DATA QUALITY &

THE DMBOK
DAMA BRASIL
SEPT 2014

DONNA BURBANK
VP, INFORMATION MANAGEMENT SERVICES

DONNA.BURBANK@ENTERPRISEARCHITECTS.COM
TWITTER: @DONNABURBANK

1 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4

› DMBOK Overview
Agenda
› Data quality management

› Benefits and impacts of data quality

› Activities relevant to data quality management

› Data quality management maturity assessment

2 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4

Introduction to Donna Burbank
› More than 20 years of experience in the areas of data management,
metadata management, and enterprise architecture.
» Currently VP of Information Management Services at Enterprise Architects
» Brand Strategy, Product Management, and Product Marketing roles at CA
Technologies and Embarcadero Technologies designing several of the leading
information management products in the market today
» Senior consultant for PLATINUM technology’s information management
consulting division in both the U.S. and Europe.
» Worked with dozens of Fortune 500 companies worldwide in the U.S., Latin
America, Europe, Asia, and Africa and speaks regularly at industry
conferences.
» President of DAMA Rocky Mountain Chapter
» Co-author of several books including:
⁃ Data Modeling for the Business
⁃ Data Modeling Made Simple with CA ERwin Data Modeler r8

Twitter: @donnaburbank
3 | MODULE 01 – COURSE INTRODUCTION | ENTERPRISE ARCHITEC TS © 2014

And When I’m Not Doing Data Management…
Pão de Açúcar, Rio de Janeiro

4 | MODULE 01 – COURSE INTRODUCTION | ENTERPRISE ARCHITEC TS © 2014

406 pages (also on CD & PDF) › Available from TechnicsPublications.com › Written and edited by DAMA members › An integrated primer: “definitive introduction” › Modeled after other BOK documents: » PMBOK (Project Management Body of Knowledge) » SWEBOK (Software Engineering Body of Knowledge) » BABOK (Business Analysis Body of Knowledge) » CITBOK (Canadian IT Body of Knowledge) 5 | MODULE 01 – COURSE INTRODUCTION | ENTERPRISE ARCHITEC TS © 2014 . What Is the DAMA-DMBOK Guide? › The DAMA Guide to the Data Management Body of Knowledge (DAMA-DMBOK Guide) › A book published by DAMA-I.com or Amazon.

deliverables and other common terminology. › To introduce widely adopted practices. build consensus and foster adoption for a generally accepted view of data management. › To identify “guiding principles”. methods and techniques. roles. › To provide standard definitions for data management functions. DAMA-DMBOK Guide Goals › To develop. Used with kind permission of DAMA-I 6 | MODULE 01 – COURSE INTRODUCTION | ENTERPRISE ARCHITEC TS © 2014 . without references to products and vendors. › To identify common organisational and cultural issues. › To guide readers to additional resources.

DATA ARCHITECTURE MANAGEMENT DATA DATA DAMA QUALITY DEVELOPMENT MANAGEMENT Framework DATABASE Functions META DATA MANAGEMENT OPERATIONS MANAGEMENT DATA GOVERNANCE DOCUMENT & DATA CONTENT SECURITY MANAGEMENT MANAGEMENT DATA WAREHOUSE & BUSINESS REFERENCE & INTELLIGENCE MASTER DATA MANAGEMENT MANAGEMENT 7 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

› Enterprise Data Modelling › Value Chain Analysis › Related Data Architecture DMBoK › › › › Specification Analysis Measurement DATA ARCHITECTURE › › › Analysis Data modelling Database Design Functions Improvement MANAGEMENT › Implementation DATA QUALITY DATA MANAGEMENT DEVELOPMENT › Architecture › Acquisition › Integration DATA › Recovery › Control DATABASE META DATA GOVERNANCE › Tuning › Delivery OPERATIONS › Retention MANAGEMENT › Strategy MANAGEMENT › Purging › Organisation & Roles › Policies & Standards › Issues DOCUMENT & CONTENT › Valuation DATA SECURITY MANAGEMENT MANAGEMENT › Acquisition & Storage › Standards › Backup & Recovery DATA › Classifications › Content Management WAREHOUSE REFERENCE & › Administration › Retrieval & BUSINESS MASTER DATA › Authentication › Retention INTELLIGENCE MANAGEMENT › Auditing MANAGEMENT › External Codes › Architecture › Internal Codes › Implementation › Customer Data › Training & Support › Product Data › Monitoring & Tuning › Dimension Management 8 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Preferences • Rituals. Heritage TECHNOLOGY ACTIVITIES • Phases. Beliefs. Tasks. Symbols. Steps • Tool Categories • Dependencies • Standards and Protocols • Sequence and Flow • Section Criteria • Use Case Scenarios • Learning Curves • Trigger Events GOALS & PRINCIPLES • Vision and Mission • Business Benefits • Strategic Goals • Specific Objectives Guiding Principles PRACTICES • DELIVERABLES & TECHNIQUES • Inputs and Outputs • Information • Recognized Best Practices • Documents • Common Approaches • Databases • Alternative Techniques • Other Resources ROLES & RESPONSIBILITIES • Individual Roles • Organizational Roles • Business and IT Roles • Qualifications and Skills Used with kind permission of DAMA-I 9 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Styles. ORGANIZATION & CULTURE Environmental • Critical Success Factors • Reporting Structures • Management Metrics • Values. Expectations Elements • Attitudes.

DAMA-DMBOK hardcopy version › 2011 – version 2 DAMA Dictionary of Data Management › 2011 Japanese version › 2012 Portuguese version › 2012 Chinese version › April 2012 –DAMA-DMBOK2 Framework › Q1 2015 – DMBoK2 Publication expected Used with kind permission of DAMA-I 10 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .Brief History of the DMBOK › First Publication of the DAMA Guide to the Data Management Body of Knowledge (DAMA-DMBoK Guide) 2009 › March 2010 .

Data Quality Management In Context EA’s Information Management Reference Architecture DQ READINESS & MATURITY INFORMATION MANAGEMENT READINESS ASSESSMENT 11 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

You may be able to drive for a long time with slowly policies. suppliers and business Management… partners. What is Data Quality Management? › Poor Data Quality Management is not equate to poor data quality “Data errors can cost a company › But when you don’t have good Data Quality millions of dollars. and make implementing new strategies difficult or even » The current level of data quality will be unknown impossible. alienate customers. poor data quality is like dirt a coherent set of on the windshield. but at some point you processes and either have to stop and clear the windshield or risk everything” supporting technology Ken Orr. The Cutter Consortium 12 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . standards. degrading vision. The very existence of an » Maintaining a sufficient level of data quality will organisation can be threatened by be a result of lots of hard work and extra effort poor data” from staff Joe Peppard – European School of Management and Technology » The risk to the business will increase › It is infinitely more sensible to ensure good data quality by having good management through “Ultimately.

13 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

from senior management and key data elements and DQM stakeholders down to users and operational • Identify key data quality areas to address and staff directives to the organisation around these key › Setting up a Data Quality Community of areas Interest can help create a common • Synchronise data elements used across the lines of understanding and provide a forum for sharing business and provide clear. use of value domains. unambiguous knowledge and best practice definitions. 1. Develop & Promote Data Quality Awareness Data Governance is Key Part of Your Job is Marketing! • Engage business partners who will work with the › Promoting and evangelising the importance data quality team and champion the DQM program of data quality as early as possible will • Identify data ownership roles and responsibilities. and data quality rules › Data Quality Management cannot survive without ownership and accountability. so close • Continuously report on the measured levels of data alignment with the Data Governance quality programme is essential • Introduce the concepts of data requirements analysis as part of the overall system development life cycle • Tie high quality data to individual performance 14 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 objectives . improve the chances of success of any Data including data governance board members and Quality programme data stewards › This needs to happen at all levels within the • Assign accountability and responsibility for critical organisation.

ANSWER: IT DEPENDS… 15 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Safety & • Effective Marketing and Security Advertising • Erratic Business Intelligence • Cost efficiencies • Amplified Corporate Risk • Improved Operational Efficiency & • Impaired Business Agility streamlining 16 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Benefit and Impact Good data quality benefit Poor data quality impact • Adherence to corporate & • Ineffectual Advertising & Regulatory acts Marketing • Improved confidence in Data • Reputational damage • Reduced “busy work” in data • Diminished Regulatory Compliance archaeology • Decrease in Customer Satisfaction • Enriched Customer Satisfaction • Uneconomical Business Processes • Better decision making • Compromised Health.

i. which are ultimately the • Specify the business rules that measure the source of many Data Quality requirements occurrence of data errors › Determining fitness for purpose requires • Provide a means for implementing measurement reporting on meaningful metrics associated processes that assess conformance to those with well-defined data quality dimensions.e. fitness for purpose • Identify key data components associated with business policies › The required level of Data Quality for a particular data component is therefore • Determine how identified data assertions affect dependent on the collection of business the business processes that interact with the component • Evaluate how data errors are categorized within › These in turn are driven by the underlying a set of data quality dimensions business policies. 2. Define Data Quality Requirements › Data Quality can only be considered within the Deriving Data Quality Requirements context of the intended use of the data . business rules 17 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

assess relative crime rates between counties. Data quality can only be considered within the context of the intended use of the data Data needs to be “fit for purpose” Data quality needs to be assessed on that basis 18 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . incident didn’t always correspond to the precise location of the crime. inflated home insurance premiums.police. for a number of reasons. the UK government In the context of creating Data fit launched a crime-mapping website for aggregated statistics to for purpose England and Wales (www. the data Unfortunately. quality is perfectly the postcode allocated to a specific police acceptable. How good does data quality need to be? Fitness for Purpose In February 2011. However. there is an issue for led many quiet residential streets to be the homeowners who receive incorrectly identified as crime hotspots.uk). if the same data is Data not fit The net result was that poor accuracy in used by an insurance for purpose the recording of geographical information company.

How good does data quality need to be? Fitness for Purpose Bad systems design can cost companies millions. it was discovered that the company had enough spare parts to last 90 years in some cases Data quality can only be considered within the context of the intended use of the data (data needs to be “fit for purpose” and data quality needs to be assessed on that basis) 19 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . However.65. so spare parts -. this is ridiculous.000 inventory items in all -. the system was hard to use. In theory. view of the cost of spare so each of the separate sites built up its for purpose parts. One pharmaceutical company had five main UK manufacturing centres. More than sufficient. own inventory of spare parts sufficient for its needs. in fact: After a data cleanup. the risk of machine downtime for purpose this is acceptable. with the holistic Data not fit But in reality.could be ordered from another location. all five sites shared a common system. each In the context of managing Data fit with its own warehouse of spare parts for the machines in the factories.

foreign keys in a RDMBS) › Timeliness– The time between when data is expected and when it is available for use. Rounding. › Referential Integrity. birthdate of 01/01/01 is valid.Data is current and “fresh”.g. Dimensions of Data Quality › Validity– Conforms to the syntax (format. number of significant digits in a number. › Privacy.Need for access control and usage monitoring.).Does the data align with operational context. floating point etc. for example. Business rules define what "100% complete" represents. the format (length. 20 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . › Currency. › Uniqueness– No value occurs more than once in the data set. maximum or contained within a set of allowable values).Constraints against duplication are in place (e. number of digits etc. but is it reasonable? › Completeness– Certain attributes always have assigned values.Level of detail of the data element. Data lifecycle is important here.) and range (minimum. › Accuracy– Data correctly describes the "real world" object or event being described. › Consistency– Values in one data set are consistent with values in another data set. Does it agree with an identified reference of correct information? › Reasonableness. e. › Precision. metadata or documentation rules as to the allowable types (string.g. e. can introduce errors. integer. type.g. Database. range) of its definition.

the CONSISTENCY UNIQUENESS represents. maximum or › Uniqueness– No thing will be recorded more than once based Data contained within a set of allowable values). The time the real world event being recorded › Consistency– The absence of occurred. Database. type. difference. COMPLETENESS › Completeness– The proportion of range) of its definition. when comparing two or VALIDITY more representations of a thing against a definition. upon how that thing is identified. The degree to which data › Timeliness– The degree to which ACCURACY TIMELINESS correctly describes the "real world" data represent reality from the object or event being described.). The Data item measured against Quality › Accuracy– The degree to which Dimensions itself or its counterpart in another data correctly describes the "real data set or database. › Validity– Data are valid if it conforms to the syntax (format. required point in time.) and range (minimum. floating point etc. world" object or event being described. Six Dimensions of Data Quality Many use a subset of these dimensions. The absence of difference. DAMA UK suggests six. number of digits etc. stored data against the potential of metadata or documentation rules as "100% complete" Business rules to the allowable types (string. define what "100% complete" integer. format (length. when comparing two or more representations of a thing against a definition Source: DAMA UK 21 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

archived. › Data › Data › Data deleted provision acquired validated. extracted. cleansed. for quality. (external). servers. › Governance › Conceptual. › Access controls analysed security etc (SOURCE DAMA) 22 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . The Information Lifecycle Information Lifecycle (DAMA) ARCHIVE CREATE & MAINTAIN PLAN SPECIFY ENABLE & PURGE ACQUIRE & USE RETRIEVE › IM strategy › Architecture › Install or › Data created. imported. edited. retained and › Define logical and networks. organised reported. converted. procedures modelling DBMSs migrated. retrieved policies and physical storage. reviewed. retention.

revealing hidden characteristics. Profile. a specialist DQ tool is normally employed › The result is an invaluable insight into the real operational data. but a “bottom up” view is crucial to identify existing issues within the data › This is achieved through an activity known as Data Quality Profiling › To conduct Data Quality Profiling as efficiently and repeatedly as possible. Analyse & Assess Data Quality Why Data Quality Profiling? › Reviewing and refining business policies provides a “top down” view of Data Quality requirements. patterns and anomalies 23 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . 3.

pattern count • Minimum. standard deviation. standard error COLUMN PROFILING • Completeness (% of non-null records) • Data type (defined v actual) • Primary key candidates • Count/percentage each distinct value FREQUENCY ANALYSIS • Count/percentage each distinct character pattern PRIMARY/FOREIGN • Candidate primary/foreign key relationships KEY ANALYSIS • Referential integrity checks between tables DUPLICATE ANALYSIS • Identification of potential duplicate records (with variable sensitivity) BUSINESS RULES • Using a preliminary set of business rules CONFORMANCE OUTLIER ANALYSIS • Identification of possible out of range values or anomalous records 24 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Typical Outputs of Data Quality Profiling • Record count. mean. mode. unique count. median. blank count. null count. maximum.

it’s difficult to define “how good is data quality” 25 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Metrics provide a common baseline › Without common metrics.

Data Governance structure with roles and 100%) represents the highest level of quality accountability for action • Ensure each DQI definition is complete and includes a › Controllability – Remedial actions are defined full description. I = Integrity. V = Validity.) › Business Relevance –Measures something of importance to the business • Wherever possible.” Guidelines for Data Quality Indicators Defined metrics should be used to assess data • Assign a unique identifier to each DQI quality using data quality indicators (DQI): • Use a consistent naming convention such as DQI-XNN › Measurability – Can be measured and where NN are two digits and X indicates the associated quantified within a discrete range Data Quality Measure (e. rationale. Define Data Quality Metrics “You cannot manage what you cannot measure. 4.g. the unit of measurement and permitted range › Trackability – Monitored over time to track progress 26 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . define each DQI as a percentage. etc. with the numerator/denominator clearly identified in › Acceptability – make sure it’s possible to define the derivation what “good” looks like • Set the polarity of each DQI such that the minimum › Accountability/Stewardship – Links to the value in the permitted range (e.g.g. 0%) represents the lowest level of quality and the maximum value (e.

Data Quality Metrics DIMENSIONS MEASURES INDICATORS VALIDITY DQI-V01 DQI-V02 ACCURACY INTEGRITY CREDIBILITY DQI-B01 Example TIMELINESS Measurement Framework DATA QUALITY CURRENCY PUNCTUALITY DQI-P01 COMPLETENESS DQI-C01 DQI-C02 DQI-C03 COVERAGE UNIQUENESS DQI-U01 27 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Define Data Quality Business Rules It’s important to make sure that data quality aligns Types of Data Quality Business Rules with the rules of the business. valid values/ranges for particular fields. etc • Format compliance › Target Rules that define the thresholds for Data • Mapping conformance Quality Indicators • Value presence & record completeness » e. notifying a data steward if a record fails a validation check. For example: • Value domain membership › Data Rules that define the precise characteristics that data needs to adhere to • Definitional conformance » e. relationships • Range conformance between fields/records. etc • Timeliness validation › Transformation Rules that define operations that should be applied to data TIP: Most DQ tools provide a rules repository so that rules can be created. automated correction of common data entry errors. 5. managed.g. and re-used consistently across the business standardisation of fields.g.g. shared » e. red-amber-green status • Consistency rules › Notification Rules that define alerts that should be • Accuracy verification fired under particular circumstances • Credibility verification » e. etc 28 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . alerting a data owner if data quality falls below a • Uniqueness verification defined threshold.g.

Test & Validate Data Quality Requirements Top Down › It’s essential that the Data Quality business rules are validated to ensure they accurately reflect the underlying Data Quality requirements › There are two complementary techniques: » Top down – formal review with business representatives to verify alignment with business expectations and ensure a common Business Rules understanding » Bottom up – inspection of exceptions occurring on real data to verify correct rule implementation › Once the Data Quality business rules have been validated. they can then be used to assess the baseline level of data quality Bottom Up 29 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . 6.

• The data quality dimensions associated with each data element conformant records are identified it’s important to establish a Data Quality Service Level • The expectations for quality for each data element for Agreement (SLA) each of the identified dimensions in each application or system in the value chain › This will set out business expectations for response and remediation and provide a • The methods for measuring against those expectations starting point for more proactive data quality • The acceptability threshold for each measurement improvement • The individual(s) to be notified in case the acceptability threshold is not met • The timelines and deadlines for expected resolution or remediation of the issue • The escalation strategy and possible rewards or penalties when deadlines are met/not met 30 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Set & Evaluate Data Quality Service Levels › The Data Quality Indicators and the business A typical Data Quality SLA should specify: rules upon which they are built are used to • The data elements covered by the agreement measure and monitor data quality • The business impacts associated with data flaws › However. in order to ensure timely resolution when thresholds are breached or non. 7.

Continuously Measure & Monitor Data Quality Effective Data Quality Monitoring is one of the 4 Key DQ Feedback Mechanisms most important aspects of DQM – a best practice capability will: • Exception Reports provide timely feedback to data stewards on the quality of data under their › Support a variety of feedback mechanisms. stewardship including interactive dashboards displaying up to date information on the level of data quality • Operational Dashboards provide data stewards. for critical data assets data owners and senior management with an interactive view of data quality within their area › Facilitate more detailed analysis to pinpoint the of responsibility underlying problem areas and support root cause analysis • Subject Area Summaries are published on a regular basis to highlight the level of data › Track changes in data quality over time to drive quality within a particular domain improvement and inform longer term data quality strategy • An Annual Data Quality Report brings together all of the data quality activities to provide a › Empower business users to take responsibility holistic assessment of data quality across the for data quality through the definition of rules enterprise and metrics › Transform existing ad-hoc data quality profiling and measurement activities into “business as usual” 31 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . 8.

Remember Activity #1: Develop & Promote Awareness 32 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Continuously Monitor & Measure Data Quality Monitoring & Reporting helps create awareness across the organisation.

classify and assign incidents as they are identified » Alerting Data Stewards to new incidents » Recording subsequent actions and outcomes from initial diagnosis through to final resolution » Handling incident escalation where SLAs have been breached » Providing management information such as statistics on issue frequency. common patterns. Manage Data Quality Issues › In order to expedite the resolution of data issues. root causes. 9. a means of recording and tracking those issues is required › A good Data Quality Incident Reporting System (IRS) provides this capability by: » Allowing users to log. supports data quality improvement initiatives and informs future Data Quality Strategy 33 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . time to fix and historical trends › The IRS helps assess performance against the SLA.

identity resolution and de- duplication) » Manual Correction – in some cases.g. even though a defect can be identified automatically. address standardisation or field substitution) » Directed Correction – less obvious defects can often be identified automatically but may require manual intervention to determine if the suggested fix is appropriate (e. the only way of resolving it is through manual inspection and correction (e. the existing data quality defect(s) need to be resolved through one of the following mechanisms: » Automated Correction – obvious defects which are well understood can often be identified and fixed by triggering an automated data cleansing routine. with no manual intervention (e. wherever possible. eliminated so that repeats of the incident will not occur › In addition to this.g.g.10. an invalid combination of fields where it’s not clear which field is at fault) › Data Quality tools use a scoring system to reflect the level of confidence in applying a correction – this can then be used to decide which defects should be corrected automatically (the cheapest and often the preferred option) and which should be flagged for directed or manual correction 34 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Clean & Correct Data Quality Defects › Detailed analysis of each data quality incident is vital to ensure that the root cause is identified and.

Data Cleansing Demystified The quiick fox jump’s over the the lazy dog 35 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Data Cleansing Demystified STANDARDISATION quiick foxjump’s The quiick fox overthe jump’sover the the the lazy lazydog dog 36 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Data Cleansing Demystified STANDARDISATION SUBSTITUTION The quick jump’s over the the lazy dog quiick fox jumps 37 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Data Cleansing Demystified STANDARDISATION SUBSTITUTION DE-DUPLICATION The quick fox jumps over the the lazy dog 38 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Data Cleansing Demystified STANDARDISATION SUBSTITUTION DE-DUPLICATION ENRICHMENT brown The quick fox jumps over the lazy dog 39 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

applying an appropriate remedy » Reporting .finding data quality issues) » Diagnosis and evaluation of remediation DEMING CYCLE (continuous improvement alternatives (i.e.monitoring ongoing performance 40 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . 11. investigating possible fixes) » Resolving the issue . Design & Implement Operational DQM Procedures › The Data Management Body of Knowledge identifies 4 key activities necessary for operationalising DQM: » Inspection and monitoring .

even the best DQM approach will need ongoing refinement to ensure it continues to serve the business as effectively as possible 41 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Monitor Operational DQM Procedures & Performance › Data Quality Management sets out good practice for ensuring data is fit for purpose – however.g. any DQM programme needs to demonstrate tangible long-term benefits › This can only be done through ongoing monitoring and evolution of the approach as the organisation matures in its management of Data Quality – this includes » Routine checking that SLAs are being met » Introducing new Data Quality Indicators as previously undiscovered DQ issues are identified » Extending the Measurement Framework to include further dimensions or measures (e. the Secondary Dimensions of Data Quality) » Developing new feedback mechanisms to satisfy the changing needs to users » Increasing the scope to include new datasets » Building in data quality at source. to succeed. 12. by improving the design of processes and systems › Just as good data will naturally degrade over time.

measurements Some basic measurements A comprehensive and A comprehensive and There is a standard set of of data quality are made on of validity and consistent set of business consistent set of business business rules defined for a routine basis and there is completeness are applied rules and data quality rules and data quality key datasets and these are no clear understanding to certain datasets. Data Stewards are supplied REPORTING No feedback is supplied Feedback on data quality Data Stewards are supplied Data Stewards are supplied with prompt feedback regarding specific issues or tends to be handled on an with prompt feedback with prompt feedback when exceptions occur. quality improvement. repository. data quality improvement. but there is no importance of data quality quality or the need for a and a common and takes a proactive common understanding and backs the drive for consistent approach. in a shared repository. satisfied. on the general level of data ad-hoc basis with no when business rules aren’t when exceptions occur and a regular schedule and on a quality. 42 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . quality. EA’s DQ Readiness & Maturity (based upon CMMI) LEVEL 1 – INITIAL LEVEL 2 – REPEATABLE LEVEL 3 – DEFINED LEVEL 4 – MANAGED LEVEL 5 – OPTIMISED There is good awareness of Everyone in the business There is some awareness of Everyone in the business DQ CULTURE There is limited awareness the importance of data recognises the crucial the importance of data recognises the crucial of the importance of data quality across the business importance of data quality quality. RESPONSIBILITY A Data Quality Champion A Data Quality Champion Data quality activities are Data quality activities tend Responsibility for data takes a lead role in ensures local adherence to handled in a reactive to be handled by the same quality activities is formally ensuring each business the DQ standards and manner with no assigned individuals. with regards to data Data Quality Community of issues. role. self-service basis. but indicators covering all indicators covering all applied whenever data is about the current level of these aren’t always applied datasets is stored in a local datasets is stored centrally received. but this isn’t a assigned through the area adopts good practice contributes to the wider responsibility for resolving formal requirement of their creation of Data Stewards. Interest. routine reporting. MEASUREMENT Few. understanding about the approach to driving data across the business. on a regular schedule. data quality. if any. key aspects. consistently.

DATA QUALITY Seven DQ Mistakes 1 2 3 4 Treating data quality Fixing data in a data Failing to consider the Confusing validity with management as a one-time warehouse rather than at intended use of the data accuracy activity source Data has to be fit for Validity is only the first step Quality data can only be Clean data for reporting purpose. no less towards accuracy ensured through a doesn’t solve the operational continuous cycle issues of poor DQ 5 6 7 Applying software quality Believing that good quality Blaming systems for bad principles to data quality data is the end goal data Data is infinitely more Deriving genuine value People and processes lie at volatile than SW and through information the heart of most DQ demands a different exploitation is the ultimate problems approach aim 43 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . no more.

If You’re Interested in Learning More www.enterprisearchitects.com/learning › EA offers a wide variety of training & consulting offerings including: » DAMA CDMP Certification Preparation » Applied Information Architecture Courses » Information Executive Overview » Business Architecture » TOGAF and Enterprise Architecture 44 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .

Burbank@enterprisearchitects.com Twitter: @donnaburbank 45 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 . Email: Donna.

Contact 46 DATA QUALITY AND THE DMBOK | ENTERPRISE ARCHITECTS © 201 4 .