Insurance Analytics High Performance Underwriting Solutions June 2009 Copyright © 2009 Accenture. All rights reserved. Accenture, its logo, and High Performance Delivered are trademarks of Accenture. Compelling Opportunity: Fiduciary Necessity While information is the primary currency of the business of insurance, many insurers lack adequate insight to operate as high performers. • Conventional business intelligence has grown out of a history of compliance reporting requirements for regulatory bureaus and shareholders, and most MIS reflects this • The legacy system environment and history of mergers & acquisitions inhibits data quality and timely access • Claim coding is not aligned with premium/policy coding conventions • Inconsistent reporting standards globally forces complex reconciliation • Exposure measurement tools (e.g., PML models) are confined to single line of business and single peril thereby constraining insurers’ ability to develop integrated, enterprise views of accumulation • Operational technologies (e.g., task management, sales/submission management, call center) are not yet broadly adopted thereby limiting insights into service metrics, sales & submission outcomes, SLA compliance and individual service performance • Third party data (e.g., peer benchmarking, industry loss experience, customer demographics) is not readily integrated into the product development and actuarial processes to ensure profitable alignment with growth expectations Copyright © 2009 Accenture All Rights Reserved. 1 Strategic Importance of Business Intelligence for Insurance Timely data access and robust analytical capabilities provide measurable, immediate value to the organization. Loss Cost Improvement • Ability to more rapidly address pricing leakage issues or market leverage opportunities • Ability to optimize capacity allocation in catastrophe exposure territories • Improved insight into rating and pricing levers to optimize product filing strategies and underwriting practices • Better alignment between premium coding and claims coding to achieve more granular and precise profitability insight • Robust trending capabilities for improved loss development management Expense Savings • More efficient and consistent data preparation facilitates data availability & reporting • Dashboard, Reporting and What If Modeling tools empower business users • Reduced exposure to regulatory fines due to incomplete or inaccurate data Market share Growth • Improved product and geographic analytics enable more focused growth strategies • Real-time insight into leading indicators of book movement , market penetration and cross-sell/up sell to proactively manage growth drivers Improved Customer Service • Timely operational metrics to address service turnaround, SLAs and customer retention • More granular distribution segmentation to optimize profitable customer segment growth Copyright © 2009 Accenture All Rights Reserved. 2 Why Now? Key Industry Trend Implication for Insurance Analytics Softening P&C Market • Increasing pressure to preserve rating and pricing adequacy • Need to leverage micro-segmentation to find profitable market segments • Growth will be fueled by differentiated products and services which are dependent on unique insights Increased M&A Activity • Insurers must efficiently use vast amounts of disparate information from their acquired companies • Core systems replacement driven by post-merger rationalization is bringing new front-end systems that capture significantly more data about risk characteristics and claimant details Advanced Pricing Methods and Tools • Pricing methods such as “Predictive Modeling” are being widely used in Personal and Small Commercial Lines and are poised to be advanced through more sophisticated analytics • Most large insurers are migrating to proprietary pricing models over purchasing third party model services • External data is being increasingly leveraged to augment proprietary/internal data Focus on Improved Productivity • Increasing need for consistency and efficiency in the core operations of Underwriting and Claims, as highly skilled workforces age and retire, and exacerbated by M&A activity • Growth imperatives are underpinned by cost optimization initiatives, thereby demanding greater capacity within current resources Changing Regulatory Environment • AM Best rating now includes a category for exposure accumulation practices • Spitzer allegations have resulted in a push for increased transparency in broker / underwriter interaction around the globe • Sarbox regulations have implications for decision accountability and risk management practices Copyright © 2009 Accenture All Rights Reserved. 3 Most insurers have focused their analytics investments on reporting capabilities and point-based predictive modeling. Business Intelligence Maturity Curve What’s the best that can happen? Optimization Predictive Modeling Forecasting/extrapolation Competitive Statistical analysis Advantage What will happen next? Analytics What if these trends continue? Why is this happening? Alerts Query/drill down Ad hoc reports Standard reports What actions are needed? Where exactly is the problem? How many, how often, where? Access and Reporting What happened? Sophistication of Intelligence Copyright © 2009 Accenture All Rights Reserved. 4 Competing on Analytics Copyright © 2009 Accenture All Rights Reserved. 5 Accenture believes the most powerful business intelligence is generated through the integration of insights across the insurance value chain Insurance Value Chain Data Analytics Benefits Enhanced Enterprise and Core Capability Decision Making Product Development • Gain insights for creating more tailored products • Improve ability to deploy new products effectively (filing strategies, training, marketing) • Enhance primary and secondary research focus • Leverage product performance metrics to improve business rules and operational efficiency Marketing & Distribution • Identify potential sales more precisely by analyzing customer purchasing patterns • Enhance agency recruiting capabilities • Optimize distribution compensation models • Improve prediction of product preferences and customer retention Copyright © 2009 Accenture All Rights Reserved. Pricing & Underwriting • Speed risk selection and quoting with better alignment between risk quality and pricing (reduced leakage) • Provide timely insight into price tiering models (eg., predictive models) to adjust models for more targeted marketing • Identify non-core underwriting operational activities that are candidates for automation, elimination or delegation • Improve utilization of underwriting services (premium audit, loss control) • Provide more accurate, timely and detailed understanding of exposure accumulation Policy Processing • Increase ability to enhance servicing for the most valuable customers and distributors • Lower costs of processing with enhanced insight into performance of operational staffing models • Gain insight into process bottlenecks and detailed transactions to identify improvements in policy automation Claims • Improve timeliness and accuracy of fraud detection • Enhance resource allocation and prioritization with better insight and predictability into claim complexity and settlement potential • Improve subrogation predictions • Enhance reserving practices • Gain operational insight to increase levels of automation for simpler claims Performance Management • Improve consistency in metric management for the organization • Improve alignment between internal and external data quality and reporting standards • Achieve greater granularity and timeliness in reporting for improved insight and regulatory compliance • Increase user sophistication in performance analytics through improved data management and analytic technologies 6 For Underwriting, sophisticated analytics have become table stakes… Product Product Research • Market Territory Analysis • Market Segment Analysis • Competitor Analysis • Distributor Analysis Product Design/ Performance • Historical Product Analysis • Rate Making Analysis • Pricing Analysis • UW Rule Analysis • Loss Experience Analysis • Contract/ Forms Analysis Product Launch • ROI Analysis • Product Launch Analysis • Impact/ Disruption Analysis • What/If Analysis • Regulatory/ Trend Analysis • DOI Relationship Analysis Marketing • Market Penetration Analysis • Strategic Demographic Analysis • Competitor Analysis • Agency/ Distribution Analysis and Performance Measurement • Campaign Effectiveness (Affinity, Advertising, Lead Dissemination, etc.) • Account Rounding Analysis • Customer Value Analysis Sales • Agent Performance/ Profitability Analysis • NB & Quote Flow • Hit and Yield Analysis • Cancellation Analysis • Retention Analysis • in-Force Strategy • Agent/Distribution Performance Analysis • Channel/Access Method Analysis • Agency Management Analysis • Sales Tool Efficiency • Acquisition Cost Analysis • Compensation Analysis • Relationship Referral Analysis • New Business Sourcing Analysis • Cross Sell / Up Sell Analysis • Lead Management Copyright © 2009 Accenture All Rights Reserved. Pricing/ Actuarial • Pricing Performance including Tool Usage • Predictive Model inputs/outputs/final • Predictive Model deviations from Baseline • Predictive Model and UW Rule Integrated Analysis • Rate Adequacy Analysis • Marketplace Analysis • Profitability Analysis • Loss Ratio Analysis • Loss Development and Trending • Rate Development and Trending • Residual Market Loads (WC) • Involuntary Book Analysis • Off-Balance Analysis • Reserve Analysis • Expense Analysis Underwriting • UW Productivity • UW Expense Analysis • SLA Mgmt • Appetite Analysis • Segmentation Analysis • Book Mix Analysis • Hit and Yield Analysis • Referral Rates • Authority Analysis • Rule Analysis • Price/Credit Analysis • Agent Performance by Account/Book • UW Service Utilization • Market Comparison Analysis • Competitor Analysis • Quality/Audit Analysis • Exposure Management • CAT Management • Loss Experience Analysis • Reinsurance Analysis (Treaty, Fac) Servicing • Service Channel Analysis & Optimization • Service Channel Segmentation • Knowledge Management • Content Management • Workflow Analysis • SLA Mgmt • Contact Mgmt • Turnaround Times • Cycle Times • Straight-through processing volume • Escalation Analysis • Reassignment Analysis • Customer Complaint Analysis • Policy Error Analysis • Span of Control Analysis • Self-service Inquiry Analysis Claims • Claims Assignment and Routing • Fraud Detection • Formula Based Reserving • Reserve Analysis • Claims Handling Effectiveness • Claims Processing Efficiency • Subrogation Analysis Corporate Performance • Planning/ Budgeting • Growth Analysis • Book Mix and Portfolio Analysis • Profitability Analysis • Marketplace Analysis • Competitor Analysis • M&A Analysis • Performance Management • Resource Attrition • Investment Analysis • Expense Analysis • ISO and Bureau Reporting • DOI/Statutory Reporting • Taxes, Boards & Bureaus Analysis • Voluntary /Involuntary Market Analysis 7 Accenture Insurance Analytics Solutions: Predictive Modeling Accenture's initial solutions for Insurance Analytics are focused on Predictive Modeling to help clients leapfrog current approaches with impactful results in the areas of Distribution, Underwriting and Claims. Enhanced Enterprise and Core Capability Decision Making Product Development Marketing & Distribution Pricing & Underwriting Distribution Insight & Optimization Pricing Insight & Optimization Policy Processing Claims Performance Management Predictive Claims Processing Common Deployment Solution Copyright © 2009 Accenture All Rights Reserved. 8 “Generation 1” Predictive Models have enabled more precise product pricing at most insurers… “Now, we've actually got predictive modeling tools that allow us to improve the overall quality of our underwriting decisions and improve the consistency of what we're doing from underwriting our book.“ JAMES LEWIS, PRESIDENT AND CEO, CNA PROPERTY & CASUALTY OPERATIONS “Our models categorize the best risks as five diamonds and the worst as one diamond. Early indications for BOP are that loss ratios for five-diamond accounts are three times better than those of one-diamond accounts, and we're writing significantly less of the one-diamond business.“ DALE THATCHER, EVP, CFO, SELECTIVE INS. GROUP, INC. “It's a data-base industry now, particularly in small face-value kind of business, and so we are very focused on innovation around product, particularly things like predictive modeling.” FRED EPPINGER, CEO, THE HANOVER INSURANCE GROUP, INC “We started with automated underwriting platform or multivariate predictive models with commercial business three - four years ago. Today we have over 80% of our business on an automated underwriting platform. We think we're one of the strongest underwriters in using predictive models here.” MIKE HUGHES, EVP - INSURANCE OPERATIONS, SAFECO CORPORATION Copyright © 2009 Accenture All Rights Reserved. 9 …But additional capabilities are needed to advance Predictive Modeling along the Maturity Curve for increased and durable benefits WAVE I Corporate Mission and Objectives Off-Line Model Process Foundation • Scaleable technology • Management process • Quality controls • Transparency • Internal competency WAVE II WAVE III Basic Integration Robust Integration Process Mastery • Enterprise scope and scale • Process and resource stability and alignment • Payback on technology investment Process Optimization Proprietary models Enriched external data sourcing Robust solution integration Automated and integrated model and rule testing Automated model/rule deployment Externalized model/rule maintenance by business Market cycle tolerance • • • Management and leadership maturity • • • Quality maturity • Create “Speed to Market’ (reduce analytics cycle time) • Create reusable process • Build credentialed and technology resource pool across the enterprise • • • • Expand Model scope and scale • Add LOB’s under model • Establish Internal competence • Grow % of premium/ claims under model • Establish scaleable • Complete reference models • Invest in Modeling model deployment Pro forma architecture technology and resources Metrics Re baseline Metrics Baseline NEAR TERM Copyright © 2009 Accenture All Rights Reserved. MEDIUM TERM LONG TERM 10 Generation 1 Predictive Model Challenges • Use of external model vs. proprietary development (dependency on industry results, lack of model transparency) • LOB vs. Account/Customer model orientation • Lack of model granularity: location, agent/broker, • Lack of model integration with other rules (underwriting, rating, forms) • Lack of integrated book of business testing facility and “what if” scenario model impact analysis tools • Lack of externalized model rule management capabilities in the hands of the business • Lack of automated deployment/release management • Limited performance insight across models and other rules Copyright © 2009 Accenture All Rights Reserved. 11 Accenture’s UW Predictive Model Solution enhances business benefits through effective lifecycle integration Accenture Underwriting Modeling Solutions Source Data 3rd Party Integration Services Data Services Claims ETL Data Cleansing Data Matching Data Profiling Data Quality Data Integration Data Augmentation Policy Sample Data Sets DW Product Palette Testing Interfaces Production Impact Analysis Trial Rates Rating Engine Displacement Trial Rules Rules Engine Rules Impact Trial Models Scoring Engine Tier/Score Impact Copyright © 2009 Accenture All Rights Reserved. Rules Application Rating Application Production Rates Premium Production Rules Production Models Monitored Data Development Applications Model Application (Transaction Systems) Score / Tier Mods / Factors Rules Fired 12 This end-to-end UW Predictive Model Solution brings additional cost and growth benefits not achieved with Generation 1 solutions. Benefit Drivers: • High percentage of underwriting decisions untouched (manual underwriting on an exception basis only) • Reduction of data acquisition costs • Reduction of administration and selling expenses on a per policy basis • UW & Pricing changes quickly brought to market • Timely cross selling of products • Positively influence channel production by improving the carrier/channel interface (easy to work Expense Reduction (5-15%) Premium Growth (varies by strategy) with) • Proactive targeting for new business • Targeted renewal pricing and improved retention • Consistent, error-free selection of individual risks • Alignment of risk selection and existing book of business with overall risk profile objective • Proactive management of non-renewals • Risk selection improved through the identification and qualification of relevant predictors of loss Loss Ratio Reduction (4-5 pts) Offline Model Basic Integration Copyright © 2009 Accenture All Rights Reserved. Robust Integration * Benefits are further enhanced as predictive modeling is integrated into other business functions 13 Accenture’s solution enables insurers to industrialize their predictive modeling capabilities throughout the organization. Model Development Objectives: – – Identify data attributes that are predictive of the outcome being measured (e.g., Loss) Derive the most reliable algorithm to calculate a predictive score, based on extensive analysis of historical data Model Integration Objectives: – – – – Copyright © 2009 Accenture All Rights Reserved. Integrate predictive scores into business processing via rules engine configuration and portal/UW/Policy system integration Develop business rules to provide further guidance & explanation of the scores Test impact of models on book of business and with other rules Deploy models/rules into run-time environment Model Refinement & Extension Objectives: – – – Establish timely operational data extracts for real-time performance insights Establish flexible platform that externalizes model/rule maintenance to rapidly change rules, explanations, and predictive scores Extend predictive score information into broader areas (servicing, sales) 14 Accenture Enables Speed to Integrate New Models Results with Accenture Common Challenges • Flexible Scoring Engine requires minimal or no new development to adapt to new model structures • 2 – 6 months lead time to integrate new model structures • Scoring Engine not flexible enough to adapt to new model structures • 9 – 18 months lead time to integrate new model structures ??? GO Accenture’s Differentiated Answers Rules-based scoring engine Robust model deployment architecture Configurable Scoring Process Architecture Accenture Assets: New Predictive Model Scoring Specifications Scoring Rules Management Rules Exceptions? Author Migrate Test Analyze Scoring Engine Wrapper (SOA) Rules Engine (Blaze,iLOG) Yes Process Rules & Process Copyright © 2009 Accenture All Rights Reserved. Author Migrate Test Analyze • Onshore & Offshore Blaze Advisor & iLOG Scoring & Rules Engine Implementation Skills • Blaze & iLOG Scoring Engine Software Accelerators Scoring Process Configuration No • Solution Approach Options & Reference Architectures Automated Model Deployment Utilities • Model Deployment Utilities • SOA Scoring Engine Wrapper Software Accelerators 15 Accenture Enables Efficient Use of 3rd Party Data Results with Accenture Common Challenges • 3rd Party Data not cleanly matched to internal data • Data refresh, scrubbing, matching, storage, and utilization options difficult to navigate ??? • Cohesive data integration & utilization • Automated matching with configurable accuracy/exception thresholds • Thoughtful, guided decisions on all key points of consideration GO Accenture’s Differentiated Answers Integrated (real-time and batch) 3rd Party Data T Thresholds 3rd Party Extensions 3rd Party Data Integration Data Utilization 3rd Party Overlay 3rd Party Extensions 3rd Party Overlay Internal Data Internal Data Cohesive data utilization Accenture Assets: Customer Policy Sophisticated data matching T Data Matching T Data Matching T Data Matching Data Matching T Data Matching T Data Matching Copyright © 2009 Accenture All Rights Reserved. Real-Time Interface Real-Time Interface Real-Time Interface 3rd Party Data Source A Data 3rd Party Source A Data 3rd Party Source A Batch Interface Batch Interface Batch Interface 3rd Party Source F • Data integration / utilization options & best practices 3rd Party Source E • Proprietary Data Matching Software Accelerators 3rd Party Source D • Onshore & Offshore 3rd Party Data integration Skills 3rd Party Source C • Third Party Data Service Provider Alliances 3rd Party Source B 3rd Party Source A 16 Accenture Enables Effectiveness of Business Rules Common Challenges • • • • Results with Accenture • Clear, timely & consistent rule requirements • Flexible & effective rules architecture established quickly • Rules consistently managed to the highest quality throughout their lifecycle Rule requirements difficult to crystallize Rules architecture not well defined Rules quality/completeness not validated Rating & UW rules not well aligned to leverage Predictive Model Scoring ??? GO Accenture’s Differentiated Answers Proven rule requirements process & templates Rule Properties Explanation Text Rule Name Rule Description Rule Type Rule Type Conditions Needed Line Of Business is <select a LOB> Program is <select a Program> State <select a State> Market Indicator <select a Market Indicator> Effective Date is between <start date> and <end date> Transaction Type is <select a Tranasaction Type> Description/Definition Business Area Identification ID Source this rule is in scope Date All input record attributes (no scoping conditions) difference in days between <start date> and <end date> <operator> <value> difference in months between <start date> and <end date> <operator> <value> Derivation difference in years between <start date> and <end date> <operator> <value> difference in days between <start date> and today <operator> <value> The rules that govern the process of updating and maintaining the difference in months between <start date> and today <operator> <value> agent hierarchies. difference in years between <start date> and today <operator> <value> The rules that govern the processing of background checks for producers such as when to order, and, ultimately, thisordering rule isfequency in scope approval or denial. All input record attributes (no scoping conditions) The rules that are used to facilitate effective management of agent education requirements. count of <derivations> within <timeframe> days <operator> <value> The rules that govern the process of developing a contract between count of <derivations> within <timeframe> months <operator> <value> internal employees and agents, as well as the rules for managing of <derivations> within <timeframe> years <operator> <value> agent/producer licenses (new, count maintenance, verification). This includes interaction with state bureaus and<derivations> any subsequent <operator> <value> count of Knock-Out obligations. In addition, Contracting includesin thedays contracted base <start date> and <end date> <operator> <value> difference between commissions. difference months between <start date> and <end date> <operator> <value> The rules that are used to determine the validinproducer code and sub-codes that are used by thedifference various systems. This isbetween specific to <start date> and <end date> <operator> <value> in years setting up a new producer in the system. difference in days between <start date> and today <operator> <value> The rules that govern the types of risks and products that can be difference in months <start date> and today <operator> <value> written for a particular agent. This can be regulated by the between type of producer, state, type of risk, authority level, etc. This includes both <start date> and today <operator> <value> difference in years between Scoping General Info Category Rule Taxonomy Sub-Category Agency/Producer Management Agency Organization Agency/Producer Management Agency/Producer Management Background Checks Education Management Agency/Producer Management Licensing and Contracting Rules Agency/Producer Management Producer Code and Sub-Code Rules Agency/Producer Management Agency/Producer Management Alliance Relationship Management Billing & Collections Billing & Collections Billing & Collections Billing & Collections Billing & Collections Billing & Collections Claims Claims Rule Architecture Requirements contractual and license eligibility. This also pertains to subProducer Eligibility producers. Writing Company The rules that are applied to determine the ultimate writing company Appointment to which an agent will be assigned. Ability to manage relationships with third party providers of products and services (e.g. TPA's, Affinity Groups, Specialty Products Carriers) Rules that determine the handling for the collection of bad debts (inBad Debt Collections house vs. 3rd Party). Billing and Cancellation Timeframe and The rules that govern the timeframes of notifications of policy bills Notice Rules and pending cancellations. Billing Options Rules that determine the valid billing options (monthly, quarterly, Eligibility annually). Billing Schedule The rules that govern the amounts to be billed based upon effective Rules date and current date of the policy. Rules that regulate the billing type options that are available for a Billing Type Eligibility risk (e.g. direct bill or agency bill or 3rd party bill) Payment Method The rules that determine the valid methods for payment (e.g. EFT, Eligibility Credit Card, Cash, Check) The rules that regulate the process of capturing first notice of loss Claim Notification information. The rules that determine the routing of a claim to a particular system Claim Routing and assignment to a team. Copyright © 2009 Accenture All Rights Reserved. Rule Flow Model Rule Methods Condition In Condition Drop-Down (Y/N) Frozen (Y/N) Y N Y N Y N Y N Y N Y N Sample N Y Y N Y N Y N Y N Y N Y N Y N N Y Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N BR Education & Change Mgt. Actions Needed <select a rule> Rule Constructs Meta-Data Approach Applicability Approach Business Rule Rule Methods Versioning OrganizationMeta-Data Meta-Data Approach Approach Approach BR Technology Director BRCC Partnership set result code to DELINE Director set result explanation to <enter explanation> Product Mgt. BR Standards/ Operations BR Management Marketing Reference Data Approach Rule Decisioning Technology Technology Rule Deployment Project Management BTI SME(s) to guide key BR Competency Partnership Technology decisions & Leader content: - Rule tech. solution design - RuleBRCC tech. vendor selection Partnership - Rule perf. optimization Director - Rules tech. stewardship BRCoE will provide a point of Rule Management Software Workstream IT Account Exec Rule Development Workstream Claims Etc. Key interaction Rule Migration Integrated Rule Testing Rule Impact Analysis Migration Grouping Rule Query Test Data Maintenance Migration Scheduling Rule Status Mgmt Test Population Maintenance Deployment Triggering Rule Locking Test Case Maintenance Enterprise Repository Rule Archiving Test Case Execution Concurrent Editing Test Results Analysis Rule History & Audit Rule Test Environment Software Testing QA Rulebase SW Test Environments SW Testing Best Practices Type-Specific Repositories Rule Management Portal Management Information Production Rule Results Analysis Rule Repository Views & Analysis Research Rule Results Analysis Data Warehouse Integration Rule Harvesting BR Product & UW SMEs Rule Librarian Accenture Assets: • Innovative rules architecture point of view • Industry leading rules management framework • Rule requirements templates & best practices • Onshore rule requirements gathering skills Rule Harvesting Rule Requirements Business Areas Business Growth Business Vision Source & topic of interaction Rule Domains Rule Testing Rule Unit Testing Rule Copy Technology Object Model Director Rule Execution The coordination across rule domains & business Change areas to: Harvesting BR Technology Software Workstream • Develop & allocate the staffing to implementArchitect & manage businessDependencies rules Methodology Rules Solution • Build awareness & expertise to guide business rules capability development & utilization Architecture Application Infrastructure • Facilitate collaboration in key business rules activities & decisions Management Management • Promote consistency in business rules best practices BR Architect Rule Rules Externalization Architecture BR BR Opportunities Manages the allocation of pool Manages the Coordinates BR Training / change Methods and tool BR BR management resources to augment projectBR Education allocationBR of Standards/ pool infrastructureBR runtime specialist to manage: Application Rule Architecture Infrastructure IT (SOA) & Management & Change Mgt. Rule Project specialist to manage: staffing: resources Operations to operations, Management Rule Technology Management - Methods, tools, Methodology Architecture augment project configuration, and Rule - Business rules analysts Rule - Skill assessments and knowledge staffing: deployment Requirements Results (rule requirements, design management Business Rules - BR education and Rule Taxonomy & support, development, - Rules software (upgrades) Competency Center training Domain Mapping - Performance • Rule Stewardship testing, migration, and and integration Metrics/ Rpts. • Rule Standards - Mentoring and maintenance) developers Business • Rule Policies OTJ training - Process Business Rules Mgmt. Teams • Rule Education Value - Verification & validation of improvement - Team building specialtiesBusiness Areas / Product Work Rule Externalization • Rule Results Ownership Config. Mgmt. Business - Administrative Commercial Markets - BR career path Opportunities Decision Sponsors Underwriting Product Mgt. Marketing Etc. - WorkshopInitiative facilitators Support support - Skill database • Strategic Planning Small Business • Prioritization • Investment Opportunities Rule Authoring Repository Architecture Rule Operating Model Y Y BR Technology Architect Other Organization Organization Approach Approach Rule Execution Rule Outcome Systems Comment RuleSystem(s) Consuming Rule Service Rule Service System(s) Technology Security Integration Partitioning Performance BR Architect SME(s) to guide key BR Governance Investment decisions & content: Confirmation Business IT - Rule logic & approach design Architecture Council - Rule capability requirements Claims - Rule taxonomy & domain mapping Underwriting - Rules stewardship Applicability Approach Rule Management System(s) IT Account Exec Partnership Rule Management Suite Business Versioning Technical Architecture Systems Reference Y Data ApproachN Sample Outcomes record this derivation set date of occurrence to <date> BTI Competency Leader Rule Constructs Constructs Action Frozen (Y/N) N Rule Flow Model Etc. Actions Document Generation … Decision Support State Data Extension & Validation Transaction Work Management Coverage Rule Architecture Rule Mgmt. Conditions Rule ID Product Rate Determination Rules Product Configuration Rule Specification Comprehensive rules management framework Robust rules architecture Business Profitability Commercial Specialty Personal Business Areas Claims, Product Management, Underwrite Risk © 2007 Accenture. All rights reserved. 17 Accenture Enables Adoption of Business Change Common Challenges Results with Accenture • Business process & organization changes not aligned with either the model design or integration, and are difficult to agree upon • Timely and appropriate communications & training are lacking • Business changes clearly aligned with both model design and integration via structured collaboration throughout implementation • Leadership & key stakeholders are informed & aligned from early on ??? GO Accenture’s Differentiated Answers Communication Plan Broad communications plan developed and vetted early with key stakeholders Plan & Analyze Actuarial Product Mgrs. Field UW Home Office Accenture IT, etc. Design Process Design Org. Design Training & Performance Mgmt. Copyright © 2009 Accenture All Rights Reserved. Process & organization changes, training and performance management developed collaboratively to align with industry best practices and unique solution aspects Build Test Deploy Accenture Assets: Structured Collaboration • Industry leading Training & Performance Management Practice Industry Best Practices • Process & organizational change reference models Unique Solution Considerations • Communication plan reference models 18 Accenture has a track record of successful solution implementation with predictive modeling Keys to Effective Predictive Modeling Solutions Clear Vision Thoughtful Design Coordinated Delivery Increased Process Automation Flexible, RulesBased Solution Clearly Defined RollOut Strategy Selective Integration Points Scalable SOA Framework Cross-Organization Dependency Mgmt. Frequency and Severity-Based Models Integrated Internal & 3rd Party Data Cross-Organization Training Analytics-Directed Appetite Robust Management Information Leadership Backing & Communications Plan Proactive Underwriting Process Changes Model Change BetaTesting Platform Calculated Offshore Delivery Model Coordinated Business Rules Management Model Development Platform Parallel Development & Testing Tracks Copyright © 2009 Accenture All Rights Reserved. 19 Getting Started: Proof of Value Accenture would work with you to conduct a short project to determine solution approach options and define potential scope and business value for the integrated solution. Accenture Responsibility • Proof of Value Business Case: Develop the business case, capability scope, and high level business and technical requirements – Technical Architecture Assessment – Data Services Survey – Vendor Services analysis • Conceptual Design – pro forma design documents to integrate current or updated predictive models with production systems and testing/analytics solutions Copyright © 2009 Accenture All Rights Reserved. Client Responsibility • Part time support (25% - 50%) from cross-functional team of IT, selected business/functional areas, analytics group and stakeholders • Weekly status meetings with key stakeholders to evaluate deliverables and provide direction. • Provide current and planned application architecture documentation • Provide current and planned data services and model deployment architecture 20 Proof of Value Approach Weeks 1-2 Base Case • Evaluate current predictive models, Model Development modeling capabilities and model update cycle Integrated Architecture and Data Organizational Readiness • Define business case for Proof of Value • Define conceptual Design for Integrated Solution Weeks 3-5+ • NA • Develop Proof of Value – basic functionality for integrating rates, rules, and predictive models in a test environment • Develop Proof of Value – basic functionality for better integration of predictive models with rules engine in a test environment • Assess Analytics Management Processes • NA and Resources Copyright © 2009 Accenture All Rights Reserved. 21 Sample Accenture Credentials • Large National Insurance Company • • Integrated financial data across disparate acquired companies to reduce financial close, consolidation and management reporting times as well as eliminate manual processes. Defined a global management reporting process and developed the EDW solution Developing Center of Excellence for Analytics & Modeling • • • Specialty Lines predictive model configuration and integration to fuel lower touch growth model Integrated with policy system and FairIsaac/Blaze rules engine (including other UW rules) Provides guidance messages • Commercial lines Business Intelligence solution, including data warehouse, product/pricing analytics (SAP/BO) for BOP, Auto, WC, predictive model analytics, operational analytics Enterprise Finance Data Mart Developed forecasting/planning desktop (MSFT BI) • • • • • • • Enterprise Data Services including Master Data Management strategy/implementation • • Data Warehouse implementation including ETL, data prep, and analytical tools Product Development solution leveraging best of breed components, including Blaze Advisor for underwriting and pricing rules and Skywire Insbridge for rating Data Warehouse/Analytics Enablement of STP for Commercial Lines Installation of Accenture Underwriting and Claims assets Implementation of new Policy system for Personal Lines Copyright © 2009 Accenture All Rights Reserved. 22 APPENDIX Copyright © 2009 Accenture All Rights Reserved. 23 Insurance Analytics Capability Blueprint Users UW / Product Claims Producer / Cust. / Vendor Actuary Sales / Service Executive Data Access Drill Down Standardized Export (Excel / Access) Real Time Batch Ad-hoc Side-by-Side User Tailored GIS / Mapping Authority / Security Enabling Capabilities Simulation Testing Report Generation Scheduling Alerts / Notifications KPI’s Trending Common Data Definitions Forecasting Analytics Data Enrichment (3rd Party) Executive Info UW / Prod / Claims Decision Support Sales / Mkt Decision Support Actuarial Info Management CAT / Exposure GIS Modeling Information Refinement Data Layer Data Marts Data Warehouse Interfacing Source Systems Claims Financials Copyright © 2009 Accenture All Rights Reserved. Legacy Policy External / 3rd Party Distribution Mgmt 24 Insurance Analytics Technology Blueprint Our solutions are supported by a target technology architecture that enables speed of predictive solution implementation while positioning the insurer with a strategic infrastructure for enterprise analytics. Staging Sources Data Warehouse Operational Data Store Reporting & Analytics Analytical Data Sources Marketing Analytical Data Model Repository External Data Operational Data Store (ODS) ETL Core Transaction Applications/Data Data Dashboards Subject Area Data Marts & Cubes Producer Underwriting Unstructured information landing area Claims Data Warehouse Master Data Repository (Producer, Product, Policy, Customer, Location) Reporting Operations Score/ Rules Engine Underwriters Data Miner Model Mgt & Applications Business Claims Mgt Product Managers Dimensional Views Policy Data Archive Match Merge Data Quality Business Metadata Management Copyright © 2009 Accenture All Rights Reserved. Actuaries ETL Repository Extract, Transform, Load (ETL) Standardize Data Miners ETL Proprietary & Public Monitor Audit Technical Balance Control Metadata Dictionary & Repository 25 Accenture’s Insurance Analytics Offering: Pre-Built Architecture Assets PHASES/ WORKSTREAMS Model Development Integrated Architecture and Data Organizational Readiness Plan Analyze Design Build Test Deploy Maintain • Model Design Specification – reusable pro forma model designs based on known problem statements (e.g. fraud) • Meta Data Dictionary – A reference data specification specific to a model type (e.g. subrogation) that identifies the individual data attributes to be used in Sample Set development • ‘Raw Data’ ETL Procedures to populate the model development sample set from the transaction data • ‘Raw Data’ data pre-processing routines to populate the model development sample set • Derived Predictors – highly predictive features created from raw data inputs, applicable to all sample data sets for a common problem statement • Insight Visualizations – typical visualizations for unsupervised/exploratory models that provide insights into unexpected patterns and data clusters • Data Quality Reporting Templates – reusable reporting and metrics for data quality relative to usable features for modeling • Validation Reporting Templates – reusable report template for validating predictive models (validation approach) • ‘Raw Data’ ETL Procedures to execute the model algorithm at prediction time with live transaction systems data • ‘Raw Data’ data pre-processing routines to execute the model algorithm at prediction time with live transaction systems data • Data Services (interfaces) – ACS interfaces to third party data (run-time model execution) • Pre-loaded 3rd Party Data Sources – preloaded tables for external data sources (free public data relevant to Model Design Specifications) • Pre-designed 3rd Party Data Interfaces – common external data interfaces to proprietary data that is typically used to improve the prediction accuracy of models • Advanced Model Maintenance Architecture – interface and functionality to dynamically improve models from the transaction data flow • Report Designs – business relevant report designs typical for management user needs • User Interface Designs – sample user interfaces presenting a likely deployment of model output to the user in the transaction system • Rule Designs – typical business rules to interpret and act on prediction outcomes for a model type • Deployment Strategies, • Change Management Approaches Copyright © 2009 Accenture All Rights Reserved. 26 Accenture’s Insurance Analytics Offering: High Value Capabilities and Accelerators Our capabilities and accelerators can reduce the cost and risk of implementing predictive models and analytics across the enterprise for underwriting, claims and distribution operations. • Methodology and Tools to accelerate the design, development and deployment of the analytical models • Experienced Insurance Analytics and Actuarial services to develop, extend, and enhance the new analytical models • Market leading analytics software suite to develop and maintain analytical models • Technology expertise to install, configure, and integrate modeling software into the insurer’s current application environment • Integration architecture to integrate Accenture’s Insurance assets (ICC, CCS, UWC) with the analytics software suite • Robust metadata dictionary based on logical and physical insurance data models to rapidly define required data • Data services to automate the connection to 3rd party data • Business solutions that address today’s business problems and also representative of predictive modeling capabilities that can be applied broadly across the enterprise Copyright © 2009 Accenture All Rights Reserved. 27 Integrating analytics into the enterprise requires clear organizational accountability and responsibility. Centralized/Enterprise Services ILLUSTRATIVE Enterprise Analytics Scope • • • • • Methods/Standards Strategy Requirements Build Deploy Market and Competitor Intelligence Customer Insight & Experience Exposure Accumulation Intelligence Distribution Insight Process Improvement Analytics Analytics Technology and Data Governance Financial Support Line of Business/ Functional Representation and Demand Management Personal Insurance Analytics Decentralized/Line of Business Focused Business Insurance Analytics Corporate Analytics Scope: • • • • Strategy Requirements Prioritization Assigning Resources Lines of Business / Functional Leadership Scope • Line of Business and Functional Reporting and Analysis • Operationalize analytics Competency Centers (Others TBD) Product Modeling Copyright © 2009 Accenture All Rights Reserved. Claims and Service Analytics UW and Pricing Functional Analytics (HR, Mrktg, Fin) 28 Data – Current and New Data Type Claims Client % of Accounts Buying Risk Services Average Claim Amount by Loss Type Average Property Rate by Client Size Client Loss Ratio Co-Brokerage Allocation Coverages Purchased by Client Size Annual Sales Insurer Services Compensation Type Form Type Head Count History LOBs brokered by Insurer # Locations Operational Detail SIC TIV Acquisitions Average Time to Payment Carrier Services Growth/Expansion Plans Quota Share/Excess Copyright © 2009 Accenture All Rights Reserved. Data Source Insurer Insurer Insurer Insurer Insurer Insurer 3rd Party Insurer Insurer Insurer 3rd Party Insurer Insurer 3rd Party Carrier 3rd Party Carrier 3rd Party Insurer Carrier 3rd Party Insurer Current New Current New 29 Data -- Current and New Data Type Carrier Location Key Individual Loss Control Services Offered Policy Error Rate Policy Issue SLA Product Profile Quoted to Bound Rating, Pricing Detail Submitted to Quoted % Sprinklered Construction Type Earthquake Zone Flood Zone Location (Lat/Long) Occupancy Type Alarm Coverage Alarm Types Distance to Fault Distance to Next Aon Client Location Distance to Water EQ PML Flood PML ISO Town Class Slope, Aspect, Gradient Total Lifetime Losses Copyright © 2009 Accenture All Rights Reserved. Data Source Carrier Carrier Insurer Insurer Insurer Insurer Carrier Insurer 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party Client Client 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party 3rd Party Current New Current New 30 Integration and modeling of this enriched data will enable enhanced statistical analysis and forecasting. Location Carrier Client Claims Benchmark/Trend Copyright © 2009 Accenture All Rights Reserved. % of Accounts Buying Risk Services Average Claim Amount by Loss Type Average Property Rate by Client Size Client Loss Ratio Co-Brokerage Allocation Coverages Purchased by Client Size Deductibles by Coverage Fee-vs-Bundled Service by Service Type by Client Size Limits Purchased LOBs Purchased by Client Size Total Lifetime Revenue per Account % of Locations Buying NFIP Coverage Commission vs. Fee by Client Type by Geography Location Propensity for CAT Claims by Client Type Account Revenue per FTE Average Cost of Acquisition Average Cost to Remarket Average Cost to Renew Cancellations/Reinstatement Trends Cost to Convert Cost to Service Frequency and Severity $$ by Loss Type by SIC FTE Productivity (Leverage Ratio) Lead Generation Metrics Net Price by Product by Client Size New Business/Renewal Hit Ratios by SIC by Carrier Payment Trends Premium Change (exposure/rate/price) Prospecting/Sales Metrics Prospects to Accounts by Carrier by Office Service Effectiveness # & Type Transactions per Account by Size Source Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Insurer Current New 31 More predictive and optimized analytics will position the insurer to create “game changing” innovations to grow its business. • New Product & Service Offerings • Enhanced Client Performance • Diversified Revenues • Increased Market Leverage • Optimized Delivery Platforms Copyright © 2009 Accenture All Rights Reserved. 32 Big Ideas Differentiated Program Offering • Target SICs, geographies, size • Tailored coverage forms • Proprietary rating/pricing methodology • Streamlined application process • Affinity/sponsor organization exclusive membership access • Special risk engineering services • Business Income consultative services (fee-based, sourced through accounting firm network) • Preferred market arrangements with carriers, including pre-negotiated servicing terms • Proprietary risk scoring tool to gauge loss propensity and risk management leakage • Branded marketing material and advertising Copyright © 2009 Accenture All Rights Reserved. 33 Big Ideas Client Dynamic Pricing • Proprietary client account scoring methodology --- lifetime value, total cost of risk, predictive risk model based on economic conditions as well as loss propensity • Real-time benchmarking during proposal/quotation process • Optimize profit margin based on trend in quote take-up by risk class, geography, product line and client profile Copyright © 2009 Accenture All Rights Reserved. 34 Big Ideas Location Predictive Model • Develop proprietary location data warehouse • Aggregate Aon client location data with U.S. business census and location/building data from third parties • Develop proprietary location scoring methodology • Develop proprietary EML/PML scores by LOB, SIC, geography • Provide fee-based service for multi-product, multiperil location modeling • Develop new risk management, catastrophe management and business expansion consultative services • Partner with third party providers for “hot site” and “cold site” contingent arrangements for clients Copyright © 2009 Accenture All Rights Reserved. 35