CA数据仓库

nullData Warehousing In The Mobile Telecommunication IndustriesData Warehousing In The Mobile Telecommunication IndustriesTerry Yeo Practice Director Wong Bak Wei Solution ArchitectWhat is Data WarehousingWhat is Data WarehousingBasic SituationBasic SituationBusinesses need more information faster to remain competitive. Businesses have lots of data, but little information. Technology is catching up to the demand.Business DriversBusiness DriversIncreased Competition Faster Business Cycles Mergers & DownsizingNeed Information!!! Advanced Analysis Capabilities!!!Data vs. InformationData vs. InformationData is there but… Fragmented Not Integrated Poor Performance Low IT Priority Only DataResulting in...Resulting in...People are still making decisions Information or no information... Single version of the truth or no single version of the truth...Technology EnablementTechnology EnablementFaster Computers More Sophisticated Tools Cheaper Cost of Data StorageSo What is a Data Warehouse ?So What is a Data Warehouse ?“a process that transforms data into information so that the full business value of this data may be realized…”Data Warehouse FunctionsData Warehouse Functions Architecture ManagementData Mart ManagementPutting It Into Perspective...Putting It Into Perspective...85% of Fortune 500 Fastest Growing Segment of Commercial IT Spending $16.8B in 1996; 19.1%Annual Growth IDC Survey: Average ROI = 401%! …but 1 in 2 are FailingWhy are People Failing?Why are People Failing?Not Business-Driven Not Partnering With the Business Wrong Expectations Set Up Front Taking Short-Sighted Approaches “Build It and They Will Come”Data Mart Only WarehousesData Mart Only WarehousesIndependent Data Marts… Seems Faster, Cheaper No common architecture; No shared reference data Once built, difficult to integrate Vs. Dependent and/or Architected Data MartsData Warehousing Market ThemesData Warehousing Market ThemesAvoiding Failure 1 in 2 are failing High Risk Return on Investment Measuring Business Value Data Warehouses vs. Data Marts Data Warehouse Management Evolve with, not after, the business does. Support growing number of users, in more places, with more data! The Critical Success Factor in Implementing a Data Warehouse in your OrganizationThe Critical Success Factor in Implementing a Data Warehouse in your OrganizationSession AgendaSession AgendaData Warehousing today 8 Reasons of Failures to Avoid !!! 10 Critical Success Factors How to measure success SummaryData Warehousing TodayData Warehousing TodayMany flavors of DW has been implemented (in Asia as well) Enterprise end-to-end DW Metadata Management Subject Area Data Marts Reporting OLAP EIS ERP users implementing DW More emphasis of Web-enablement However …. 1 of 2 DW projects fails !!!!8 Reasons of Failure8 Reasons of FailureSuccess is hard to measure…Failure is easy !!! Reasons why DW projects fail No more funding Bad data quality Users unhappy with query tools Only a small percentage of users use the DW Poor performance Inability to expand Data is not integrated ETL process does not fit batch window Critical Success FactorsCritical Success FactorsCommon Data Definitions Consolidate different sets of departmental definitions (Extremely difficult ….) These definitions are rarely documented !!! Each project should have a glossary of business terms to support the projectCritical Success FactorsCritical Success FactorsWell-defined transformation rules Data from source systems will be transformed in one way or another. Data will always be specifically selected, recorded, summarized and integrated with other data Transformation rules are critical to do this correctlyCritical Success FactorsCritical Success FactorsProperly trained users Regardless of how easy to use a tool is, users must be trained Training should be geared to the level of user and the way they use the data warehouse Types of training how to use the tools how to use any custom developed applications availability of predefined queries & reports the data itself, and data structures for more powerful users Critical Success FactorsCritical Success FactorsExpectations communicated to users Performance. Availability of the data warehouse. Functions and what data is accessible, what predefined queries and reports are available, the level of detail data and how data is integrated and aggregated. The expectations of simplicity and ease-of-use. The expectations of accuracy in both data cleanliness and what the data means. Timeliness of when data will be available, and frequency of refreshing the data. Schedule expectations (system delivery). Where support comes from.Critical Success FactorsCritical Success FactorsEnsured user involvement Solicit requirements input from users. Have the users involved throughout the project. Best scenario - Business and Technical users Critical Success FactorsCritical Success FactorsThe project has a good sponsor The best sponsor is from the business side, not IT. Should be willing to provide ample budget. Should be able to get resources needed for the project. Should be accepting problems as they occur, and not use them as an excuse to kill the project. Should be in serious need of DW capabilities to solve a problem, or gain some advantage.Critical Success FactorsCritical Success FactorsThe team has the right skill set Resources with the right skill sets should be dedicated to the team. Critical roles should report directly to the PMCritical Success FactorsCritical Success FactorsThe schedule is realistic Unrealistic schedule - most common cause of failure. Project schedules should be imposed with the concurrence of the PM and team members. Schedules must include task and effort required.Critical Success FactorsCritical Success FactorsProper project control procedures (change control) The scope will always change. Changes in the project must be managed and controlled. Critical Success FactorsCritical Success FactorsThe right tools must be chosen Decide on the right categories of tools Tools must match requirements of the organization, users and project. Tools must work together without the need to build interfaces or special code.How to measure successHow to measure successFunctional quality Do the capabilities of the data warehouse satisfy the user requirements? Does the data warehouse provide the information necessary for the users to do their job?How to measure successHow to measure successData quality Ask the users if their reports are accurate Maintain a scorecard on the quality of the data.How to measure successHow to measure successComputer performance Query response time Report response time Time to load/update/refresh the data warehouse Machine resourcesHow to measure successHow to measure successUser satisfaction Is the DW solving their business problem? Does the DW make their jobs easier? Do they access the DW often to obtain information? Are they asking for more information to be put into the DW? Data Warehouse ApplicationsData Warehouse ApplicationsIs a process … not a one-time project. Is business driven … not product/technology. Effort 80-20 : 80% back-end, 20% front-end. Implementation required … not buy-and-install !!! Is different for every organization. Must be expandable, and more importantly, avoid repetition. Should be built, using integrated technology and tools.SummarySummaryMore organization realize the importance of DW. DW is becoming part of the Internet. Avoid the known failures before even thinking on how to achieve success. Success is not always measured by numbers DW is a solution which must consist of integrated tools, and services. Our Approach and OfferingsOur Approach and OfferingsOur ApproachOur ApproachEnterprise Data WarehouseData TransformationInformation Access Analyst, Executive, MarketingTrend Analysis, etc.Scalability over TimeFast ReturnMiningOLAP EISProcess-Oriented ApproachProcess-Oriented ApproachData Warehousing is a process…not a project. Ensures ability to: Address both known and ad hoc information needs Dynamically respond to changing business environment Support iterative development -- incremental benefit for incremental costArchitecture-Based ApproachArchitecture-Based ApproachDon’t have to choose between doing it right and doing it now Blueprint for long-term evolution Minimize risk Departmental Efforts Changing NeedsMetadata Repository ApproachMetadata Repository ApproachData Warehouse OfferingsData Warehouse OfferingsData Warehouse Planning Data Architecture Definition Data Warehouse Construction Data Mart Construction Data Warehouse Management Data Warehousing Tool Support & Training DecisionBase InfoBeacon InfoReports Forest & TreesSummary of CA’s Key StrengthSummary of CA’s Key StrengthExperience-based Methodology - A reusable methodology End to End - IT to Business Understanding End to End Solution - Building Ground to Roof Fast Return on Investment Scalability, Expandability & Maintainability Openness of Solution No Compromise Data Warehouse Solution DecisionBaseNo Compromise Data Warehouse Solution DecisionBaseTransformation & Movement IssuesTransformation & Movement IssuesFinding the right data to satisfy end user needs Designing the warehouse/mart Moving the right data into the warehouse/mart Building the population jobs Scheduling & monitoring transformation and movement Providing visual access to the transformation and movement process Linking transformation and movement metadata with all other metadata activity The Problems with Hand Coding The Problems with Hand Coding Writing COBOL to access multiple data sources as a coherent whole Implementing a solution before user community needs change Gaining confidence of users by providing the right data from the right sources No easy method for ensuring data consistency over time Huge effort to change target DBMS The DecisionBase Solution The DecisionBase Solution The DecisionBase SolutionThe DecisionBase SolutionGenerates code automatically Combines data from multiple relational and non-relational sources Defines mappings quickly and graphically Records transformations in Repository for future reference Ensures data consistency through Repository technology Leverages ERwin data modelsDecisionBase TechnologyInfoBeacon InfoReports Forest &Trees Business Objects Brio * Cognos *DecisionBase TechnologyData Sources & Targets Data Sources & Targets Oracle Sybase Microsoft SQL Server DB2/MVS and DB2 UDB Informix Flat files (both EBCDIC and ASCII) SAP R/3 (source only) AS/400 Various others via ODBCGraphical MappingGraphical MappingDecisionBase WorkflowDecisionBase WorkflowEnd User ToolsEnd User ToolsForest & Trees InfoBeacon InfoReposrts Business Objects* Cognos* Brio* Single, integrated web browser–metadata, queries, Word documents all in one placeDecisionBaseDecisionBase ArchitectureDecisionBase ArchitectureTARGET WAREHOUSE Competitive AdvantagesCompetitive AdvantagesSingle vendor solution Data design: Erwin Data movement: DecisionBase Metadata: DecisionBase (based on the Microsoft Repository standard) Data access: Forest & Trees, InfoBeacon, Integration with third party tools Business Objects, Brio, CognosQ & AQ & AThe Business ValuesThe Business ValuesChallengesChallenges Liberalization Of Telecommunication Industries in Year 2001 Merges and Acquisitions Increase in Competition Increase in Customer Sophistication Uncertainty Economic Climate Which Way is the Industry Heading ?Monopolistic MarketLiberalized Market Open MarketChallengesChallenges The Rate of of Customer Acquisition is slowing down High Cost of Acquisition High Churn RateWhy The Interest In Data ?Why The Interest In Data ?Turning Data into Information to make Business Decisions Anticipate Problems Anticipate Trends PredictionInformation Helps toTo Increase Business Competitive Advantage Through the Use of IT and InformationInformation is an ASSET to the OrganizationWhy The Interest In Data ?Why The Interest In Data ? THE BOTTOM LINE UNDERSTAND THE CUSTOMER BETTER Call Pattern Service Item Usage Payment Behavior Customer Care Response to Loyalty ProgramThe Information Warrior The Information Warrior Knowledge Customer Centric View Identify New Services / Products Identify and Control Escalating CostsAble to Deliver the Right Information to the Right Person at the Right Time in the Right FormatWhat Does Data Warehouse Offer ?What Does Data Warehouse Offer ?What Does Data Warehouse Offer ?What Does Data Warehouse Offer ? Bringing together the data from disparate sources Creating a single consistent view of corporate information Transforming data into useful information Delivering information to the right peopleKnowledge Management‘ Transforming Corporate Data into Information and Deliver the Information to the right people at the right time ‘ What Can You Do With The Information What Can You Do With The InformationSample of Key Performance IndicatorsChurn Call Usage Call Destination Network Availability Activation (Deactivation, Reactivation) Coverage Subscriber Analysis Responsiveness to Orders Analysis of Faults Revenue FraudDimensionTime Geographical Customer Segment Product Call PlanData Warehouse AnalysisMultidimensional Analysis Trend Analysis Geographical Analysis Comparative Analysis Exception Analysis Ranking Analysis Cause and Effect Analysis What If AnalysisInformationAnalysisData Warehouse AnalysisData Warehouse AnalysisWhat is Data Warehouse Analysis and What Can It Do ? Data Warehouse AnalysisData Warehouse Analysis“Data Warehouse Analytical Capability will help Decision Makers to Identify and Rapidly Respond to Performance Problems and Market Opportunities”Data Warehouse AnalysisMulti Dimensional Analysis At Any Aggregate or Detail Level, Analyze Key Performance Indicators At Different Business Dimension. (Example By Customer Segment, Time, Product, Geography) Data Warehouse AnalysisData Warehouse AnalysisData Warehouse AnalysisTrend Analysis Identify Trends and Potential Outcomes using Both the Goal Based Projection Methods and Modeling Techniques.Data Warehouse AnalysisData Warehouse AnalysisGeographical Analysis Analyze Business Performance Geographically. (Example Cell Utilization, Number of Subscribers in each Geographical Segment)Data Warehouse AnalysisData Warehouse AnalysisComparative Analysis Performance Measurements Must Be Comparable or Benchmarks to be Meaningful. Data Warehouse AnalysisData Warehouse AnalysisRanking Analysis View a “best to worst” Ranking Of products, Customer by Selecting Key Performance Indicators, Business Dimensions, and Measurement Units. (For Example User can identify the best and worst performing products in a particular customer segment, calling plan, time period etc.Data Warehouse AnalysisException Analysis Exception Analysis and Report Techniques Help Users Quickly Identify “out of norm” conditions that requires immediate attention.Data Warehouse AnalysisData Warehouse AnalysisData Warehouse AnalysisCause-and-Effect Analysis Quickly Determine Whether One Measured Performance (e.g Revenue) is Being Driven by Another (e.g Network) By Analyzing Cause-and-Effect Relationships Between Key Performance Indicator. Data Warehouse AnalysisWhat-if-Analysis Immediately Test Alternate Performance Scenarios Dynamically modifying exception analysis thresholds and comparison Data Warehouse AnalysisData Warehouse Analysis - Our SolutionData Warehouse Analysis - Our SolutionOnline Analytical Process Technology Trend Analysis; Exception Alert; Forecasting; Productivity Measurement; Data Mining Technology Cluster Similar Behavior; Market Segmentation; Customer Targeting; Prediction and Impact Analysis;Decision Support SystemDecision Support SystemMonopolistic MarketLiberalized Market Open MarketHow Can Decision Support Help ?Increase Competition From External Carrier Increase Customer Sophistication Resulting from CompetitionIssues & ChallengesSituationCounteractDifferentiate & Improve Customer Segmentation Enhance Target MarketingData Warehouse At WorkData Warehouse At WorkTELCO IT ArchitectureTELCO IT ArchitectureBilling InformationPayment InformationCustomer InformationCustomer Care InformationFinancial InformationCall InformationContract InformationService Package InformationService Item InformationContract HistoryIn Bound CallOut Bound CallRate Plan InformationBalance Sheet Income StatementTransactional InformationWarehouse Conceptual Model - The ArchitectureWarehouse Conceptual Model - The ArchitectureCustomerContractBillingPaymentCustomer DemographicContract HistoryService PackagesReason Status Call RecordsCall DemographicPeriodCustomer CarePromotionDecision Support Application - Customer ProfilingDecision Support Application - Customer ProfilingCustomerService PlanningNetwork UtilizationCustomer CareCustomer AcquisitionCustomer RetentionCustomer Billing/ PaymentFraudCustomer LoyaltyDecision Support Application - Customer ProfilingCustomer Care Customer Complain and Response; Customer Complain Analysis; Customer Ranking Analysis; Customer Complain Growth Rate; Response to Complain Analysis.EXAMPLE Proactively anticipate growing trends of problems Anticipate and increase Quality Of ServiceDecision Support Application - Customer ProfilingDecision Support Application - Customer ProfilingService Planning Service Item Utilization; Market Basket Analysis; Price and Rate Plan Analysis.EXAMPLE Product Offering and Service Packaging Cross Selling of ProductDecision Support Application - Customer ProfilingDecision Support Application - Customer ProfilingNetwork Utilization Roaming Analysis; Revenue and Profitability Analysis; Call Pattern Analysis; Network Utilization Response to Promotion and Marketing Campaign; Call Volume Analysis; Subscriber Usage Pattern Analysis; Outgoing and Incoming Traffic Analysis; Customer Ranking Analysis; Cell Utilization and Growth Patterns.Decision Support Application - Customer ProfilingDecision Support Application - Customer ProfilingCustomer Loyalty Customer Reward Program Analysis; Response to Marketing Campaign Analysis.Decision Support Application - Customer ProfilingCampaign Management Productivity Measurement Decision Support Application - Customer ProfilingFraud Potential Fraud Analysis; Call Tracking Analysis ( Alert and Exception Tracking ).Decision Support Application - Customer ProfilingIs there a potential fraud ? Is this call make normal ?Decision Support Application - Customer ProfilingCustomer Billing/ Payment Customer Billing By Service Packages; Customer Aging Analysis; Customer Payment Analysis; Bad Debt Management Analysis; Over Credit Limit Analysis; Dunning Analysis; Credit & Rebate Analysis; Profitability and Revenue Analysis.Decision Support Application - Customer Profiling Focus in Bad Debt Management Which customer is most likely to fraudDecision Support Application - Customer ProfilingCustomer Acquisition Customer Acquisition by Promotion and Marketing Campaign; Customer Acquisition by Service Package; Customer Acquisition by Area; Customer Acquisition by Customer Demographic; Growth Rate Analysis (Trend); Customer Stay Duration Analysis; Actual VS Plan Analysis (Comparative Analysis); Profitability Analysis.Decision Support Application - Customer ProfilingWho are my customer Buying Trend Average life span of my customerDecision Support Application - Customer ProfilingCustomer Retention Churn Rate Analysis; Suspend Subscriber Analysis; Customer Complain Analysis; Network and Service Quality Analysis; Customer Deactivation Analysis; Service Duration Analysis; Growth Rate; Churn Cause Analysis; Profitability Analysis.Decision Support Application - Customer ProfilingDecision Support Application - Other ExampleDecision Support Application - Other ExampleChurn Management Campaign Management How to Profile a Prepaid Customer ?Decision Support Application - BenefitsDecision Support A