extreme scoping 2006 cincinnati - swoc dama · data warehouse role in bi enterprise data warehouse...
TRANSCRIPT
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
“Extreme Scoping”What iterative development really means
presented by
Larissa T. MossPresident, Method Focus, Inc.
DAMA – CincinnatiOctober 25, 2006
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
OutlineWhy traditional project management (PM) does not work on DW projectsCommon DW failures and PM challengesSpiral DW methodologySoftware release conceptSelf-organizing project teamsDifferent project planning processBI program/portfolio managementBest practices/critical success factorsOrganizational culture change
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Business Intelligence …
…is a cross-organizational disciplineand an enterprise architecturefor an integrated collection of
operational as well as decision support applications and databases,
which provide the business communityeasy access to their business data, and
allows them to make accurate business decisions.
… is not a system or a product
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
From Chaos to Architecture
DataMart
DataMart
DataMartOperational
Data Store Enterprise
Data Warehouse
OperMart
OperMart
OperMart
CRMAnal
ExplorationWarehouse
BI Access & Delivery
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Data Warehouse Role in BI
EnterpriseEnterpriseDataData
WarehouseWarehouse
OperationalOperationalDataDataStoreStore
Dat
a M
anag
emen
t
Dat
a D
eliv
ery
Order Entry
ERPBack Office
SalesAutomation
CRMFront Office
AccountsReceivable
OtherLegacy
&Historical
CustomerDemogr.
Web&
External
ExplorationExplorationWarehouseWarehouse
MarketingMarketingData MartData Mart
Data Mining
CRM Analytical
OperationalOperationalMartMart
Other BI apps.
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
DW Goals and Objectives
1. Data Management• data integration• data cleansing
Get control over theexisting data chaos
2. Data Delivery• data access• data manipulation
Provide intuitive accessto business information
Business Intelligence Framework
80% 20%
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Traditional Development MethodOperational
Systems(Legacy)
Decision SupportSystems Business Units
L
L
L
L
FinanceFA
EngineeringEN
Product SalesPS
Customer SupportCS
MarketingMK
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Uncoordinated Development
Business Technology
paired up
BusinessManager
BusinessManager
ITManager
ITManager
... ...
... ...
ChiefExecutive
Officer
ChiefOperating
Officer
ChiefInformation
Officer
Business Units
Marketing
Financial
Product
Custom
er
Distribution
Inventory
Sales
KW KW KW KW KW KW KW
IT IT IT IT IT IT IT
Information Technology Units
Do You Know Your Business ?
• data redundancy• process redundancy• dirty data
“swim lane development”
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Waterfall Methodologies
Post-MortemReview
Business Need
Project Plan
FunctionalRequirements
SystemAnalysis
Development
Testing
ImplementationProductionSystem Maintenance
ClientApproval
System Design
External Design
Internal Design
“Silos”
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Using the Waterfall Approach
Mistakes are less expensive to fix early in the development process!
Requirements are well definedScope is manageableScope and users are confined to one business unit
(department, business function) Technology infrastructure is known and provenData volumes are relatively smallDevelopment activities are the same on every projectDevelopment activities are not cross-organizationalProject schedules are relatively easy to estimate
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Industrial-Age Mental Model
TIMESCOPE
BUDGET
PEOPLE
QUALITY
1 2 3 4 5
highest to lowest priorityPr
ojec
t Con
stra
ints
• Cheaper, faster, better • Automate as quickly as possible
Priority
Cost-based value proposition
“Scrap and rework”
(Larry English)
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Time versus QualityEveryone wants quality, but rarely is the extra time given or taken to achieve it.
Quality and time are polarized constraints.
The higher the quality the more time it takes to deliver.
Companies are driven by shorter and shorter schedules.
SCOPE
TIMEYAH DDD
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
The Time Trap
Redundancy increases
Developmenttime increases
Employeeburnoutincreases
Quality decreases
Usersatisfaction decreases
Maintenance increases
Need for morestaff increases
Business value decreases
Productivity decreases
Inability to support business drivers
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Proliferation of RedundancyLegacy Data Warehouses
Data Marts
Marketing
Finance
Product Sales
Engineering
Users
L
L
L
L
DM
DM
EDW
DM
DM
transformation?cleansing? Customer Support
BI ?BI ?
Framework?
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Adding to the Data Chaos
DW andData Marts
DM
DM
EDW
DM
DM
transformation?cleansing?
DMDM
DM
DMEDW
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
What’s different about DW projectsData Warehouse is not a system or productDatabases are structured differentlyUsers are more directly involvedDirty data becomes more visibleMore integration is requiredStandards are neededDifficult to justify costTimeframes are tighterPossibility of failure is increased
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Over budgetSlipped scheduleFunctions not implementedUnhappy usersPoor performancePoor availabilityPoor or no scalabilityBenefits never materializedCost exceeded benefits
Common DW Failures
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Too complicated for users Poor quality data/reportsDirty data was underestimatedDW development was treated like any other system development projectTraditional methodology did not workRAD meant skip analysis and design
Common DW Failures
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Common DW Failures
DW techniques were not understood or deployedETL process did not fit into batch windowMeta data is not available to usersDirty data continues to existData is still not integratedEveryone is building stovepipe “Lega” Marts
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Unrealistic scheduleUnrealistic user expectationsBudget too smallNo management commitmentNo user involvementLost or changed sponsorLost or changed staff (high turnover)Untrained or unavailable staff
Project Management Challenges
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Management Challenges
Disjointed DW administration (no BI Program Management)Changing business prioritiesUnclear or changing requirements (scope creep) New technology not understoodNo resources for data standardizationNo dispute resolution procedure
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
The Lesson?You cannot keep doing
what you have always doneand expect the results to be different.
“That wouldn’t be logical”Spock, Star Trek
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Information-Age Mental Model
QUALITYTIME
PEOPLE
BUDGET
SCOPE
1 2 3 4 5
Proj
ect C
onst
rain
ts
Priority
highest to lowest priority
• Reassemble the entire enterprise • Reuse assets from inventory
Investment-basedvalue proposition
“Reassemble reusable components”
(John Zachman)
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Coordinated Development
ODSDM
Discover, Communicate, Integrate, Document, Coordinate
Operational Environment
EDWOM
BI Databases
Decision Support Environment
Operational Systems
“cross-organizational development”
Business Units
Marketing
Financial
Product
Custom
er
Distribution
Inventory
Sales
KW KW KW KW KW KW KW
IT IT IT IT IT IT IT
Information Technology Units
EnterpriseInformationManagement
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Cross-Organizational DevelopmentCommitment to data integration embedded in the methodologyCross-organizational program managementEnterprise information management groupStandards that include a common information architecture (enterprise data model)Coordinating the development/ETL processes
Disallowing stovepipe developmentExtracting and cleansing source data only onceReconciling data transformations and storing the reconciliation totals as meta data
< principles
< governance
< resources
< policies
< enforcement
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Fundamental Changes1. Stop working in isolated “swim lanes”
Standardize and “share” data2. Centrally manage data like a business asset
(Enterprise Information Management [EIM])Assemble data as needed from the data inventory (enterprise data model and meta data)Standardize and reconcile data transformations Standardize business metrics for BI applications
3. Scale down project scopes to incorporate EIM activitiesUse the software release concept
4. Modify your methodology to be based on cross-organizational disciplines
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Engineering Stages
Deployment
Justification Readiness
Planning Preparation
BusinessAnalysis Integration
Construction Solution
Design ProductIterative
Flexibleentry points
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Development Steps
1Business CaseAssessment
Justification
Business Analysis
4Requirements
Definition
5Data
Analysis
6ApplicationPrototyping
7Meta Data RepositoryAnalysis
2Enterprise
InfrastructureEvaluation
3Project
Planning
Planning
Design
10Meta Data Repository
Design
8Database
Design
9ETL Design
Construction
14Meta DataRepositoryDevelopmt
13Data Mining
12ApplicationDevelopmt
11ETL
Developmt
15Implementatn
16Release
Evaluation
Deployment
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Business Case Assessment
Business problem or business opportunity is definedReadiness and feasibility of a BI solution is assessedThe proposed BI application is cost-justifiedRamifications of not addressing the business problem or business opportunity are documentedBusiness sponsorship is obtained
Justification
Business needsBusiness needs
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Enterprise Infrastructure Evaluation
Technical infrastructure gap analysis is performedTechnology platform is upgraded– Hardware, middleware, network components– Software, operating systems, tools
Non-technical infrastructure gap analysis is performedNon-technical infrastructure components are enhanced– Standards, procedures, business rules– Enterprise architecture, methodology– Scope management (change control)– Roles and responsibilities, team structure
Planning
Technology gapsTechnology gaps
Procedural gapsProcedural gaps
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project PlanningHigh-level BI project requirements defined for– Data, Functionality– Infrastructure components (technical and non-technical) Roles and responsibilities are defined and assignedTeam structures are createdProject constraints, assumptions, and risks are identifiedThe project charter is writtenThe project plan is developed– Source data is analyzed to get realistic ETL estimates– Cross-organizational activities are taken
into consideration when estimating tasks
Planning
Game plan Game plan
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Requirements DefinitionProject requirements are compared to project constraintsThe BI application is separated into multiple releases for iterative deployment (scope management)An application requirements document is written, describing the project scope in terms of – Infrastructure requirements– Access and analysis requirements– Data requirements– Meta data requirements– Security requirements
BusinessAnalysis
Product definitionProduct definition
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Data AnalysisExtensive businessbusiness analysis of data is performedA normalized logical data model (business view of data) is created (not database design!)Business meta data is defined and capturedData-related business rules are defined and capturedSource data is analyzed during modeling sessions Project-specific data model is merged into the enterprise data model (enterprise information architecture)Data definitions and data domains are standardizedData disputes are resolved with data owners and information consumers
BusinessAnalysis
Information architectureInformation architecture
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Application PrototypingUsed to be called systemssystems analysisFunctional project deliverables are studiedImplementation solutions are tried and tested– Critical or difficult deliverables are partially
designed, developed, and tested (prototyped)– Technology and tools are tested– User interfaces are testedBI project requirements are again compared to project constraintsBI project requirements are revised
BusinessAnalysis
FeasibilityFeasibility
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Meta Data Repository Analysis
Business meta data components are identifiedTechnical meta data components are identifiedSources of meta data (DBMS and tool dictionaries) are identified and analyzedA meta model of required meta data components is created Meta-meta data is created
BusinessAnalysis
ContextContext
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Database DesignCommonly called “logical database design” (not to be confused with logical data modeling!)Appropriate design schema for BI target databases are selected (relational or multi-dimensional)A denormalized physical data model (database view) is created, either entity-relationship or star-schemaTechnical meta data is defined and capturedPhysical database design is performed– Indexing strategy– Dataset and index placement– Partitioning– Clustering
Design
Data storage Data storage
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
ETL DesignSource data is mapped into BI target databasesThe process flow is designed for extracting, sorting, merging, cleansing, transforming, and loading dataAn ETL process flow diagram is createdThe ETL staging area is defined or enhanced (can be centralized or decentralized, but all ETL activities must be coordinated)Common reconciliation and error handling routines are designedData transformation and cleansing requirements are designed
Design
MigrationMigration
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Meta Data Repository DesignDecision to build or buy a meta data repository is madeIf the decision is to build, its components are designed – Meta data repository database (E-R or OO)– Meta data migration (load) process– Meta data application (reports, user interface)– Meta data online help functionIf the decision is to buy, a meta data repository product has to be evaluated and selected– Products have to be evaluated– Vendors have to be evaluated– RFP or RFQ is issued
Design
Context storageContext storage
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
ETL Development
Three ETL processes are developed – Initial load– Historical load– Incremental loadETL processes are tested– Integration or regression tested– Performance or stress tested– QA and acceptance testedETL technical meta data is defined and captured
Construction
SourcingSourcing
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Application Development
Can be final iteration of an “operational prototype”Access and analysis deliverables are finalized for this BI project releaseApplication programs are tested– Integration or regression tested– Performance or stress tested– QA and acceptance testedAggregation, summarization, and other algorithmic technical meta data is defined, standardized, and captured
Construction
DeliveryDelivery
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Data Mining
Data mining tool is installed and testedData mining databases are designed and createdAnalytical data models are developed and “trained” Source data is prepared (consolidated, cleansed, filtered, aggregated)Data mining databases are loadedData mining operations are selected and runData mining results are compared to industry models
Construction
Knowledge discoveryKnowledge discovery
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Meta Data Repository Development
The meta data repository database is built (or enhanced)The meta data repository programs are developed and tested – Integration or regression tested– QA and acceptance testedETL-specific meta data programs are included in the ETL job schedule– Data quality statistics– ETL reconciliation totals– ETL errors and rejections
Construction
Context deliveryContext delivery
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
ImplementationCreate BI target databases in productionCreate the meta data repository database in productionCreate production program libraries (ETL, application, meta data repository)Move all programs into production librariesApply appropriate security measuresLoad BI target databasesLoad the meta data repositoryConduct training
Deployment
SolutionSolution
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Release Evaluation
A post-implementation review (PIR) is organized– PIRs are conducted about 2 months after
implementation– All project team members are invited to participate– Meeting should be held off-site– Meeting should be facilitated by third-party– Meeting should be documented by third-party scribeAfter the PIR, action items are followed up onSome action items may be included as requirements in future BI application releases
Deployment
OptimizationOptimization
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Business Integration Activities
Cross-organizationalCross-organizationalCross-organizationalProject-specificProject-specificCross-organizationalProject-specificCross-organizationalCross-organizationalCross-organizationalCross-organizationalCross-organizationalProject-specificCross-organizationalCross-organizationalProject-specificCross-organizational
Development Steps1. Business Case Assessment ........................2.A Technical Infrastructure Evaluation ...........2.B Non-Technical Infrastructure Evaluation ...3. Project Planning ..........................................4. Requirements Definition ..............................5. Data Analysis ...............................................6. Application Prototyping ...............................7. Meta Data Repository Analysis ...................8. Database Design .........................................9. ETL Design .......................................….......10. Meta Data Repository Design ....................11. ETL Development .....................................12. Application Development .........................13. Data Mining ..............................................14. Meta Data Repository Development ........15. Implementation .........................................16. Release Evaluation ...................................
Enterprise InformationManagementTouch Points
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Parallel Development Activities1
Bus. CaseAssess.Justification
Go/No-Go decision Project kick-off2
EnterpriseInfrastr.
3Project
Planning
4RequiremtDefinition
Planning(Analysis)
5Data
Analysis
7MDR
Analysis
10MDR
Design
9ETL
Design
14MDR
Develop.
13Data
Mining
12ApplicationDevelop.
11ETL
Develop.
6Prototyping
8DB Design
BackEndETL
FrontEnd Appl.
MetaDataRep.
Analysis
Design
Construction
15Implement.
16Release
Eval.Deployment
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project DynamicsBusiness CaseAssessment Planning and
Requirements
Data Analysis
ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta Data RepositoryDesign
ETLDevelopment Application
Prototyping
Meta DataRepositoryDevelopment
Database Design
Data Mining
ApplicationDevelopment
Release Implementation
ETLTesting
Meta DataRepositoryTesting
Application Testing
BI/DW Project
ETL Design
< parallel< parallel
< iterative< iterative
< dynamic< dynamic
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Concept
Project = ApplicationProject = Application//
“Refactoring”- Kent Beck
“Extreme scoping”- Larissa Moss
“feels like prototyping”
SecondRelease
FirstRelease
FourthRelease
Reusable &Expanding
FinalRelease
BI Application
FifthRelease
ThirdRelease
Projects
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Using the Software Release Approach
Mistakes are less expensive to fix early in the development process!
Unstable requirements can be tested and enhanced in small incrementsScope is very small and manageableTechnology infrastructure can be tested and provenData volumes (per release) are relatively smallProject schedules are easier to estimate because the scope is very smallDevelopment activities can be iteratively refined, honed, and adapted
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Using the Software Release Approach
And the quality of the release deliverables (and ultimatelythe quality of the DW applications) will be higher!
And the development process will get faster and faster!
Unstable requirements can be tested and enhanced in small incrementsScope is very small and manageableTechnology infrastructure can be tested and provenData volumes (per release) are relatively smallProject schedules are easier to estimate because the scope is very smallDevelopment activities can be iteratively refined, honed, and adapted
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Guidelines
Deliver every three to six months (first release will take longer)Strictly control the scope and keep it very smallKeep expectations realisticThe enterprise infrastructure must be robust (technical and non-technical)Meta data must be an integral part of each release; otherwise, the releases will not be manageableDesigns, programs, and tools must be flexible
SecondRelease
FirstRelease
FourthRelease
FinalRelease
BI Application
FifthRelease
ThirdRelease
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Team Organization
EXTENDED
TechnicalServices
StrategicArchitect
Auditor
SecurityOfficer
End userSupport
BusinessSponsor
Operations
ProjectManagement
TechnicalSkills
EIMSkills
COREBusiness
Participation
ETLTrack
Appl.Track
MDRTrack
Programmers
TechnicalAdvisoryBoard
BusinessAdvisoryBoard
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Team StructureProject core team
Self-organizing SWAT teamOptimum team size: 3 – 5 people; never more than 7Full-time involvement on the project from beginning to endMeet daily to control project activities
Development track teamSelf-organizing XP (extreme programming) teamOptimum team size: 2-3 people; never more than 5Full-time involvement on their development trackMeet daily to control development activities
Extended teamInvolvement on an as-needed basisSupport roles
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Team StaffingProject core team
One hands-on project managerOne business representative with some decision-making authorityOne business analyst from IT (data administrator or business liaison)One technical lead person from IT (senior programmer or senior systems analyst)
Development track teamOne senior programmerOne programmer or programmer/analyst
Extended team Business sponsor Technical support teamOperationsEtc.
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Core Team RolesApplication lead developerBI infrastructure architectBusiness representative Data administrator (EIM)Data quality analyst Database administrator/designer ETL lead developerMeta data administrator (EIM) Project managerSubject matter expert
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Application Lead Developer
Designs and leads the development of the BI application’s access and analysis componentsMentors the “front-end” application developersWrites and tests the more complex queriesSupervises data access analysis and prototypingactivities of the data delivery componentsPerforms stress testing of the BI applicationUpdates the technical meta data in the repository
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
BI Infrastructure Architect
Identifies the technology requirements to support the BI goals and objectivesArchitects, installs, and maintains the technical infrastructure componentsSets technology priorities based on the strategic direction of the organizationParticipates in tool evaluation and selectionMonitors BI application use, trends, performance
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Business RepresentativeHelps with project planning Determines deliverablesParticipates in tool evaluation and selectionIdentifies data and functional requirementsParticipates in data and process modeling sessionsParticipates in prototyping sessionsProvides data definitionsParticipates in testing activitiesValidates the quality of dataValidates the accuracy of reports and queriesIdentifies/validates transformation rules for source dataResolves disputes among business units
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Data AdministratorPerforms cross-organizational data analysisCreates and manages the logical data modelEnsures the quality and integrity of the data modelManages the integrated enterprise data modelCaptures the business meta dataDevelops data standards Administers, manages and controls the CASE toolsAssists data quality analyst with dirty data discoveriesHelps the meta data administrator with populating themeta data repositoryParticipates with DBA in logical database design activities
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Data Quality Analyst
Analyzes the source files and identifies dirty dataPerforms source to target data mapping (ETL specs)Creates specifications for ETL reconciliation totalsLooks for opportunities to re-use and integrate dataProvides system verification of data definitionsProvides system verification of business rulesHelps with capturing meta data into the repositoryHelps with training business people
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Database Administrator/Designer
Designs the BI target databasesCreates physical database structuresResponsible for database maintenance(Monitoring, tuning, reorganization, backup and recovery)
Participates in extracting and loading datafrom the source systems to the DW databasesImplements database securityProvides internal support for DBMS softwareAssisting and enhancing the use of query toolsto access the DW databases
DatabaseDesign
DBMS environment
Applicationsupport
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
ETL Lead Developer
Designs and leads the development of the ETL process Mentors the “back-end” ETL developersHelps to evaluate, select, install, and test the ETL toolCreates and maintains the process flow diagram for the ETL processHelps in the selection of which source files to use when there are multiple sourcesPerforms stress testing of the ETL processUpdates the technical meta data in the repository
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Meta Data Administrator
Creates the meta model for the meta data repositoryBuilds or selects a meta data repositoryImplements the meta data repositoryProvides access to the meta data repositoryAdministers, manages and controls the meta data repositoryMerges and loads meta data from the other tools into the meta data repositoryParticipates with DA in establishing data standards
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Manager
Overall responsibility for the BI projectDefines, plans, controls, reviews the project activitiesTracks and reports on project progressStaffs the project with the right peopleEstimates and measures the costs and benefitsActs as a quality reviewer throughout the projectNegotiates with tool vendors, business representatives, and the business sponsorResolves technical issuesRaises disputes to the advisory boards
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Subject Matter Expert
Provides business knowledge for a subject areaMakes suggestions on the required format of the data
(detailed, summarized, aggregated, derived and historical)
Writes test cases and participates in all testing activitiesActively participates in application prototypingHelps establish criteria for query tool selectionTrains business people in the use of query toolsIdentifies source data quality problems and establishes processes to resolve them
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Extended Team RolesApplication developersBI support staff (helpdesk)Business sponsorData MinerETL developersIT auditor or QA analystMeta data repository developerNetwork servicesOperationsSecurity officerStakeholders
Technical services TestersTool administratorWeb developersWebmaster
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
592 Write query script specifications 587 593 Write access interface programming specifications 588 594 Write online help function programming specifications 589 595 Step 11: Extract/Transform/Load Development 596 Build and unit test the ETL process 542, 549 597 Code the ETL programs 598 If using an ETL tool, write instructions for ETL tool modules 599 Capture the ETL technical metadata for the metadata repository 600 Write code to produce reconciliation totals, quality metrics, and load statistics 601 Unit test each individual program module 597, 600 602 If using an ETL tool, unit test each ETL tool module 598, 600 603 Write the scripts to execute the ETL programs and the sort, merge, and load utilities 601, 602 604 Integration or regression test the ETL process 596 605 Create a test plan with test cases for the ETL process 606 Create test data for the ETL programs 607 Integration or regression test the entire ETL process 606, 606 608 Log the actual test results and document any test issues
Methodology
from Business Intelligence Roadmap (Moss & Atre)
• activities• dependencies• deliverables• roles
Work Breakdown Structure (WBS)
describes
• tasks• sequence• schedule• resources• milestones• checkpoints
Traditional Project Plan
defines
refines
Software Releases & Project Activities• number of releases• day-to-day work activities• progress tracking and status reporting• change control and impact analysis
drives
Based on source: TDWI
overall effort,resources,
cost
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
2. Break BI application into software releases2. Break BI application into software releases
SecondRelease
FirstRelease
FourthRelease
FinalRelease
BI Application
NthRelease
ThirdRelease
20062006
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process
3. Create milestones from DDD to YAH for first release 3. Create milestones from DDD to YAH for first release
1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
2. Break BI application into software releases2. Break BI application into software releases
FirstRelease
Milestone
Activity
Task/SubtaskTask
Deliverable
PROJECT
Highest level of projectpartitioning
Major work effort of 1-3 weeks consisting of one or more tasks
A unit of work that may be measured in days or hours produces
produces MilestoneDeliverable
reported
managed
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process
4. Organize and assign parallel development tracks4. Organize and assign parallel development tracks
3. Create milestones from DDD to YAH for first release 3. Create milestones from DDD to YAH for first release
1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
2. Break BI application into software releases2. Break BI application into software releases
1Bus. Case
Assess.
2Enterprise
Infrastr.
3Project
Planning
4RequiremtDefinition
5Data
Analysis
7MDR
Analysis
10MDR
Design
9ETL
Design
14MDR
Develop.
13Data
Mining
12ApplicationDevelop.
11ETL
Develop.
6Prototyping
8DB Design
BackEndETL
FrontEnd Appl.
MetaDataRep.
15Implement.
16Release
Eval.
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process
5. Create detailed work assignments for all milestones5. Create detailed work assignments for all milestones
4. Organize and assign parallel development tracks4. Organize and assign parallel development tracks
3. Create milestones from DDD to YAH for first release 3. Create milestones from DDD to YAH for first release
1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
2. Break BI application into software releases2. Break BI application into software releases
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process
6. Create Gantt chart showing milestones6. Create Gantt chart showing milestones
5. Create detailed work assignments for all milestones5. Create detailed work assignments for all milestones
4. Organize and assign parallel development tracks4. Organize and assign parallel development tracks
3. Create milestones from DDD to YAH for first release 3. Create milestones from DDD to YAH for first release
1. Estimate development effort for entire BI application1. Estimate development effort for entire BI application
2. Break BI application into software releases2. Break BI application into software releases
ID Milestone Name DurationDetermine problem/opportunity
Analyze cost/benefit of new projectDetermine critical success factorsAnalyze resources available
Approval of project
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Software Release Planning Process
6. Create Gantt chart showing milestones6. Create Gantt chart showing milestones
5. Create detailed work assignments for all milestones5. Create detailed work assignments for all milestones
4. Organize and assign parallel development tracks4. Organize and assign parallel development tracks
3. Create milestones from DDD to YAH for next release 3. Create milestones from DDD to YAH for next release
1. Review development effort for entire BI application1. Review development effort for entire BI application
2. Review and adjust BI application software releases2. Review and adjust BI application software releases
SecondRelease
FourthRelease
FinalRelease
NthRelease
ThirdRelease
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Iterative BI Application DevelopmentPlanning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Planning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Planning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Planning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Planning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Planning
Requiremts& Data Analysis
Requiremts& ApplicationPrototyping
Meta DataRepository Analysis
ETLDesign
ApplicationPrototyping
Meta DataRepository Design
ETLDevelopment
Meta DataRepositoryDevelopment
Data Analysis
Data Mining
ApplicationDevelopment
ETLTesting
Meta DataRepositoryTesting
ApplicationTesting
ETL Design
ApplicationPrototyping
Release Implementatn
BusinessCase
Assessment
Post-Impl.Review
Release 1
Release 2
Release 3
Release 4
Release 5
Release 6
BIApplication
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Flexible Entry and Exit Points
Third release of a Marketing data martDeliverables:
20 new data elements from a new internal source 2 new reports
Data will be added to the existing BI databaseExisting report writer will be used (no new tool)Meta data repository is not in scope (separate project)Online help function is not in scope (last release)Data mining is not part of the requirements at allBudget and resources are not an issueNo post-implementation review necessary after this minor release
Example:Example: SecondRelease
FirstRelease
FourthRelease
FinalRelease
BI Application
FifthRelease
ThirdRelease
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Entry and Exit Criteria
1Business CaseAssessment
Justification
Business Analysis
4Requirements
Definition
5Data
Analysis
6ApplicationPrototyping
7Meta Data RepositoryAnalysis
2Enterprise
InfrastructureEvaluation
3Project
Planning
Planning
Design
10Meta Data Repository
Design
8Database
Design
9ETL Design
Construction
14Meta DataRepositoryDevelopmt
13Data Mining
12ApplicationDevelopmt
11ETL
Developmt
15Implementatn
16Release
Evaluation
Deployment
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Organization
3Project
Planning
9ETL
Design
11ETL
Develop.
8DB Design
BackEndETL
Front-End Applicatn.
15Implmnt.
4RequiremtDefinition5
DataAnalysis
4RequiremtDefinition
6Prototyping
12ApplicationDevelop.
Planning(Analysis)
Analysis
Design
Construction
Deployment
ThirdRelease
Milestone
Activity
Task/SubtaskTask
Deliverable
PROJECT
Highest level of projectpartitioning
Major work effort of 1-3 weeks consisting of one or more tasks
A unit of work that may be measured in days or hours
produces
produces
MilestoneDeliverable
reported
managed
ID
Milestone Name
Duration
Determine problem/opportunityAnalyze cost/benefit of new project
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
BI Program/Portfolio Management
• Common Development Methodology• Common Enterprise Standards • Common Ratified Business Rules• Common Enterprise Infrastructure
• Coordinated ETL Processes • Common Meta Data• Common Enterprise
Information Architecture
ProjectManagement
TechnicalSkills
EIM Skills
COREEXTENDED
TechnicalServices
StrategicArchitect
Auditor
SecurityOfficer
End userSupport
BusinessParticipation
BusinessSponsor
ETLTrack
Appl.Track
MDRTrack
Programmers
Operations
ProjectManagement
TechnicalSkills
EIM Skills
COREEXTENDED
TechnicalServices
StrategicArchitect
Auditor
SecurityOfficer
End userSupport
BusinessParticipation
BusinessSponsor
ETLTrack
Appl.Track
MDRTrack
Programmers
Operations
ProjectManagement
TechnicalSkills
EIM Skills
COREEXTENDED
TechnicalServices
StrategicArchitect
Auditor
SecurityOfficer
End userSupport
BusinessParticipation
BusinessSponsor
ETLTrack
Appl.Track
MDRTrack
Programmers
Operations
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
BI Program Management ComponentsBI steering committee made up of business executives and senior business managers Collective sponsorship from business executivesCommunicate enterprise-wide data integration principles to all lines of business (LOB) [ongoing]Create a BI strategy appropriate for the organizationCreate an enterprise information management groupIdentify data owners and data stewards in all LOBAssign business representatives to CORE teams
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
BI Program Management ComponentsHeaded by a BI program manager or BI directorPerform periodic readiness assessments
Identify new information needsAscertain user satisfaction
Create and enforce a common non-technical infrastructure (“Core Competencies”)Prioritize DW projects in alignment with strategic goals (BI application management)Determine DW project interdependenciesCoordinate DW project resources and activities (Software release management)
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
BI Maturity Model
1. Prenatal 2. Infant 3. Child 4. Teenager 5. Adult 6. Sage
“Production Reporting”
“Spreadmarts”
“Data Marts”
“Data Warehouses”
“Enterprise DW”
“Analytic Services”
GULF CHASM
Source: Wayne Eckerson, TDWI
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Best Practices/Critical Success FactorsStop building silo point solutions for individual usersCreate a BI Program with executive business sponsorsNever sacrifice quality for time or scopeProfile your source data before committing to a project delivery date (deadline)Involve enterprise information management groupUse a spiral, cross-organizational DW methodologyPrototype extensivelyBuild and deliver each BI application in multiple 90/120-day software releases Organize the project work into parallel (simultaneous) tracksCreate small self-organizing SWAT teamsInsist on daily user participation on project activities
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Over budgetSlipped scheduleFunctions not implementedUnhappy usersPoor performancePoor availabilityPoor or no scalabilityBenefits never materializedCost exceeded benefits
Common DW Failures
Software releases
User participation on CORE team
Spiral methodology
BI Program Management
Software releases
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Too complicated for users Poor quality data/reportsDirty data was underestimatedDW development was treated like any other system development projectTraditional methodology did not workRAD meant skip analysis and design
Common DW Failures
User participation on CORE team
Spiral methodology
Enterprise Information Management
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Common DW Failures
DW techniques were not understood or deployedETL process did not fit into batch windowMeta data is not available to usersDirty data continues to existData is still not integratedEveryone is building stovepipe “Lega” Marts
Software releasesand prototyping
BI Program Management
Spiral methodology
Enterprise Information Management
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Unrealistic scheduleUnrealistic user expectationsBudget too smallNo management commitmentNo user involvementLost or changed sponsorLost or changed staff (high turnover)Untrained or unavailable staff
Project Management Challenges
User participation on CORE team
BI Program Management with collective sponsorship
Self-organizing SWAT team
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Project Management Challenges
Disjointed DW administrationChanging business prioritiesUnclear or changing requirements (scope creep) New technology not understoodNo resources for data standardizationNo dispute resolution procedure
BI Program Management
Software releases and user participation on CORE team
BI Program Management
Prototyping
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Organizational ImpactNo BI program management in placeCross-organizational tasks and responsibilities are not well definedNew project dynamics are not understoodTeams are still organized the traditional wayBusiness people still don’t participate on project activitiesData quality responsibility is not clear or ignoredDW projects are usually cost justified using the industrial-age mental model (no ROA)No reward for data sharingResistance to change
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Inevitable Culture Shift
Business and IT collaboration (“partnership”)Business and business collaboration (“partnership”)IT and IT collaboration (“partnership”)Shifted roles and responsibilitiesCommon architecture Enterprise standardsCross-organizational activitiesSoftware release concept
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Power Shift
Business people must play an increased role and assume responsibilities for:
Data ownership – Information ArchitectureEnforcement of standardization – IncentivesIntegration and standardization across allbusiness units – Cross-Organizational DisciplinesDispute resolution across all BI projects – LeadershipGovernance – BI Program Management
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Charge-Back PolicySystem ownership – by business unit
Automate a business process from scratchCost-based value propositionBusiness units pay for their own systemsSystem = programs coupled with data
old
new Information ownership – by the enterpriseReassemble the entire enterprise quicklyInvestment-based value propositionOrganization pays for enterprise dataBusiness units pool to pay for shared BI applicationsBI application = BI tools and programs decoupled from data
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
New Incentive Policy
Applicable to the business side as well as IT.
Management and staff should be evaluated on:Team work and collaboration (IT and business)Data sharing (business)Data quality (business)Adherence to standards (IT and business)Adherence to cross-organizational development approach (IT and business)
Rewards (salary increases and bonuses) should be tied to above performance measures
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
New LeadershipCEO
CIOCKOCOO
DG ...EA,EIM,
MDM,DW,
BI,CRM,
CDI,PIM,
…
LOB Execs IT Execs
ChiefKnowledge
Officer
CFOcollaboration collaboration
Copyright 2004-2006, Larissa T. Moss, Method Focus, Inc.
Thank You
Larissa T. MossMethod Focus, [email protected]
ISBN 0-201-61635-1
ISBN 0-201-78420-3
ISBN 0-201-76033-9
ISBN 0-321-24099-5