FID MSB
CAT Analysis
Cancellation
Reporting
Address
Data
Agency
Profile
Analysis
Claims
Liability
Partner
Experience
Reporting
Data Warehouse StatisticsData Warehouse Statistics
1997 policies used to seed warehouse: ~700,000
Total policies Jan 1998 thru Jun 2007
Total units Jan 1998 thru Jun 2007
Average Number of Coverages per policy: 5
Average number of policies in-force per month: 800,000
Average number of claims per month: 8,000
Data Warehouse BenefitsData Warehouse Benefits Single version of the truthSingle version of the truth
Data integrated at the lowest levelData integrated at the lowest level
High-end hardware platformHigh-end hardware platform
Codes translated to “English” termsCodes translated to “English” terms
Resolve source system problemsResolve source system problems
Data quality review and correctionData quality review and correction
Integration of external informationIntegration of external information
Data Mart ThemesData Mart Themes
modernLINK quote modernLINK quote ExposureExposure RetentionRetention ExperienceExperience Loss CostLoss Cost ClaimsClaims UnderwritingUnderwriting
Technology Enablers….Technology Enablers…. IBM RS6000 AIX processorsIBM RS6000 AIX processors
EMC data storageEMC data storage
Oracle DBMSOracle DBMS
COGNOS for reporting utilizing query, report, COGNOS for reporting utilizing query, report, mapping and analytical tools mapping and analytical tools
Websphere PortalWebsphere Portal
LDAP for single sign-onLDAP for single sign-on
Showcase AgendaShowcase Agenda Background/Business Case Background/Business Case
20 minutes Sandy Wagner20 minutes Sandy Wagner
Data Warehouse – AIIM Data Warehouse – AIIM 20 minutes Latha Subramanian20 minutes Latha Subramanian
Data Model – AIIM Data Model – AIIM 20 minutes Duke Ganote20 minutes Duke Ganote
Information Management – AIIMInformation Management – AIIM20 minutes Dan Daly20 minutes Dan Daly
Q& A – Duke/Sandy/Latha/Dan Q& A – Duke/Sandy/Latha/Dan 20 minutes20 minutes
Data Model Data Model Provides a common, integrated way for the Provides a common, integrated way for the
corporation to view and to communicate corporation to view and to communicate about its businessabout its business
Allows the business to drive the systemAllows the business to drive the system
Creates standard definitions/documentationCreates standard definitions/documentation
Provides structure to new development Provides structure to new development projectsprojects
Enterprise Data ModelEnterprise Data ModelPeoplePeople PlacesPlaces ThingsThings
InsuredsInsureds
OperatorsOperators
LienholderLienholderss
ClaimantsClaimants
GeographyGeography
AddressAddressQuotes/PoliciesQuotes/Policies
ClaimsClaims
Coverages Coverages
Accidents/ViolationsAccidents/Violations
Homes/VehiclesHomes/Vehicles
UW rulesUW rules
Makes/ModelsMakes/Models
Jump Start Enterprise Data ModelJump Start Enterprise Data Model
Acord Standards
Generic Model based on Insurance Industry Practices
AMIG Enterprise Data Model
TransformAMIG Specific Requirements
Integrated View: Common Data DefinitionsAcross business
Manufactured HomeSite BuiltMotorcycleMotor HomeTravel TrailerClassic AutoFIDCommercial
Data Model BenefitsData Model Benefits Foundation for:Foundation for:
modernLINK rate & quote applicationsmodernLINK rate & quote applications Data warehouse/data mart/analytic designData warehouse/data mart/analytic design mLP3 Operational Data Store (ODS) mLP3 Operational Data Store (ODS)
designdesign New projects simply add to the modelNew projects simply add to the model
Insurance scoreInsurance score Claims liabilityClaims liability
Development of data standards and a Development of data standards and a common “language”common “language”
Inmon, InitiallyInmon, Initially
Data warehouse built using Inmon Data warehouse built using Inmon approach:approach:
Source (non-
relational)
Data Warehouse(normalized)
DataMart(star)
End of month
End of month
“Corporate Information Factory Components”, W. H. Inmon http://www.inmoncif.com/view/26
ConformanceConformance
Conformed Dimensions:Conformed Dimensions:
Data Warehouse(normalized)
Loss CostDataMart
(star)
Conformed Dimensions
PricingDataMart
(star)
RetentionMart(star)
“The 38 Subsystems of ETL”, Ralph Kimball http://www.intelligententerprise.com/showArticle.jhtml?articleID=54200319
ChallengesChallenges
Multiple sourcesMultiple sources LatencyLatency StewardshipStewardship
Multiple SourcesMultiple SourcesOPPORTUNITIESOPPORTUNITIES::
Daily claims/catastrophe feedsDaily claims/catastrophe feeds
3rd party Claim data (claims cost 3rd party Claim data (claims cost standards)standards)
Huon (an new Insurance ERP)Huon (an new Insurance ERP)
Munich RE (pending merger with Munich RE (pending merger with reinsurer)reinsurer)
Multiple SourcesMultiple Sources
RESPONSESRESPONSES:: Pull dataPull data: generally from relational : generally from relational
DBMS, e.g. DB2, Informix, SQL DBMS, e.g. DB2, Informix, SQL ServerServer
Push dataPush data: generally from non-: generally from non-relational DBMS: DMS II (Unisys)relational DBMS: DMS II (Unisys)
Latency ChangesLatency Changes
OPPORTUNITYOPPORTUNITY: Daily information: Daily information
Catastrophe reporting; e.g. Hurricane Catastrophe reporting; e.g. Hurricane Katrina 2005, “Fab Four” of 2004Katrina 2005, “Fab Four” of 2004
Financial Institutions requesting daily Financial Institutions requesting daily account information on insureds.account information on insureds.
Latency ChangesLatency Changes
Source (OLTP)
CATastropheDataMart
(star)StagingArea
daily daily
Daily Conformed Dimensions
dailydaily
RESPONSERESPONSE: Kimball architecture: Kimball architecture
“Kimball Design Tip #34: You Don’t Need an EDW”, Ralph Kimball http://www.kimballgroup.com/html/designtipsPDF/DesignTips2002/KimballDT34YouDontNeed.pdf
Latency ChangesLatency Changes
Kimball ArchitectureKimball Architecture““The staging area is exactly like the kitchen in a The staging area is exactly like the kitchen in a
restaurant. The kitchen is a busy, even dangerous, restaurant. The kitchen is a busy, even dangerous, place filled with sharp knives and hot liquids. The place filled with sharp knives and hot liquids. The cooks are busy, focused on the task of preparing the cooks are busy, focused on the task of preparing the food. It just isn't appropriate to allow diners into a food. It just isn't appropriate to allow diners into a professional kitchen or allow the cooks to be professional kitchen or allow the cooks to be distracted with the very separate issues of the fine distracted with the very separate issues of the fine dining experience. ”dining experience. ”
Two Powerful Ideas: foundations for modern data warehousing, Ralph Kimball Sept 17, 2002: http://www.intelligententerprise.com/020917/515warehouse1_1.jhtml
Data StewardshipData Stewardship
OPPORTUNITYOPPORTUNITY: : Daily instead of monthly reference data Daily instead of monthly reference data needed. However, for example, no needed. However, for example, no dailydaily system of recordsystem of record automated for: automated for:
Claims AdjustersClaims Adjusters
Catastrophe name/detailsCatastrophe name/details
Data StewardshipData Stewardship
RESPONSERESPONSE::
Data stewards maintain master data / Data stewards maintain master data / system of record.system of record.
Over night ETL uses master data for Over night ETL uses master data for building dimension.building dimension.
Referential integrity always enforced with Referential integrity always enforced with fact table, so data stewards cannot fact table, so data stewards cannot “delete” required for integrity.“delete” required for integrity.
Showcase AgendaShowcase Agenda Background/Business Case Background/Business Case
20 minutes Sandy Wagner20 minutes Sandy Wagner
Data Warehouse – AIIM Data Warehouse – AIIM 20 minutes Latha Subramanian20 minutes Latha Subramanian
Data Model – AIIM Data Model – AIIM 20 minutes Duke Ganote20 minutes Duke Ganote
Information Management – AIIMInformation Management – AIIM20 minutes Dan Daly20 minutes Dan Daly
Q& A – Duke/Sandy/Latha/Dan Q& A – Duke/Sandy/Latha/Dan 20 minutes20 minutes
Information Management BenefitsInformation Management Benefits
Single BI ArchitectureSingle BI Architecture Provides a consistent view of our Corporate DataProvides a consistent view of our Corporate Data Allows for common product training & supportAllows for common product training & support Volume license pricing provides flexibility and Volume license pricing provides flexibility and
cost savingscost savings
Converting Data Collectors to Information Converting Data Collectors to Information ConsumersConsumers
Corporate Portal IntegrationCorporate Portal Integration Delivering specific information to specific Delivering specific information to specific
business usersbusiness users Providing pre-emptive alerts to users based on Providing pre-emptive alerts to users based on
specific (data) eventsspecific (data) events
Single BI ArchitectureSingle BI Architecture (Consistent View, Common Training & Support & Volume Pricing)(Consistent View, Common Training & Support & Volume Pricing)
Using Cognos 8.2 for our Enterprise Using Cognos 8.2 for our Enterprise Reporting PortalReporting Portal Report Studio, Analysis Studio, Query Studio, Report Studio, Analysis Studio, Query Studio,
Event Studio, Metric StudioEvent Studio, Metric Studio
All Cognos Content Provided in ThemesAll Cognos Content Provided in Themes modernLINK quote modernLINK quote ExposureExposure RetentionRetention ExperienceExperience Loss CostLoss Cost ClaimsClaims UnderwritingUnderwriting
Single BI ArchitectureSingle BI Architecture (Consistent View, Common Training & Support & Volume Pricing)(Consistent View, Common Training & Support & Volume Pricing)
Converting Data Collectors Converting Data Collectors to Information Consumersto Information Consumers
Corporate Portal IntegrationCorporate Portal Integration
Converting Data Collectors Converting Data Collectors to Information Consumersto Information Consumers
Delivering specific content to specific usersDelivering specific content to specific users ‘‘Bursting’ Experience & Exposure information Bursting’ Experience & Exposure information
directly to our Business Partners (Agents)directly to our Business Partners (Agents)
Converting Data Collectors Converting Data Collectors to Information Consumersto Information Consumers Providing pre-emptive alerts to users Providing pre-emptive alerts to users
based on specific (data) eventsbased on specific (data) events
So What’s Next?So What’s Next?
Spend more time Spend more time executing strategy executing strategy & less time & less time gathering datagathering data
Manage to Manage to Corporate Corporate Scorecards / Scorecards / Performance Performance MetricsMetrics
Showcase AgendaShowcase Agenda Background/Business Case Background/Business Case
20 minutes Sandy Wagner20 minutes Sandy Wagner
Data Warehouse – AIIM Data Warehouse – AIIM 20 minutes Latha Subramanian20 minutes Latha Subramanian
Data Model – AIIM Data Model – AIIM 20 minutes Duke Ganote20 minutes Duke Ganote
Information Management – AIIMInformation Management – AIIM20 minutes Dan Daly20 minutes Dan Daly
Q& A – Duke/Sandy/Latha/Dan Q& A – Duke/Sandy/Latha/Dan 20 minutes20 minutes
Q & A sessionQ & A session
Wrap UpWrap Up Enterprise Data Warehouse now in its 7th year
Business units embrace the DW
Holistic view of information in one place
Next phase: deliver similar functionality to our external business partners
Our case study has been placed in National Archives
The copy of the case study can be found on the following web page: http://www.cwhonors.org/viewCaseStudy.asp?NominationID=54
SWOC DAMA 2008 Showcase atSWOC DAMA 2008 Showcase at American Modern Insurance American Modern Insurance
February 21, 2008