SAP Data Quality

Developing a Data Quality
and Integration Strategy

Jonathan G. Geiger
Intelligent Solutions, Inc.
April 28, 2010
Sponsor
Topics
 Complexities
 Expectations Setting
 Assessment
 Improvement
 Strategy
3
Copyright © 2010 Intelligent Solutions, Inc., All Rights Reserved
Topics
 Complexities
 Typical Complexities
 Why They Exist
 Assessment
 Improvement
 Strategy
4
Data Integration
. . . and a miracle Operational

Data Store
Internal
happens here . . .
Data
Data
Warehouse
External
Data
Operational Data
Data integration is very complex!
5
Data Integration - ETL
Process
 Capture data from source-of-  Consolidate data
record Persistent  Translate to target semantics
 Filter operational detail Staging  Calculate derived and summary data
 Filter only changed data Area  Transport to target platform
 Ensure data Quality  Load target data warehouse
Integrate Operational
Data Store
Internal Capture Validate Load
Data Transform Data
Warehouse
External
Data
Operational Data
Transport
 Audits and control metrics

Metadata  Source of record mapping
 Quality metrics
 Performance metrics
6
Capture Complexities
 Data requirements
 Best source of the data
 Business rules for capturing the data
 Data meaning
 External data requirements
 History requirements
 Currency requirements
 Privacy and security requirements
 Audit and control requirements
 Metadata
7
Cleansing, Transformation &
Integration Complexities
 Data quality
 Data integration
 Data transformation
 Data enrichment
 Error handling
 Metadata
8
Load Complexities
 Currency
 Metadata
9
Why Complexities Exist
Data deficiencies are often

Problem Recognition not recognized
Responsibility Overlaps and gaps
Data is called an asset but

Discipline not managed as such
People are getting work

Benefit Recognition done`
10
Topics
 Complexities
 Quality Definition
 Governance and Stewardship
 Prioritization
 Assessment
 Improvement
 Strategy
11
Quality - Definition
 Quality is conformance to requirements

 Quality is not
.... (necessarily) zero defects
Defect Rate
Target
Time
12
Quality - Definition
 Quality is conformance to requirements
 Conformance to what?
 Whose requirements?
 How are requirements set?
 What degree of conformance?
13
Stewardship
 A steward is one who is called upon to exercise

responsible care over possessions entrusted to
him or her
 (adapted from Webster’s dictionary)
 The steward does not own the possessions
 The steward has a responsibility affecting the

processes that impact the possessions
14
Data Steward
 Exercise responsible care over the data

resources of the enterprise
 The steward does not own the data
 The steward impacts processes that affect the data
and its use
 Acquisition
 Management, maintenance and storage
 Dissemination
 Disposal
15
Data Planning Roles
 Stewardship responsibilities
 Provide input to the subject area model
 Provide input to the business data model (business
rules, definitions, etc.)
 Establish metadata management strategy
 Custodianship responsibilities
 Develop the subject are model
 Develop and maintain the business data model
 Establish metadata management strategy
16
Data Acquisition Roles
 Establish the quality expectations
 Define business processes for acquiring data
 Establish the authority levels for creating, updating,
and deleting data
 Establish the validation rules for ensuring that the
data meets quality expectations
 Understand the quality of existing data
 Provide input to the quality expectations
17
Data Management Roles
 Establish and monitor data demographic expectations
 Establish archival and disaster recovery rules
 Provide metadata content
 Transform the business data model into system and
technology models
 Establish (technical) data naming standards
 Manage metadata
 Manage data storage (design, reliability, security,
recoverability, archival and restoration, etc.)
18
Data Dissemination Roles
 Establish privacy and security policies
 Define standard query and reporting requirements
 Establish capability requirements
 Establish quality expectations
 Establish policies and guidelines for information use
 Provide metadata content
 Ensure adherence to privacy and security policies
 Providing input to the quality expectations
 Manage and provide metadata
19
Data Disposal Roles
 Establish retention rules
 Establish erasure rules
 Provide input to retention rules
 Provide input to erasure rules
20
Making It Real
 Two ways to approach stewardship

 Data stewards are assigned to a specific data subject
area – customer, product, order, etc.
 Data stewards are assigned to a particular function –
sales, marketing, finance, etc.
 There are benefits and drawbacks to each
approach
 In either case, good communication is
mandatory
21
Executive Oversight
 Cross-functional committee
 Provides authority to the data stewards
 Provide resources for data stewardship and

information management
 Resolve conflicts
22
Prioritization
 Too many data elements to do at once
 Need to categorize data

 Criticality
 Visibility
 Usage
 Sanctioned Projects
23
Topics
 Complexities
 Assessment
 Continuous Improvement
 Data Profiling
 Symptoms vs. Root Causes
 Improvement
 Strategy
24
Continuous Improvement
Process
Reactive actions Proactive programs

typically start here start here
ACT PLAN
CHECK DO
Some companies start

Data profiling typically
here, following existing
starts here
processes
25
Data Profiling Framework
DO
PLAN
CHECK
Data Cleanup and Business Process Adjustment

Framework courtesy of SAP ACT 26
Data Profiling in Context
 Diagnostic step to understand data meaning

and quality
 Priorities dictate scope
 Business data model provides business rules
 Quality expectations provide perspective
 Data profiling reveals conditions
 Analysis determines actions
 Expectations adjustment
 Corrective actions
 Preventive actions
27
Root Cause Analysis is
Performed
Major Cause Major Cause
# %
Characteristic
Major Cause Major Cause
B A D C OTHERS
28
Topics
 Complexities
 Assessment
 Improvement
 Data warehouse implications
 Upstream implications
 Strategy
29
Data Warehouse Implications
 Data handling options

 Accept
 Reject
 Fix
 Adopt default value
 Error handling options

 Suspend data awaiting correction
 Transmit correction to source
 Transmit need for corrections
30
Process

ACT PLAN
CHECK DO

starts here
processes
31
Topics
 Complexities
 Assessment
 Improvement
 Strategy
 Major components
32
Framework & Business Drivers
 Relate to Enterprise Quality Management

Approach
 Formal or informal
 Goals
 Understand business drivers and needs
 Business intelligence / operational systems
 Strategic / tactical / operational
 Declare strategy
 Mission statement
 Guiding principles
33
Methodology
34
Data Models
Subject Area Model
Business Data Model
Operational Data Warehouse

System Model System Model
Technology Models 35
Process

ACT PLAN
CHECK DO

starts here
processes
36
Data Profiling Framework
Data Cleanup and Business Process Adjustment

Framework courtesy of SAP
37
Roles and Responsibilities
 Executive Oversight
 Data Stewardship
 Data Custodianship
 Data Providers
 Data Users
38
Tools and Technology
 Database management system
 Data modeling
 Data profiling
 Metadata management
39
Data Quality Metrics
 Data usage
 Data quality improvement
 Benefits attained
40
Strategy Communication
DATA
QUALITY
STRATEGY
DATA
QUALITY
STRATEGY
41
Topics
 Complexities
 Assessment
 Improvement
 Strategy
42
About Intelligent Solutions
 Founded in 1992 by Claudia Imhoff
 Received outstanding recognition for client satisfaction
(Dun and Bradstreet survey of our clients)
 Internationally recognized industry expertise
 Full line of Corporate Information Factory and CRM
courses
 BI and CRM Consulting services
 Mentoring
 Assessment and Planning
 Management
 Design and Implementation
 International client base in all industry verticals
43
SAP Solution Overview
Enterprise Information Management
Kristin McMahon
Director, Enterprise Information Management
SAP
April 28, 2010
Poorly Managed Information
Leads to Inefficiency and Risk
“ Over 51% of organizations estimate data

related issues cost their company over
$5 million. Forbes Insight
“ 90% of all businesses still do not have

sufficient policies in place to meet data
governance regulations.”
IT Policy Compliance Group
Build an Information Driven Organization
Improve business insight Increase operational Meet compliance and

and decision making efficiency and reduce costs regulatory requirements
Provide all users with data that Provide high quality data to all Enhance information
is complete, accurate and business processes governance via policy-based
accessible data management
SAP Provides Best-In-Class EIM Solutions
Deliver Information That Is Complete, Accurate, and Accessible
Data Integration & Quality Management: Master Data Management:

 SAP BusinessObjects Data Services  SAP NetWeaver Master Data Management
 SAP BusinessObjects Data Federator  SAP Master Data Governance for Financials
 SAP BusinessObjects Text Analysis  SAP Data Maintenance by Vistex
 SAP BusinessObjects Data Insight
 SAP Data Migration services
Content & Information Lifecycle Management: Enterprise Data Warehousing:

 SAP NetWeaver Information Lifecycle Management  SAP NetWeaver Business Warehouse
 SAP Extended ECM by Open Text  SAP NetWeaver Business Warehouse Accelerator
 SAP Document Access by Open Text  SAP BusinessObjects Rapid Marts
 SAP Archiving by Open Text  SAP BusinessObjects Metadata Management
© SAP 2007 / Page 49

What Are the Sources of Bad Data Problems?
Employee
Data Entry
Purchased
or Rented Customer
External Self-Service
Data
Enterprise
Information
IT Data
Application Migration
Updates Projects
© SAP AG 2010 / 50
The Data Quality Framework
Continuous Monitoring
Data Assessment
CONTINUOUS
MONITORING
MEASURE
CONSOLIDATE
ANALYZE
Match &
Consolidate
MATCH YOUR DATA
PARSE
ENHANCE
STANDARDIZE
Enhance CORRECT
Data Cleansing
© SAP 2009 / Page 51

Data Quality Approach
The Three Rules of Data Quality
Rule 1: Analyze your data • What is the definition of “clean” data?

Profile, query, extract and in every other way • Who defines “clean”?
become intimately familiar with data content at a • Who owns it over time?
detail level. If you take a high-level approach to • Which entities have the most issues?
data quality, you will waste time discussing what • Where are the issues originating from?
the data might look like.
Rule 2: Define your scope • Which business processes are affected?

All data quality projects uncover hidden issues. • What business benefit can be achieved?
Be very clear about what is, and is not, relevant • How clean does it need to be?
to your current effort. • People, process, and tools?
Rule 3: Cleanse your data and track your • Define stakeholders to analyze and clean
results • Define processes to clean, monitor and
Data quality is not a one-time process. It is an maintain cleanliness
ongoing process of monitoring and correcting • Acquire necessary tools to assist
your data. You should know that: 1) new quality
needs are being met and 2) new business
processes are being monitored.
© SAP AG 2010 / 52
Business and IT collaboration through
visualizing information governance metrics
IT can easily share data quality

metrics to business users and
involve them in owning the data
problem
Business users can easily see how

their information measures up
against information governance
rules and standards
© SAP 2009 / Page 53

Building a Roadmap for Enterprise Information
Management is Key for Success
4. Data
GOVERNANCE
Technology enabling
people to implement a
repeatable process to
3. Data
manage the use, quality
CONSOLIDATION and lifecycle of
2. Data
Value
Consolidate diverse information

INTEGRATION & master data landscapes
CLEANSING and increase trust and Govern
reliability in information
Deliver trusted
information repeatable
1. Data and reliably at the right Consolidate Consolidate
READINESS form, to the right place at
Understand what data the right time
assets you have and Cleanse Cleanse Cleanse
how they are being
used
Understand Understand Understand Understand
People & Process Maturity

Data Quality Provides Value Throughout
Portfolio maps to the People and Process Maturity
Full Enterprise END-TO-END Data

COVERAGE Management
4. Data
Value
GOVERNANCE
3. Data MDG
CONSOLIDATION
2. Data
INTEGRATION & MDM MDM
CLEANSING
1. Data Dash Integrator Dash Integrator Dash Integrator

READINESS
Dash Quality Dash Quality Dash Quality Data Quality
People & Process Maturity
© SAP 2009 / Page 55

Why SAP? Time to Value: Fast and cost-
The Best Choice for EIM effective integration with existing
SAP and non-SAP systems
Proven Customer Value: Mature

offering and large install base of
customers supporting critical
business scenarios
Market Leadership: Analyst

recognition and customer
implementation success
Comprehensive Solutions for EIM

Strategy
One-stop for end-to-end information
governance and management
© SAP 2007 / Page 56
Questions??
Contact Information
If you have further questions or comments:
Jonathan G. Geiger, Intelligent Solutions

jggeiger@earthlink.net
Kristin McMahon, SAP

kristin.mcmahon@sap.com

SAP Data Quality

Uploaded by

Copyright:

Available Formats

SAP Data Quality

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SAP Data Quality

Uploaded by

Copyright:

Available Formats

Developing a Data Quality

and Integration Strategy

. . . and a miracle Operational

Data integration is very complex!

 Audits and control metrics

Data deficiencies are often

Responsibility Overlaps and gaps

Data is called an asset but

People are getting work

 Quality is conformance to requirements

 Quality is conformance to requirements

 How are requirements set?

 What degree of conformance?

 A steward is one who is called upon to exercise

 The steward does not own the possessions

 The steward has a responsibility affecting the

 Exercise responsible care over the data

 Two ways to approach stewardship

 Provides authority to the data stewards

 Provide resources for data stewardship and

 Too many data elements to do at once

 Need to categorize data

Reactive actions Proactive programs

Some companies start

Data Cleanup and Business Process Adjustment

 Diagnostic step to understand data meaning

Major Cause Major Cause

Major Cause Major Cause

 Data handling options

 Error handling options

Reactive actions Proactive programs

Some companies start

 Relate to Enterprise Quality Management

Subject Area Model

Business Data Model

Operational Data Warehouse

Reactive actions Proactive programs

Some companies start

Data Cleanup and Business Process Adjustment

 Database management system

 Data quality improvement

“ Over 51% of organizations estimate data

“ 90% of all businesses still do not have

Improve business insight Increase operational Meet compliance and

Data Integration & Quality Management: Master Data Management:

Content & Information Lifecycle Management: Enterprise Data Warehousing:

© SAP 2007 / Page 49

© SAP 2009 / Page 51

Rule 1: Analyze your data • What is the definition of “clean” data?

Rule 2: Define your scope • Which business processes are affected?

IT can easily share data quality

Business users can easily see how

© SAP 2009 / Page 53

Consolidate diverse information

People & Process Maturity

Full Enterprise END-TO-END Data

1. Data Dash Integrator Dash Integrator Dash Integrator

Dash Quality Dash Quality Dash Quality Data Quality

People & Process Maturity

© SAP 2009 / Page 55