reflections on alignment with togaf 5 th november 2008 james dawson - consulting enterprise...
TRANSCRIPT
Reflections on alignment with TOGAF
5th November 2008
James Dawson - Consulting Enterprise Architect
“Unlocking the Information Assets”
Monitoring, explaining, predicting and optimising all aspects of the business
AOEGA AGM and Case Studies
Purpose of Today’s Talk
Presentation on UNSW Data Warehouse Project Covers strategy, business, technical and many other
aspects The thoughts and thinking on this have evolved over
several years Reflection on where elements of TOGAF can be
found and how the ADM process and methodology might have shaped and improved the outcomes
First, lets start with the presentation as is…………..
3
Overview
This presentation will examine the history and development of the UNSW data warehouse.
The presentation is less about the technical aspects and more about the business aspects of selling an enterprise solution to a broad and varied set of stakeholders, battling the NIH syndrome and what to do when the idea is actually sold.
Agenda
About the UNSW
Background on how UNSW implemented a SUN/SAS DW/BI solution Selling the idea and Gaining sponsorship Managing different stakeholders from different aspects of the business Juggling long term strategy versus immediate outcomes Different ways to tackle Project Reporting
Planning for the future Managing and developing the project as a product Developing processes and governance structures Development processes to manage non-IT developers (& IT too )
5
UNSW facts in brief
Member Go8 and Universities 21
Students (2006) 37,836
International Students (2006) 7,411
Staff (2006) 6,209
Degrees/Diplomas awarded (2005) 9,832
Total awarded 204,167
Alumni 182,401
Faculties 9
Schools 75
Centres 69
Institutes 6
Principal teaching hospitals 4
Residential colleges 8
University College (ADFA) 1
Undergraduate programs available 603
Postgraduate programs available 327
Kensington site: area (ha) 38
Permanent buildings 85
Field stations 5
Items in University Library 2.7m
Number of reporting solutions - unknown
6
UNSW facts in brief
But perhaps best know for its But perhaps best know for its
solar challenge car and football solar challenge car and football
playing robots.playing robots.
Definitions (we can skip these?)
What is a data warehouse?
A database geared towards the business intelligence requirements of an organisation. The data warehouse integrates data from the various operational systems and is typically loaded from these systems at regular intervals. Data warehouses contain historical information that enables analysis of business performance over time. (Webster)
A data warehouse is the main repository of an organization's historical data, its corporate memory. It contains the raw material for management's decision support system. The critical factor leading to the use of a data warehouse is that a data analyst can perform complex queries and analysis, such as data mining, on the information without slowing down the operational systems. (Wikipedia)
Some data modelling definitions Relational databases are optimised for online transaction
processing (OLTP). OLTP systems. For day-to-day operational needs and the database performance is tuned for many fast small transactions. can retrieve a small number of records quickly, but it can be slow if you
need to retrieve a large number of records and summarise data on the fly
data may not be consistent across the various OLTP systems access to the data can be complicated data organised around specific processes (such as order entry)
A dimensional database is designed and tuned to support the analysis of business trends and projections. This type of informational processing is known as online analytical processing (OLAP) or decision-support processing. data is “de-normalised” optimised for data retrieval and analysis new data loaded into the database is usually updated in batch and from
multiple sources a dimensional database tends to be subject oriented and aims to
answer questions such as, "What courses are most popular?" "At what time of year do courses sell best?" "In what regions are prospects and recruits weakest?"
International
Extr
act
, tr
ansf
orm
and load
Managem
ent
Report
sM
anagem
ent
Report
s
Busi
ness
Inte
lligence
Report
ing C
apabili
ty
“Operational” “Strategic”
“Base” Data(Raw, cleansed, exceptions)
Summary data
Data from systems“read only” access
ServiceManagement
Faculty(eg Medicine)
Research
Student
Others
HR
Finance
Data arranged in the DW differently. Arranged in subject areas and designed for
customisable and DIY reporting
(Examples to illustrate)virtual
History at UNSW
12
PeopleSoft@UNSW
UNSW was one of the first universities in Australia to deploy enterprise resource planning ERP software. One of the original business case items for PeopleSoft was it’s promised reporting capability.
A typical business case item for the myriad of ad hoc reporting solutions around the University since then……….
“is the lack of effective strategic reporting”, difficult access and inability to cross-link reports from PeopleSoft” :o)
Unlock the Information Assets 2003
• Extensive business analysis, ‘what if’ and judgmental content.
• Highly consolidated current and historical data.
Decision Support
• Extensive statistical and modelling content used for assessing trends and exceptions.
• Detailed and consolidated current, and usually requiring a large span of historical data.
Planning & Analysis
• Provides routine analysis & reporting.
• Consolidated current and historical data.
Monitoring & Control
Operational Support
• Provides day to day business operations and operations reporting.
• Highly detailed current data.
• Interactive operation
• Used for on-line processing with real time updates.
• Highly detailed current data.
• Time critical, highly interactive.
Transaction Processing
University Information Need Analysis, 2002
Unlock the Information Assets 2003
Enrolment Data Repository
1st Phase Data Warehouse
Executive Information System
Originally Planned Deployment Timeline 2003 2004 2005
Key Elements:
Improve data quality
Enforce “system of record” principles
Develop data warehouse infrastructure
Empower user community by deploying intuitive tools to analyse data
Triggered project “University Business Intelligence and Knowledge (UBIK)
Several earlier attempts at improving reporting and access to information (UBIK example)
Needs analysis conducted early 2005 with progressively inclusive interviews
A consistent set of problems emerged Problems linked to a set of key themes (and root causes) to
address Developed Model to address these themes and emerging
requirements
Unlock the Information Assets 2005
Typical Issues Access to information and getting effective reports or
information for decision support is a general issue Some complain that reports they need are non existent, not
known, hard to get, not always sufficiently accurate or up-to-date.
Many cite data being captured is not available or if it is they are unaware of what is there and how to get at it.
Part of this problem came from not assigning responsibility and ownership to data as there was no general policy on data management at that time.
There was no widely known general service that people could use to identify data, report problems and provide reports or even obtain lists of what is available.
Business information was and still is stored in multiple data stores with little linkages between them
Lots of stakeholders involved, each having their own requirements (and ways of stating them).
Selling the idea
Gaining common understanding and sponsorship
Seeking and gaining sponsorship
Operational decision making Are we currently on budget this month? Did this student enrol? Has the student paid their fee? Is a supervisor assigned to this student researcher?
Strategic decision making Does our new course structure attract more overseas students? What cities and countries need to be targeted more effectively? Was our last advertising campaign successful? How does our research performance compare to other Universities? Are we gaining more research grants now then five years ago?
Something in-between How many expressions of interest did we get from our last advertising campaign? Based on our current rate of enquiry, can we expect an increase or decrease in
International enrolments in four months time? What can we do to address this situation and generate more interest in time for VISA application and processing?
Is our DEST report accurate? Why don’t these reports match?
Selling the idea (1) – putting it in context
Recruitment Admissions Enrolment
ProcurementAccounts Receivable/Payable
GL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-award Ethical clearance Post AwardProject
Management
Da
ta D
oma
ins
Student
HR
Finance
Research
Facilities
Da
ta S
tore
s (r
ep
osito
ries
)
SecurityServices
Building Works
SafetyServicesHazard
Reporting
Fire and Hazard Prevention
Recruitment Admissions Enrolment
ProcurementAccounts Receivable/Payable
GL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-award Ethical clearance Post AwardProject
Management
Da
ta D
oma
ins
Student
HR
Finance
Research
Facilities
Da
ta S
tore
s (r
ep
osito
ries
)
SecurityServices
Building Works
SafetyServicesHazard
Reporting
Fire and Hazard Prevention
Selling the Idea (2) – process context
Selling the idea (3) - root cause analysis
Recruitment Admissions Enrolment
ProcurementAccounts
Receivable/PayableGL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-awardEthical
clearancePost Award
Project Management
STUDENT
HR
FINANCE
RESEARCH
Ca
mp
us
Co
mm
un
ities
Pe
op
leso
ft
Inte
rfa
ces
Operational Reporting Areas
Data capture opportunities Systems and interfaces
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service Current extracts
External sources
Additional Need
other other other otherOther
(eg Raisers Edge, Facilities)
Operational Reporting AreasOperational Reporting Areas
Operational Reporting Areas
Operational Reporting Areas Operational Reporting Areas
Recruitment Admissions Enrolment
ProcurementAccounts
Receivable/PayableGL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-awardEthical
clearancePost Award
Project Management
STUDENT
HR
FINANCE
RESEARCH
Ca
mp
us
Co
mm
un
ities
Pe
op
leso
ft
Inte
rfa
ces
Operational Reporting Areas
Data capture opportunities Systems and interfaces
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service Current extracts
External sources
Additional Need
other other other otherOther
(eg Raisers Edge, Facilities)
Operational Reporting AreasOperational Reporting Areas
Operational Reporting Areas
Operational Reporting Areas Operational Reporting Areas
Real versus intangible value?
The assessment of value of information needs to be made using different criteria to standard accounting metrics such as ROI and NPV.
Management Focus
Knowledge Capital
Value
Added Value
AutomationLow
High
Low High
Tangible, direct benefits
Intangible, indirect benefits
The IT Value ParadoxSource: Butler Group
Cost reduction - mainly through labour displacement. Dominant use of IT for last 30 years
Value Creation - barely understood because measurement is so difficult. Examples include new channels to market, enhanced services to trading partners, increased collaboration…
Knowledge Capital - there is a strong correlation between the information and knowledge orientation of an organisation and its market valuation
IT Value Domains
“The difficulty in dealing with the value created through investments in knowledge capital has caused organisations to shy-away from these types of investment. However, with the benefits of business process automation largely harvested by many organisations, the challenge will be to create new value propositions around IT investments which enhance knowledge Capital.”
“The degree to which benefits are tangible and also provide value are inversely related.” (Butler Group)
The IT value paradox:
Recall this slide…..
Recruitment Admissions Enrolment
ProcurementAccounts
Receivable/PayableGL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-awardEthical
clearancePost Award
Project Management
STUDENT
HR
FINANCE
RESEARCH
Ca
mp
us
Co
mm
un
ities
Pe
op
leso
ft
Inte
rfa
ces
Operational Reporting Areas
Data capture opportunities Systems and interfaces
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service Current extracts
External sources
Additional Need
other other other otherOther
(eg Raisers Edge, Facilities)
Operational Reporting AreasOperational Reporting Areas
Operational Reporting Areas
Operational Reporting Areas Operational Reporting Areas
Recruitment Admissions Enrolment
ProcurementAccounts
Receivable/PayableGL
Recruiting Induction Maintenance
Assessment
IncomePayments
Pre-awardEthical
clearancePost Award
Project Management
STUDENT
HR
FINANCE
RESEARCH
Ca
mp
us
Co
mm
un
ities
Pe
op
leso
ft
Inte
rfa
ces
Operational Reporting Areas
Data capture opportunities Systems and interfaces
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service Current extracts
External sources
Additional Need
other other other otherOther
(eg Raisers Edge, Facilities)
Operational Reporting AreasOperational Reporting Areas
Operational Reporting Areas
Operational Reporting Areas Operational Reporting Areas
Now Simplify to illustrate Target Environment
EIS
Executive Dashboard
Inte
rfac
es
Operational Reporting Support
Localised Support and Functions
•Canned
•Lists
•Extracts
•Service
•VALIDATION
External sources
En
ha
nce
d D
ata
Ca
ptu
re
Ex
tra
ct
Tra
ns
form
an
d L
oa
d (
ET
L)
Dat
a W
areh
ouse
Current extracts
Reporting Area
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service•
VALIDATION
Data
Mart
Data
Mart
Data
Mart
Data
Mart
DIY
•DIY
•web access
•download
Future extracts/reports
validation
OTHER
EIS
Executive Dashboard
RESEARCHRESEARCH
Inte
rfac
es
Operational Reporting Support
Localised Support and Functions
•Canned
•Lists
•Extracts
•Service
•VALIDATION
External sources
En
ha
nce
d D
ata
Ca
ptu
re
Ex
tra
ct
Tra
ns
form
an
d L
oa
d (
ET
L)
Dat
a W
areh
ouse
Current extracts
Reporting Area
Support and Functions
•Tools
•Canned
•Lists
•Extracts
•Service•
VALIDATION
Data
Mart
Data
Mart
Data
Mart
Data
Mart
DIY
•DIY
•web access
•download
Future extracts/reports
validation
OTHER
STUDENT
HR
FINANCE
STUDENT
HR
FINANCE
Funding and Business Case
Seed funding for proof of concept Use of Butler’s “value paradox” Risk and compliance Specific savings on targeted solutions
Student churn Savings through elimination of paper based
systems Others
Strategic funding – “reporting tax”
Implementation Decisions
Reviewing tools and vendors
All have their good and bad points - reviewed according to needs/gaps of current deployments – how progressed is the UNSW in 3 major areas? Data generation (ETL, data sourcing) Data management (DB, DM tools) Information access (presentation, analysis, data mining)
Balance between choosing individual “Best of breed” versus ease of integration through smaller numbers of vendor solutions (BOB versus solution suites) Warehouse generation tools - used in the design, cleansing,
transformation, loading, and administration of the data warehouse. Warehouse management tools - database management software
(DBMS) used to manage data in the data warehouse. Warehouse information access tools - used to enable end users to
access and analyse information stored in the data warehouse. Although important, the tool set is not the primary
consideration
Reviewing skill sets
Strong SUN/Solaris experience Less strong Windows Server (at that time - situation now
has changed significantly) Fragmented skills in VB, Pivot Tables, Business Objects,
Cognos and MSBI around the traps More ubiquitous skills in SPSS and SAS through
teaching (SAS), research (SAS & SPSS) and institutional reporting area (SAS)
After review and reports (Gartner, IDC, etc) we chose SAS We are also be able to leverage good students, post grads and
existing skill sets. Training still required
Sent both users and developers of the system on relevant training
Data integration
WebBrowser
MicrosoftOffice
HTTP / Java Servlet / W
ebdavScaleable Intelligent Server Analytical & Business Intelligence
ETL Studio Management Console
Enterprise Guide
Analytics Enterprise Guide
PeopleSoft
PeopleSoft
Access
ODBC / OLE DB
Oracle
Metadata Repository
Integration TechnologiesXMLSAS/SecureSAS 4GL SAS/STAT
LEVEL 3DataMarts
LEVEL 2DataWarehouse
LEVEL 1Staging/ ODS
Enterprise Data Warehouse (EDW) WebServer
Portal
Web Report Studio
IntrNet
Stored Processes
AppDevStudio
Information Maps
ApplicationServer
Stored ProcessesServer
Workspace Server
OLAP Server
MetadataServer
SPD Server
SAS Foundation SAS/ETS
Information Map Studio
WebBrowser
WebBrowser
ETL
Data Connectivity
DataMigration
DataQuality
DataSynchronization
DataFederation
SAS Enterprise Suite
Data generation environmen
t
Data management environment
Analytics, data mining, information
access
clients
Architecture
Client Tier
Middle Tier
ServerTier
SAS ETL StudioSAS OLAP Cube StudioSAS Management ConsoleSAS Information Map Studio
Web Infrastructure Kit Web Infrastructure Kit
Java Servlet Container
Java Servlet Container
webDAVServer
webDAVServer
HTTP ServerHTTP Server SDKSDK
SAS Enterprise GuideSAS Add-In for Microsoft Office
SAS Web Report StudioSAS Information Delivery Portal
Workspace Server Workspace Server
Stored Process Server Stored Process Server
OLAP Server OLAP Server
Metadata Server Metadata Server
SAS®9 Foundation SAS®9 Foundation
SAS/CONNECT Server SAS/CONNECT Server
Supports programming environments
Stores and manages meta data repositories that store data on servers, libraries,
stored processes and user policies
Stores and runs canned programs
Delivers data cubes (pre standardised tabulations)
to enterprise guide or OLAP clients
Super users
End user accessDevelopers
Reviewing Infrastructure
Most vendors have Windows and UNIX solutions (some only windows)
UNSW has strong local capability in UNIX environments especially SUN hardware and Solaris
Belief that UNIX is currently more scalable and fits well into the current managed environment
Architecture needs to handle future growth and the rate of growth not clearly understood
Looked at Solaris 10 containers/zones to create more flexible and scalable deployment
Chosen Infrastructure
SAS Application Tier Mid Tier
Client (Administration)
Client (Ad-Hoc Analysis)
Client (External Customer)
AUTHENTICATION DOMAIN- DefaultAuth
AUTHENTICATION DOMAIN- DefaultAuth
NON-SAS SOFTWARE- Tomcat 4.1.18- Xythos Web File Server 4.0.48- Xythos Darabase- Sun Java Development Kit 1.4.2_05
MS IE 6+PortalWRV
EG4.1WRS
SMCIMS
University of NSW: Sun Solaris SAS Enterprise Intelligence Platform
Sun Fire V490 (UltraSPARC IV)64-bit Solaris 104 CPU, (Dual Core)Memory: ??Mb
~ 20 External Users
~ 20 Internal Users
SAS 9.1.3 Foundation,Enterprise BI Server,Enterprise DI Server,Enterprise Miner,Platform Suite for SAS
Peoplesoft
Adhoc Flat Files~ 5 Internal Users
SAN
SAS Metadata Tier
AUTHENTICATION DOMAIN- DefaultAuth
SAS 9.1.3 Foundation
Zone 2MidTier
Server
Zone 1Application
Server
Zone 3MetaTier
Server
Running the Project
Implementation
and Project Reporting
The big vision versus immediate outcomes (maintaining momentum)
Set up project steering group and identified a number of “quick wins”
Used to quickly showcase capability QW1 – OLAP cube of student load QW2 – Budget report (with forecasting) QW3 – HR staff report made more friendly
Delivered on time and used for senior management presentations and demonstrations to further sell the idea
Used high level plans and different progress reporting styles for flexibility
Project reporting a new way
OperationalSystems Data
Feeds Data&
Applications
Reporting&
Analytics
AccessUser uptake
Project reporting a new way
Infrastructure
Goals/Objectives
ActualProgress
FunctionalityLayers
Outcomes
Showing Scope and Progress
External
HR
Student
Finance
Research
Facilities
Others
DATAGENERATION
DATAMANAGEMENT
INFORMATIONACCESS
CHANNELSLocal
o Unified development environmento 3GLs
o Data movement and replicationo Data management facilitieso Data feeds
o Data feeds
o Relational and dimensional datao Information maps
o Unified data managemento Data cleansingo Performance monitoring and optimisations
o Data security models
o Reportso End-user query, reporting and analysis
o Reportso Data miningo Packaged data marts
o Reportso Spatial information management.
o General authenticated access via webo Developer access via GUI tools
o Specific authenticated access via webo General and broad access via web (info kiosk)o Broader usage GUI tools & Office
o Specialist information services
Data quality process
ExceptionReports &Feedback
1
2
2+
2
The Project Report Card
System Outputs
Total number of reports and reports by data domainInformation maps and report themes (capability)
Organisational Contribution
Financial implication Time saved. Revenue generated. Costs reduced.
Process
Client Satisfaction (CSI) .Ease and convenience of use. Responsiveness to new requests and
requirements Security breaches and information leaks
People
Client Satisfaction (CSI) Data quality culture. More use of decision support data to make
decisions Learning organisation
Reports Capability Financial CSI Other??
Report Card Examples
Hybrid service modelData policy (DBOR, ROR)
Traceability matrix to root cause problems and associated
solutions with progress bar
PIR on business case items
User uptake and adoption.Numbers trained
Development Process Artifacts
High Level Data Acquisition Plans
ETL Flows
Data Schemas
InformationMaps
Rhubarb Rhubarb Rhubarb Rhubarb
Rhubarb Rhubarb
Reporting Requirements
Early screen shots
Achievements
ISSUES ACHIEVEMENTS
Information unavailable or unknown Self service and 40+ standard (user customisable) reports with descriptions and data dictionary export capability (also provided training to GMs etal)
Difficult to measure our real business performance
Produced VC report scoring our performance in key areas and in comparison to other universities.
Project Optimum (Research) Reports
Commencing AUQF and Phase 2 of Optimum (web based access)
Unable to track students/staff through their lifecycle
Created data feeds and reports linking admissions and enrolments
Just completed HR and Student data feeds and first reports
Now adding International data (for campaigns, conversion rates, predictive follow up) and IT service management. data for tracking SLAs
100’s of reports from different systems with errors and difficult to understand
Replaced with self service and 40+ canned reports from trusted data. Introduced data and text mining tools to glean more information from existing data. CATEI survey system saving $1M in annual expenses and being enhanced with text mining capability (automated trend analysis).
Unable to link data from different systems (eg HR, Student, Research, Finance, other) and external sources
Combining data in warehouse for any type of analysis. HR, Student, Medicine, International, Financial, ITSM data is being added. Reports have ability to span across different data domains – eg staff/student/researcher/lecturer. Live link to external UAC data provides 15 years of historical information
Unable to monitor our business performance
With strong internal/external data linkages trend and causal relationships can finally be monitored. We are now ready to provide executive dashboards for current and future defined KPIs and measures.
Early Achievements
Managing different stakeholders from different aspects of the business
Mistake number one Involving too many stakeholders during requirements
gathering and analysis phase == delays
Mistake number two Not involving enough stakeholders during requirements
gathering and analysis phase == turf battles
Major Lesson – understand your stakeholders!!!
Planning for the futureor….The BHAG
A truly integrated view of information.
Timely and relevant to inform and enable effective business decisions.
Meaningful measures of corporate performance and competitive business intelligence.
All aspects of key business lifecycle events are tracked, stored and optimised.
Guest Prospect Applicant Admitted Student Active Student Graduand
Recruitment units GradsStudent AdminAdmissions offices
Administration Com
mun
ityResearch S
tudent
Staff
UNSWsystems
Trustworthy information and reports that are contextual, complete and correct.
The Big Hairy Audacious Goal
Planning for the Future
More realistically it is managing the ongoing project as a product intended to satisfy new and emerging requirements
Planning for the future – “data maturity”(product view)
Optimisation What is the best that could happen?
INTELLIGENCE
Predictive Modelling What will happen?
(What if?)
KNOWLEDGE
Descriptive Modelling Why did it happen? INFORMATION
Integrated and cleansed data, standard and adhoc reports, OLAP
What happened? DATA
RO
I and
Bus
ines
s P
ower
Pro
gres
s
Prioritising Efforts
We can’t wait to complete each row before moving up Lose business interest Miss out on early benefits May never get past the first row!
Three step (VCP) process to define projects – value, complexity, priority Classify the Value of each initiative Classify the Complexity of each Prioritise the highest value lowest complexity
implementations first
Classify valueValueValue
Raw operational data
Cleansed, modelled, understood, integrated data
Data modelled for descriptive analysis
Data modelled for predictive analysis
Data modelled for what-if analysis and optimization
Management Cockpit
Corporate Performance Management Financial
reporting
Product Profitability
Human ResourcesResearch
Competitive Intelligence
Combine financial data with other domains
Link Admission and Enrolment
More detailed student data
Trends reporting
Balanced Scorecard
Analytical CRM
Campaign Management
Target Marketing
Customer Lifecycle Value
Budgeting and Forecasting
Enrolment Forecasting
Business Activity Monitoring
Production Planning
Activity Based Costing
Operational CRM
Data Integration MaturityData Integration Maturity
Complexity to Implement
Corporate Performance Management
Financial reporting
Product Profitability
Research
Combine financial data with other domains
More detailed student data
Trends reporting
Balanced Scorecard
Analytical CRM
Campaign Management
Target Marketing
Activity Based Costing
Operational CRM
Production Planning
Business Activity Monitoring
Competitive Intelligence
Budgeting and Forecasting
Customer Lifecycle Value
Management Cockpit
Enrolment Forecasting
Link Admission and Enrolment
Classify ComplexityB
usin
ess V
alu
e
Examples only!
Consider the styles of user
Farmers
Know what they want
Predictable Small queries Expect fast
responses Almost always
gets an answer Rarely find gold
Explorers
Have many ideas Unpredictable Large queries Do not always
get useful answers
Occasionally find gold
Miners
Generate hypotheses
Uses statistical methodologies
and models Seek unknown
patterns
Tourists
Have wide knowledge
Often use metadata
Internet capable ”Look things up”
“Building the data warehouse” Bill Inmon
BI Strategy: Development Plan
2007 2009 2009
Time
Create the “product” road map(reviewed annually)
Pro
jects
Financial reporting
Combine financial data
with other domains
More detailed student data
Campaign Management
Enrolment Forecasting
Link Admission and Enrolment
Trend reporting
Balanced Scorecard
Target Marketing
Competitive Intelligence
Budgeting and Forecasting
Depending on the situation, projects may be categorised against
data/problem domains or types of reports etc
Cust
omer
Life
cycl
e Va
lue
Management Elements
Program Steering Group Sets direction Manages Finances Review/approve all business cases
Reference Groups Represented by each major data (problem) domain Feed requirements up into PSG
Data Principles and Policy Identifies major data sources (DBORs) and owners Sets rules and approvals “Ombudsman” role to settle disputes Designated “official reports” (ROR)
Hybrid Service Model IT & Business - each support what they are good at Business provide incremental capability IT provide major capability jumps
Communications – report and celebrate wins regularly
61
Where to now?
Expanding capability for more users and data.Expand scope to lifecycle events such as cohort trackingIntegration with other reporting systems.Formalise governance structures.Continue to leverage skill sets
Original Architecture
SAS Application Tier Mid Tier
Client (Administration)
Client (Ad-Hoc Analysis)
Client (External Customer)
AUTHENTICATION DOMAIN- DefaultAuth
AUTHENTICATION DOMAIN- DefaultAuth
NON-SAS SOFTWARE- Tomcat 4.1.18- Xythos Web File Server 4.0.48- Xythos Darabase- Sun Java Development Kit 1.4.2_05
MS IE 6+PortalWRV
EG4.1WRS
SMCIMS
University of NSW: Sun Solaris SAS Enterprise Intelligence Platform
Sun Fire V490 (UltraSPARC IV)64-bit Solaris 104 CPU, (Dual Core)Memory: ??Mb
~ 20 External Users
~ 20 Internal Users
SAS 9.1.3 Foundation,Enterprise BI Server,Enterprise DI Server,Enterprise Miner,Platform Suite for SAS
Peoplesoft
Adhoc Flat Files~ 5 Internal Users
SAN
SAS Metadata Tier
AUTHENTICATION DOMAIN- DefaultAuth
SAS 9.1.3 Foundation
New Architecture Just Deployed
Immediate Improvements for 2008 Completed performance tuning exercise Bigger faster server and storage to handle the increase in
users and analytic applications SAS upgrade plus Oracle for fast static ad-hoc reporting Increased storage and own NSS copy (max 24 hours old) Additional resources to accelerate progress in reporting Structured lifecycle approach to create assignments for
students of varying skill sets and creating BI courses and qualifications – Institute of Analytics Professionals (www.iapa.org)
Projects to execute or assist Performance management Dashboards, scorecards Cohort and lifecycle tracking
Full operational handover
65
Lessons Learned
What we would do differently
Avoid over-thinking future needs – eg a small windows POC may well have produced similar “selling” outcomes to a POC built with the future in mind
Avoid overly fast tracking internal processes – only comes back to bite later (still need to make them do their part for buy in and ownership)
Communicate more about the project internally and externally (aka “best kept secret in UNSW”)
Ensure solution architectures agreed and approved with a greater number of stakeholders and there is buy-in where it matters
Allocate more resourcing to project management aspects
What we would do the sameUse SUN with SAS for production environment Build in-house skillsUse Hybrid Service Model
Sounds like a success story and in some ways it is BUT in some ways it isn’t Some stakeholders now trying to refocus it to operational and
reduce the vision even though others are won over Some stakeholders still don’t know about the project! Internal and external stakeholders still differing on requirements Constant re aligning and selling even now Resources stretched given enterprise wide view – “n x clients” IT received blame for business generated mistakes
Project longevity relying on a small number of key influential stakeholders rather than perceived as an Enterprise Imperative
Is there any TOGAF in here? Yes!
Phase P – Planning and Principles
Business Goals •Strategic•Operational•Regulatory
Drivers •Information available, monitored, track lifecycles, quality, integration
Overarching Principles•Information should be trustworthy•Information should be boundary less•Information should be easily available
Principles•User driven report generation and customisation•Centralised Dictionary environment•Centralised Business Performance measurement & Monitoring•Centralised tracking of Student / Staff data•Few trusted reports than many erroneous ones•Centralised Data Warehouse
Is there any TOGAF in here? Yes!
Phase A – Architecture Vision
ScopeLocal and External reporting – DW / BI Architecture – Stage 1HR, Student, Finance – DW / BI Architecture – Stage 2Research, Facilities, Others – DW /chitecture – Stage 3 Architecture VisionUser Self Service for reportsExecutive Dash BoardData Warehouse / Business Intelligence
Phase B – Business ArchitectureBusiness Scenario Workshops – Stake Holder Management, Buy In and Prioritisation of Concerns Business PrinciplesUser driven report generation and customisationCentralised Business Performance measurement & MonitoringCentralised tracking of Student / Staff data Business ArchitectureOrg Chart with new Business Processes if any
Is there any TOGAF in here? Yes!
Phase C – Information ArchitectureInformation Principles•Centralised Dictionary environment•Fewer trusted reports versus many erroneous reports•Centralised Data Warehouse
Target Information ArchitectureSlide number 22 : Target Environment
Phase D – Technology ArchitectureSlide no. 30 – Chosen Infrastructure
Phase E – Opportunities & SolutionsSolutions Architectures - SASSlide no. 28 – Architecture
Phase F – Migration PlanningSlide no. 35 – Showing Scope & Progress
Phase G - Implementation & GovernanceGovernance Structure to be set up
Phase I – Change ManagementChange management processes to be set up
1. Technical Definition
2. Data Modelling
3. Data Architecture
4. History
5. Issues
6. Selling the idea – whose idea?
7. Real versus Intangible value problem
8. Business Case Mechanisms
9. Tools and Vendors
10. Skill sets
11. Software technology
12. Infrastructure Project Management
13. Development Processes
14. Screen shots of the system
15. Achievements
16. Stakeholder Issues
17. Technical Definition
18. Data Modelling
19. Data Architecture
CONCLUSION – All good stuff but not always the best order and not always the right time = missed stakeholder engagement opportunities
TOGAF is in here – so what is the problem?
20. Finally - THE VISION!21. Strategy
22. Types of Users
23. Future work and techo and architecture stuff
24. Lessons Learned
The value TOGAF would have provided this project………
Focused on the vision earlier and kept this focus Tracked and aligned (changing) business requirements
continually throughout the process (centre piece of ADM) Able to use underlying principles to assist and ensure
people behaviour Share and common vision and SOW through ADM with
continual customer alignment would have reduced these problems
Helped keep the varying kinds of clients and needs by providing the common and agreed principles rules and guidelines by which the architecture and solution continuum is developed and sustained. 72
Conclusion
TOGAF helps you bring out the good work at the right time, working with the business with joint objectives and agreed deliverables
It provides the frameworks and support materials but works equally well with those you develop in house
Higher chance of success!
QUESTIONS?