Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
An Engagement Model for Master Data Consumers
2015-‐05-‐21 David Loshin
Knowledge Integrity, Inc. loshin@knowledge-‐integrity.com
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350 1
What is Master Data? • Master data encompasses the models represenIng the core
business enIty objects used in the different applicaIons across the organizaIon, along with their associated metadata, aOributes, definiIons, semanIcs, roles, connecIons, and taxonomies.
• Examples include: – Customers – Products – Parts – Vendors – Employees – Suppliers – LocaIons
• EnIty concepts are idenIfied in relaIon to the business context
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
2
The Master Data Environment • A Master Data Environment provides a set of services
enabling data consumers with accessibility to a composite view of uniquely idenIfiable enIIes
• Requirements are solicited from data producers and data consumers for providing or using master data: – CollecIng source informaIon about enIIes – Resolving enIty idenIIes – Indexing enIty data – Establishing connecIons among source data records associated with
uniquely idenIfied enIIes – IngesIng data from the sources into the master data environment – Project planning for consuming applicaIons to integrate the
consumpIon of master data
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
3
Mistaken Perceptions of Master Data • Typical pitches for master data management use plaItudes to
moIvate adopIon: – “Golden record” – “360° view of the customer” – “Single source of truth”
• These concepts are somewhat misleading: – IntegraIon into a single “golden” record implies transformaIons that
may be inconsistent with operaIonal use – MDM, by its very nature, cannot be a “source” except under very
constrained circumstances – “Truth,” from a business process and applicaIon perspecIve, is
malleable – Dependencies on structure, intent, semanIcs, and context cannot be
ignored © 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
4
Multi-‐Domain Master Data Use
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
5
Human Resources Sales Customer
Support MarkeIng
Purchasing/Materials
Manufacturing Fulfillment
AccounIng/Finance
R & D
ForecasIng
IT
Business Func=ons
Customers Vendors Suppliers Employees Parts Products
Business En==es
Criteria for Usability
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
6
• Profiles capturing shared master aOributes • IdenIfying aOributes selected for indexing Model completeness
• Algorithms for similarity and matching • Methods of implementaIon IdenIty resoluIon
• Structural consistency • SemanIc consistency
Conformance consistency
• List of master data “events” • Inventory of services Master data services
• SolicitaIon of interest • Requirements, design, develop, deploy
Process of engagement
• Provision of master data • ConsumpIon of master data
Methods of integraIon
MDM Engagement Model
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
7
• Stages in transiIoning an opportunity into a producIon implementaIon for – Data consumers – Data producers
• EvaluaIon will help determine if there is a need for applicaIon data provision or applicaIon integraIon
Produc'onaliza'on, O & M
Tes'ng
Development Plan
Cos'ng Model
Data Governance
Facilitate Requirements Analysis
Ini'al Evalua'on
Iden'fy Opportunity
Initial Evaluation
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
8
• An iniIal discussion between the business process owners and the MDM team to assess suitability, consider scope, and clarify roles and responsibiliIes
Understand the Business Process
IdenIfy Master Domains
Specify Use Cases
Review Performance
Metrics
IniIal Scope DeterminaIon
• IntegraIon planning may indicate the business process incorporates data producers, data consumers, or both
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
9
Sample MDM Usage Scenarios Use Context Example
IdenIty resoluIon and registraIon
OperaIonal workflow ValidaIng that a customer aOempIng to enroll in a markeIng promoIon is eligible to parIcipate
IdenIty resoluIon and registraIon
AnalyIc workflow Linking payments to vendors to idenIfy where vendors have been paid more than once for providing the same item
IdenIty management as a service
OperaIonal or AnalyIc Providing a unique enIty idenIfier that links all employee records across benefit applicaIons
RelaIonship and hierarchy linkage
OperaIonal workflow Determining where physicians are associated with mulIple hospital organizaIons
RelaIonship and hierarchy linkage
AnalyIc workflow SupporIng reporIng of rolled-‐up metrics associated with regional sales by customer type
Data quality Informing synchronizaIon and quality within the master catalog
Publishing updated material item data as records are modified
Facilitating Requirements Analysis
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
10
• Consuming business process owners provide the requirements for using the MDM shared services: – Data requirements – FuncIonal requirements – Performance criteria – User acceptance criteria – TesIng plajorm and data
needs – Accessibility
• For publishers, the MDM team provides an Interface Control Document (ICD)
• The MDM team examines whether the exisIng integraIon paOerns are sufficient or if there is a need for customizaIon or addiIonal services
Requirements Analysis
AddiIons to Services Stack
CustomizaIon Needs
Auxiliary Processing
Data IntegraIon
Access paOerns
Business Rules
Performance ExpectaIons
Requirements Analysis: Access Patterns
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
11
• The Business Process owners evaluate how their applicaIons seek to use the master data
• The MDM team will share MDM usage paOerns to help drive the analysis process
• Examples include: – Direct queries via MDM tool interface (“inspector”) – Direct queries using SQL – Federated/Virtualized access – BI Tools – Web services – Extracts – Real Ime or batch?
Data Governance
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
12
• System of record – Have the systems of record been idenIfied? – Is the master data environment the system of record?
• Authorship and oversight – How are new items added to the MDM system?
• SynchronizaIon – How frequently is master data entered into the MDM system? – How frequently is the master data synchronized with the consumer
systems?
• Stewardship workflows – What happens when an issue with the source data needs to be
resolved?
• Responsibility for Data Quality
Costing Model
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
13
• Enterprise stakeholders subsidize the development of shared services
• Business data consumers benefit in the use of shared services via their investment
• Costs include: – Design and development
staffing – Dev/Test resources – AmorIzaIon of services costs – Chargeback models – O & M plan
• Currently: – Consumer costs require new
money – Producer costs may be covered
under exisIng budgets
$
Development cosIng
ConsumpIon cosIng
O & M Plan and cosIng
Development & Deployment • Detailed planning and execuIon of development tasks and
costs • Details regarding use of data integraIon tools, development,
implementaIon • Agreement to specificaIon and project plan • Resourcing & staffing
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
14
Testing
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
15
• Development and tesIng plajorms that scale with real applicaIon performance sizing
• Scalable data test plan (ensure accessibility or create data lab or generate test data)
• Simplify migraIon into producIon
Unit tests Full tests User acceptance ProducIon
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
16
Productionalization and O & M • ProducIonalize using standard deployment framework:
– Hardware plajorm – Database plajorm – IdenIty resoluIon tool/techniques – Data integraIon tools – Data virtualizaIon tools – Rendering engine – Business intelligence/reporIng
• ProducIonalizaIon – Move to designated producIon environments
• OperaIons & Maintenance: – General MDM OperaIons – General System Maintenance – Periodic Review and Revision of Business Rules
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
17
Performance Criteria and Measures Criterion Descrip'on of Measure
Persistence Volume The amount of data and number of enIIes that can be managed within the profile repository and index.
TransacIon Volume The number of master data transacIons per Ime period.
Index size The extent to which the index can grow.
User Load The number of simultaneous interacIve and system users.
Query Response Time The speed at which queries are saIsfied
Refresh Cadence The Ime period in which the master index is updated.
Ease of deployment How simple it is for users to adopt the use of the master data environment.
Time to value How quickly a prospecIve consumer can adopt the use of the master data environment.
Batch size The number of batch transacIons that must be saIsfied in a Imely manner.
Result set size The maximum number of records to be returned from a search.
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
18
Developing the Roadmap and Plan • Clearly explain the engagement cycle with the prospecIve
customer • Devise common deployment paOerns ahead of Ime to speed
Ime to value • Specify how funcIonal requirements address performance
expectaIons • Provide template cost model, interoperability model, and
integraIon plans for the prospecIve consumer
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
19
Developing the Roadmap and Plan • Clearly explain the engagement cycle with the prospecIve
customer • Devise common deployment paOerns ahead of Ime to speed
Ime to value • Specify how funcIonal requirements address performance
expectaIons • Provide template cost model, interoperability model, and
integraIon plans for the prospecIve consumer
Questions & Suggestions • www.knowledge-‐integrity.com • www.dataqualitybook.com • www.decisionworx.com • If you have quesIons, comments,
or suggesIons, please contact me David Loshin 301-‐754-‐6350 loshin@knowledge-‐integrity.com
© 2015 Knowledge Integrity, Inc loshin@knowledge-‐integrity.com (301) 754-‐6350
20