Data quality sunz 2012

1

16 February 2012

Zeeman van der Merwe Manager: Information Strategy & Planning, ACC

[email protected]

DATA QUALITY: Getting Investment for a Weird Area

2

Why a weird area … ?

• Everyone talks about it

• Very few understand what it is about

• Lots of interpretation and confusion

• Mysticism ?

• How organisations understand & interpret data quality has bearing

• It is more about evolution

• This is a journey … that began in 2008

• With an assessment …

3

How did ACC Justify

Data Quality Tools?

4

State of BI in ACC - 2008

5

Background: ACC in 2008

• Inconsistent statistics for Ministerials

• Inconsistent data terminology

• Duplicate/inconsistent datasets

• Inconsistent business rules

• Unmanaged data in Data Warehouse

• Uncoordinated Data Analyst community

• Minimal standards & processes

Common Interpretation: Data Quality is bad

6

Intelligent Business Vision

7

• Oversees all aspects of data and information management

• Exercise authority and control (planning, monitoring, and enforcement) over the management of ACC’s data and information assets.

To be achieved by:

• Defining, and communicating strategies, policies, standards, architecture, procedures, and metrics relating to data and information

• Developing regulatory procedures for data and information

• Overseeing data management projects

• Providing governance and oversight of ACC data and information related issues

• Communicating the value of ACC’s data and information assets

Why Data Governance ?

Source: ACC Data Governance Terms of Reference (Version 1)

Data users have the right to know what the data means 1. The right to know the definition of the data.

2. The right to know where the data came from.

3. The right to know how the data was calculated or manipulated.

Data users have the right to know how risks to the data have

(or have not) been managed 4. The right to know what Security risks weren't eliminated.

5. The right to know what Quality risks weren't eliminated.

6. The right to know what Privacy risks weren't eliminated.

7. The right to know what Compliance requirements influenced data

processing and usage.

Data users have the right to know who made decisions about

managing the data, according to what rules 8. The right to know who made data-related decisions.

9. The right to know what decision-making checks-and-balances were in place.

10. The right to know how issues have been and will be resolved.

THE DATA USER’S

BILL OF RIGHTS

9

• Data Governance The execution and enforcement of authority over the management of data assets and the performance of data functions

• Data Stewardship The formalization of accountability for the management of data resources

Steward: Old English “Sty Ward”; “Keeper of the sty”

10

11

Data Management

Document & Content

Management

Data

Warehousing

& Business

Intelligence

Management

Reference &

Master Data

Management

Data

Security

Management

Data

Development

Meta Data

Management

Data

Quality

Management

Data

Architecture

Management

Database

Operations

Management

Data Governance

• Specification

• Analysis

• Measurement

• Improvement

• Enterprise Data Modelling

• Value Chain Analysis

• Related Data Architecture

• Architecture

• Integration

• Control

• Delivery

• Acquisition & Storage

• Backup & Recovery

• Content Mgmt

• Retrieval

• Retention

• Architecture

• Implementation

• Training & Support

• Monitoring & Tuning

• Analysis

• Data Modelling

• Database Design

• Implementation

• Acquisition

• Recovery

• Tuning

• Retention

• Purging

• Standards

• Classification

• Administration

• Authentication

• Auditing

• External Codes

• Internal Codes

• Customer Data

• Product Data

• Dimension Mgmt

• Strategy

• Organisation & Roles

• Policies & Standards

• Projects & Services

• Issues

• Valuation

12

Data Governance

Assessment

Assess maturity

• Executive awareness

• Data “ownership”

• Data “stewardship”

• Organisation structures

• Processes & procedures

Establish requirements

Develop

Data Governance

Strategy

Establish

Data Governance

Structures

Develop

Data Stewardship

Functions

Mandate

Data Governance

Strategy

Create

Data Excellence

Awareness

Data Governance

Committee

The Data Council

Define/Recommend

• Approach

• High Level Structures

• Roles/Responsibilities

• Change Management

• Budget

• Educate/Present

• Reporting line

• Agree strategy

• Get mandate to implement

• Identify Business & Technical Sponsor

Launch

Data Excellence

Educate

Organisation

“Recruit”

Data Stewards

Training &

/Mentoring

Identify members

• Subject Areas

• Responsibilities

• Identify/Appoint Implementation Team

• Appoint Principal Data Steward

Messages

• This is serious

• Long term

• Everyone responsible

By area

• Responsibility

• Impact on others

Change Management Plan

• Importance

• Commitment

Define/Agree

• Responsibilities

• Operations

• Sanctions

• Roles

• Areas

13

Data Governance Council

Data Stewardship

Steering Committee

Executive

Data Stewards

Coordinating

Data Stewards

Data Stewardship Teams Business

Data Stewards

Strategic

Tactical

Operational

< 5%

< 20%

< 80 - 85%

Data Governance Office

DMBOK Data Governance Issue

Resolution

14

History in ACC

• 2006 Information Governance Committee

– Failed: no terms of reference

• BI Strategy identified data governance

– CSF for Strategy

• Data Quality Effectiveness Review

• Data Quality Working Group

• Recommended Data Governance

– Received Mandate

– Formed Data Governance Committee

– Identified and appointed Data Stewards

– Formed Data Council

All because of data quality!

15

ACC Executive Mandate

• Defining & implementing a data governance framework across ACC

• Management of data

• Oversee the implementation of the ACC Business Intelligence strategy where it supports the aims and objectives of data governance

• Responsible for all data quality initiatives and resolution of data issues across ACC

• Incorporating existing initiatives where appropriate

16

Data Governance Structures

ACC BoardACC Board

Executive Management TeamExecutive Management Team

Data Governance CommitteeData Governance Committee

Data CouncilData Council

Data Quality

Coordination Group

Data Quality

Coordination Group“Data in Action”

Workgroup(s)

“Data in Action”

Workgroup(s)

Data Governance

Office

Data Governance

Office

Coordinates data management

initiatives across all projects

Resolves data related issues

Reviews and ratifies

recommendations

Overall responsibility for

data governance

Coordinates and supports

data governance in ACC

Data Related

Steering Committees

Data Related

Steering Committees

Governs data related projects

Chief Executive OfficerChief Executive Officer

17

ACC Data Quality Guiding Principles

1. ACC will manage data as a core organisational asset with decisions made based on value and the greater good of ACC and its stakeholders

2. The Data Council is mandated as the only forum for ratifying semantics, definitions and business rules for the use of data

3. Use industry and international data standards whenever, and as current, as possible

4. Data quality will be measured across the value chain and all data consumers will have a voice in specifying data quality service level agreements

5. Business process owners will agree to and abide by data quality Service Level Agreements

6. Data masters will be the primary source for any further use of that data

7. Validate data instances and data sets against defined business rules

8. Apply data corrections at the original source, if possible

9. Data entry and system integration will be automated whenever possible with validation applied on entry

10. System user interfaces will be designed to assist and encourage data quality

11. Reference data is current and relevant

12. All data entities will be semantically unique and defined

13. Metadata will be available to all data consumers

14. All report development will be peer reviewed

18

Data Quality Initiatives

Improving Client Information • Reducing number of client duplicates

created

–350,000 duplicates (60,000 P.A.)

–Staff were trained and Eos improved

–Rate reduced from 18% to 8%

• Improving NHI, Date of Death (DOD)

–Collaboration with MOH

– Improve valid NHI from 49% to 81%

–Provide DOD for 345,000 clients

19

RFI: Data Quality Tools

Data Quality Tools & Processes

• To be used for:

– Analysis of data quality

– Address cleansing, standardisation and validation

– Duplicate identification and elimination

– Monitoring and managing data quality

• First application Client address standardisation and verification

20

Data Quality Management is a process …

SOURCE: SAS/Dataflux

21

Evaluation Criteria (Gartner) 1. Connectivity/Adapters 2. Data Quality Assessment and Visualization 3. Parsing 4. Standardization and Cleansing 5. Matching/Relationship Identification 6. Monitoring 7. Subject Area — Specific Support 8. Address Validation/Geocoding 9. International Support 10.Data Quality Workflow 11.Enrichment 12.Metadata 13.Configuration Environment 14.Deployment Modes and Runtime Environment 15.Operations and Administration 16.Architecture and Integration 17.Service — Enablement Vendor/product image, relationship with ACC & Support Total Cost of Ownership

22

The Business Case

• Financial Savings

– NZ Post

– Rework

– Error recovery

• Reputational Risk Mitigation

– Minister/Public/Clients

• Process Support

– Data cleansing/verification

– Data matching/lookup (Master Data)

• Ease of development/reuse/monitoring

This is on paper …

23

Not on paper • Data governance was working

– Used as a ―sanctioning committee‖

• Data quality must be addressed

– Tools will help to do this

– We must have tools

• Confusion

– Data Quality vs Text mining

– What do the tools actually do?

– How do you use them (development)

– How can they help?

• Current Image: Anything to do with data can be fixed with the tools!

24

Summary

• Procurement highly influenced by:

– Non-paper justification

– Reputation of Data Governance

– Data quality is a ―big‖ problem

• Lots of confusion

– Needs lots of education/mitigation

– Evaluation panel needed lots of coaching

• The organisation was ―ready‖

• The business case … well …

25

Useful Sites

• Data Management Association (DAMA) www.dama.org

• The Data Administrator Newsletter www.tdan.com

• The Data Governance Institute (DGI) www.datagovernance.com