informatica world 2006 - mdm data quality
DESCRIPTION
Database Architechs is a database-focused consulting company for 17 years bringing you the most skilled and experienced data and database experts with a wide variety of service offering covering all database and data related aspects.TRANSCRIPT
Creating Data Quality Rigor for Your Core Data Categories
Paul BertucciEnterprise Data Architect
2
Agenda
• Initiative-based data strategy
• What must be done to execute on this strategy
• A data architecture to support you
• A data category example (Customer data)
• Making the strategy a way of life
• Q & A
3
Initiative-based data strategy
6
Initiative-based Data StrategyStrategy• Identifies key benefits• Seeks out alignment• Sets direction and priorities.
Initiatives• Short duration• Specific ROI• Incremental
Foundationalactivities• Mandated• Enterprise-wide• Ensure business
alignment• Focused on data
management and infrastructure.
. . .
DataStrategy Data Quality
Governance
Architecture
Implementations creating incremental value
Initiative-based Activities
Foundational Activities
7
What must be done to execute on the strategy
8
What Must Be Done
• Focusing on enabling key business initiatives with the data they need!
• Introduce data governance for all critical data (orders, product, employee…)
• Enhance/increase data quality across the board (rules, gates, process, tools)
• Move to a data services approach (highly sharing/leveraging data)
• Provide data SLA’s for data availability, integrity/quality,
9
What Must Be Done (cont’d)
• Eliminate data redundancies (across the board) -decrease P2P’s
• Put into play data integration capabilities to enable M & A, accelerate current systems consolidations (merger), and support other group or divisional data acquisition in line with the business speed
• Move to data hub concepts for key enterprise data (Customer, Product,..) and other enabling tools (e.g. HM) to elevate your ability to do Master Data Management
10
What Must Be Done (cont’d)
• Roll-out a data certification process for all data sources across the enterprise
• Create and maintain an enterprise reference view of data (reference layer) and leverage industry based models where ever possible (e.g. Party–based models)
• Protect/secure the data (security/privacy guidelines, roles, DR, backup, archiving)
11
A data architecture to support you
12
Strategic Data Architecture
META DATA
DataWHSE
CustomerDimensions
DWServices
MetaData Services
ODS Services
ERP 2
ERP 1
Web Services * Business Objects * Portal * Other
ERP – Next Acquisition
NextDW
Data Analytics
Mas
ter D
ata
Man
agem
ent
Applications/ComponentsTransactional System
s
MetaData
Other Transactional
ODS SFA
ODS ERP 2
ODS ERP 1
ODS CDH
NextODS
ODS
Extranet TransactionalNext Extranet System
xODS RC
Next xODS
ETL
REPL
REPL
EAIBackbone
Application Integration Services
Data Enrich
Hier Mgmt
Abstracted Services
Data Hub Services
Product/PricingData Hub
Customer Data Hub
NextData Hub
License KeyGeneration
MMDServices
Next Common Service
13
A data category example(Customer data)
14
Customer Data Strategy
ENTERPRISE CUSTOMER
MID-MARKET CUSTOMER
CONSUMER CUSTOMER
SMALL BUSINESS CUSTOMER PARTNER
BUSINESS CUSTOMER
Government Education SoftwareDevelopers
Other Segments
“Defining and setting how you will effectively identify, manage and leverage customers and their core attributes across all
segments to best serve the business now and into the future.”
TBD. . .
15
Some of the Problems (Symptoms)
Orders
Finance
SFA
CRM
Edu
ERP
DataWarehousing
FinanceReports
SalesReports
Partners
Customers
Direct
EDI
Partner Svcs
EduCRM ERP
Transaction Processing
Reporting
Leads
= “data hygiene/correction/reconciliation”
• Can’t recognize your customers completely (or not at all sometimes)
• Burn lots of energy/$ with duplicate data entry, consolidations, roll-ups & reporting . . . . and still don’t have good information.
• Must apply data hygiene, corrections, reconciliation in multiple places (not scalable, not consistently applied, out of control).
16
Some of the Problems (Symptoms) (cont’d)
Orders
Finance
SFA
CRM
Edu
ERP
DataWarehousing
FinanceReports
SalesReports
Partners
Customers
Direct
EDI
Partner Svcs
EduCRM ERP
Transaction Processing
Reporting
Leads
= “data hygiene/correction/reconciliation”
• Don’t share a common view of your customers/partners, and can’t provide one to THEM, even when they ask.
• Don’t know what customers own (licenses, maintenance, subscriptions), and can’t assess compliance, coverage or cross-sell opportunities.
• Struggle to append external information (enrichment)
• Have difficulty measuring sales effectiveness.
17
CONTACT LEAD OPPORTUNITY
PROSPECT
CUSTOMER
PARTNERCHANNELS CUSTOMER
PARTNER
How Do You Identify a Customer?
Account_ID, Email Address, Per_ID, Order_ID, Login ID, Name [+], Renewal_ID, others?
Contact ID, Party ID, Portal_ID,Company ID, Customer Nbr, DUNS Nbr, Name, Canonical ID,Support ID, others?
Partner Nbr, Party ID, others?
CONSUMER CUSTOMER
ENTERPRISE CUSTOMER
MID-MARKET CUSTOMER
SMALL BUSINESS CUSTOMER
BUSINESS CUSTOMER
PARTNER
CONSUMER CUSTOMER
18
Customer Data Dilemma
CustomerA
CustomerB
CustomerC
CustomerB
CustomerD (B)
CustomerE
CustomerB
CustomerF
CustomerG
ERPERP(M&A)
CRM
CustomerB
CustomerH
CustomerI
SFA
ERP ID Party ID Contact IDCRM ID
CustomerB
CustomerH
CustomerX
D&B(enrichment)
DUNS #
No strategy or consistency within a silo, or across silo’s
19
A Customer Data Strategy Should Provide:• Consistent customer identification & recognition
• A single, consistent technique for recognizing and enumerating customers (identification abstraction), sophisticated matching capabilities (Fuzzy, AKA’s, so on), de-duping, merging, etc…
• Model-driven (party-based models, so on)
• Customer relationships & hierarchies• Enables complex associations to our other customer data
(services, sales, opportunities, support, marketing, renewals, so on) to provide the needed 360-degree views of customer data
• Support multiple customer hierarchy views for different lines of business (Fin, Sales, …)
20
A Customer Data Strategy Should Provide: (cont’d)• Customer data enrichment (internal/external)
• Enables any critical data expansion or data enrichment from both internal systems (i.e. “customer segment classification”) and external sources (D&B, HH, Axxiom, so on)
• Customer data stewardship (reconcile/resolve/publish/ownership)• Group with sole customer data management
responsibility with appropriate counterparts out in each line of business (extended/federated model)
21
• Customer data quality/consistency/full life cycle management• Single “stable” approach to applying data standards, data
cleansing, data quality metrics measurement, auditing, and exceptions processing across the full life cycle for this core customer data
A Customer Data Strategy Should Provide: (cont’d)
22
Aligned With the Business• Supporting prospecting (lead, opportunity)• Supporting order quoting • Supporting order capture (all channels)• Supporting marketing campaigns• Supporting customer service/support • Supporting cross-sell/up-sell opportunities• Supporting customer loyalty programs• Supporting licensing/entitlements• Supporting renewal• Resolve financial reporting inconsistencies• Compliance evaluation/customer G2• Enabling 360-degree views that span different systems
Market Contact/Response Lead Opportunity Quote Order Fulfill Service Support
Marketing Sales Service Sales
Renewal
Fulfillment
23
Customer Data Management
Customer Data Across the Enterprise
Customer Support
MDM
Customer Order (ERP) Customer Sales (SFA) Customer
Customer Intelligence Partner Master Marketing Customer
• Customer ID• Customer Type• Initial Source• Primary Contact Details• Hierarchy Info (D&B)• Classification Details
• Customer ID• ERP Customer Number• ERP Cust Master Details
• Customer ID• SFA Customer Number• SFA Cust Mast Detail• Sales Classifications
• Customer ID• DUNS Info• Customer Profile Data
(Harte Hanke, D&B 1784, SFA Intelligence)
• Customer ID• Marketing Cust Details
LOCATION GROUP
SALES ENTITYPERSON
ORGANIZATION
EQUIVA-LENCY
NAMERELATION-
SHIP
CONTRACTROLE
IDENTIFIER
CONTACT METHOD
ADDRESS GROUPCONTACT METHOD
GROUP
ADDRESSROLELOCATION
ROLEIDENTIFIER
ROLELOCATIONPURPOSE
MACROROLE
GROUP
• Customer ID• Partner ID• Customer Profile Data
(Harte Hanke, D&B, SFA Intelligence)
• Customer ID• CS Customer Number• Titan Cust Master Details• Support Classifications• Support Entitlements
24
Making the strategy a way of life
25
Business (CDM)
IT/Data Architecture
Model-driven Customer Data Management
LOCATION GROUP
SALES ENTITYPERSON
ORGANIZATION
EQUIVA-LENCY
NAMERELATION-
SHIP
CONTRACTROLE
IDENTIFIER
CONTACT METHOD
ADDRESS GROUPCONTACT
METHOD GROUP
ADDRESSROLELOCATION
ROLEIDENTIFIER
ROLELOCATIONPURPOSE
MACROROLE
GROUP
Systems/Applications
Customer Model
Standardization
Market Contact/Response Lead Opportunity Quote Order Fulfill Service Support
Marketing Sales Service Sales
Renewal
Fulfillment
26
How the Strategy Becomes Reality
CRM
CommonParty-based
Model
ERP
Drives
Consistent with
Project-level Customer Models
LOCATION GROUP
SALES ENTITYPERSON
ORGANIZATION
EQUIVA-LENCY
NAMERELATION-
SHIP
CONTRACTROLE
IDENTIFIER
CONTACT METHOD
ADDRESS GROUPCONTACT
METHOD GROUP
ADDRESSROLELOCATION
ROLEIDENTIFIER
ROLELOCATIONPURPOSE
MACROROLE
GROUP
27
Movement to Data Hubs (MDM)
ERPCustomer DB
CRMCustomer DB
PartnerCustomer DB
ERP CRM Partner
FinanceReporting
Sales Reporting
Other
Customer Data Interactions
Data Quality
Integration Services
CustomerData Hub
ERP CRM
Partner Other
First up, a customer data hub
28
Data Hub Criteria (To Qualify)
Data that is created/updated/deleted in more than one place
Data that has a need to be highly consistent (across many sources)
Data that requires many views (e.g. 360 view of Customer)
Data/Attributes that must live on their own
Data that must be correlated with other sources (e.g. D&B)
Data that must be highly available
Data that must be readily accessible (high performance)
Data that must have the high integrity
Data that requires a formal change management process
Data that requires abstracted (enterprise) rules enforcementsuch as Global Customer ID's (canonical ID's).
29
BPELWS EAI ETL/EII
AnalyticsViews
Integration services
ODS
360 °CustomerTransaction
Views
CustomerData Hub
HistoricalAnalytics
Real TimeAnalytics
CustomerService
CustomerID Mgmt
CustomerLoyalty
Etc.
DW
Business Objects/Portal/ApplicationsDM
. . . .
Customer Data Hub
ERP DataAuditing
CustomerData Model
DataVersioning
Data Standardization
DataCleansing
DataPurge/Arch
Data Recognition
Data Enrichment
BusinessRules
CRM
SFA
SVC M&A
30
• ERP• Customer support• Services• Partner systems• Consulting services• Sales force automation• CRM• Contacts/leads• Data enrichment (D&B, Harte Hanks, …)
360-degree View of Customer
31
Customer Abstraction
Sales Entity
Role/Relationship
Specific Reference
1N
NN
“ERP System”
“342990667”
“100022” [“General Electric”]
“29903689”
“DUNS System”
“CRM System”
“118902”
• Provides the insulation from moving parts (“n” customer sources)• Provides a consistent representation to apply data rules, standards, and guidelines• Provides a strategic basis for tools or systems (Data Hubs, ERP, CRM, Reporting…)• Highly flexible for M & A and data leveraging (exposing customer views)
32
Summary
• Make sure you are aligned with what the business needs
• Go after one core data category first !
• Leverage industry tools/models if possible
• Establish a data quality paradigm/group
• Be initiative based with incremental value
33
Abstract
Trying to solve the data quality issues across multiple divisions, acquisitions, and user realms often leads to failure. Fundamental process and tooling can greatly reduce these failures across the board if they are focused on the primary (core) data categories of your business. Raising the quality of this core data has a ripple affect throughout the organization. In this session, you’ll learn how to identify what the data quality problems are, what needs to be fixed, what type of organization structure is needed, what type of data guidelines and data strategy must be present, and which tools of the trade you need to be successful in delivering all the benefits of high-quality data to your organization.