Most of you will have heard of the term metadata but how can we leverage a metadata strategy to advance data quality management?
Anne Marie is an experienced Enterprise Information Management (EIM) consultant for corporations and consultancies. Accomplished data management strategist, published author, speaker, facilitator and instructor.
Data Quality Pro: By way of an introduction to our readers, could you briefly describe your background Anne Marie?
Anne Marie Smith: I am a data management professional with over 20 years experience in areas such as data governance and stewardship, metadata management, data warehousing requirements and planning, data modeling, and data quality management. I have a PhD in MIS and hold certifications as a CDMP (Certified Data Management Professional) and CBIP (Certified Business Intelligence Professional). Also, I am a member of DAMA International (data management professional organization) and a speaker / instructor in data management areas.
Data Quality Pro: You obviously consult with organisations regularly so do you think the term “metadata” is fully understood by the business community?
Anne Marie Smith: No, I don’t think the business community really understands the term “metadata” and the implications of missing / incorrect / poor quality business or technical metadata.
Data Quality Pro: How do you typically describe metadata to clients that you meet for the first time?
Anne Marie Smith: I don’t simply give the standard definition “data about data”. That tells people nothing.
I explain that metadata is the context and descriptions of the data (the type, what it means, where it is located, how it is used, etc.)
Data Quality Pro: How does technical and business metadata differ in your experience?
Anne Marie Smith: Technical metadata is used by IT staff to manage the data and to know if it is “correct” according to the technical specification (is it really in date format, is it updated according to schedule, etc.)
Business metadata is not well-understood since it is the business definition/context/usage/etc. of that data. Technical metadata is much easier to capture than business metadata and MOST organizations in my experience concentrate solely on the technical metadata if they pay any attention to metadata at all.
Data Quality Pro: What should be some of the core goals of a metadata strategy?
Anne Marie Smith: Core goals of a metadata strategy should be the development of an enterprise understanding of the intrinsic value of metadata across the organization, recognition of the value of metadata to improving data quality, understanding of the different types of metadata and choosing those categories in each type that are important to that organization, how the org will use the metadata and what technologies will support their metadata activities.
Data Quality Pro: How important is re-use in a metadata strategy and what tips can you offer for ensuring its adoption?
Anne Marie Smith: Re-use is essential – there is no need to constantly create the same metadata over and over.
Capturing it once, in a centrally controlled environment and managing it properly with data governance should reduce the need for re-creation dramatically.
Data Quality Pro: In your experience, how beneficial is a corporate data management strategy for sustaining a metadata strategy?
Anne Marie Smith: Essential – you can’t do a real metadata strategy without corporate data management.
If you look at the DAMA DMBOK Functional Framework and the EIM Institute’s EIM Framework you will see that both show the need for corporate data management as part of the foundation.
NOW, you don’t need an army to create a sustainable corporate data management strategy or program, but you do need the strategy if you want to have properly managed metadata and good data quality.
Data Quality Pro: What are some of the key challenges organisations typically face when launching their metadata strategy for the first time?
Anne Marie Smith: Doing it without expert assistance to learn and incorporate best practices (the right ones for their organization).
Trying to do too much at the start and also focusing on technology before identifying all the non-technical aspects like rationale, goals, components, etc.
Data Quality Pro: Can you briefly describe some of the typical “tools of the trade” that are required to deliver a metadata management initiative?
Anne Marie Smith: Repository (meta data database), access tools, data modeling tools (and the data models), ETL tools.
Data Quality Pro: Let’s talk specifically about data quality, how can good metadata management directly improve data quality? What are some of the dimensions of data quality that are affected for example?
Anne Marie Smith: Good metadata management can lead to good data quality since having and relying on the metadata can identify poor data / incorrect data / missing data.
Also, having good metadata shows an understanding of data management and shows that the organization is committed to good data – hence an improvement in data quality almost always follows.
Data Quality Pro: What importance and role does a metadata strategy have in terms of data governance?
Anne Marie Smith: Data governance is concerned with both the data and the metadata, so it is essential to a good metadata program to have a functioning and active data governance program.
Conversely, it is almost impossible to do a good data governance program without good metadata management, since again, data governance is heavily involved with metadata.
Data Quality Pro: As data governance and metadata management are intricately linked, what kind of activities would you expect the data stewards in this relationship to fulfil?
Anne Marie Smith: This can vary but the following activities are important:
- Interviewing subject matter experts and discovering needs for data, meta data and process
- Organizing complex information into appropriate categories and subjects
- “Translating” technical language into business language and vice versa
- Ensuring appropriate stakeholder involvement at all levels of effort
- Drafting clear and concise written documentation for users and technicians
- Working successfully with multidisciplinary teams
Data Quality Pro: Finally, what would you cite as some of the main components of a metadata strategy?
Anne Marie Smith: The main components of a meta data strategy could include:
- Organizational meanings of meta data and its role in the organization
- Business challenges and issues that can be addressed by improved meta data
- Approach to data governance
- Data stewardship roles for all main subject areas and key data
- Meta data usage guidelines
- Identifying sources of meta data
- Overview of process to determine the quality of the meta data sources (absolute, relative, historical, etc.)
- Methods to consolidate meta data from multiple sources
- Identifying where meta data will be stored
- Determining responsibility for proper use, quality control and meta data update procedures
- Establishing meta data standards and procedures
- Measuring the use and effectiveness of the meta data
Data Quality Pro: Finally, what are some of the key building blocks an organisation must put in place in order to convert a metadata strategy into benefits on the ground?
Anne Marie Smith: I would say the following key points are applicable:
- Recognizing the enterprise value (business and technical) of having managed metadata and demonstrate examples of that value – and the costs of not having had good metadata management
- Developing the right approach to metadata and identifying the appropriate best practices for metadata management to the organization
- Developing and sustaining a data governance and stewardship program that is part of the metadata management program
- Creating and sustaining a managed metadata environment for the organization (technology and business processes)
Summary of Recommendations for Data Quality Through a Metadata Strategy
- Business community have limited knowledge of the true meaning and benefits of metadata
- Most organizations focus on technical metadata, not business metadata (if they have any metadata strategy at all)
- Core goals of a metadata strategy are to increase understanding of metadata value across the enterprise, demonstrate the benefits on data quality, educate around the different types of metadata and how each category will be used and identifying what technologies are requried
- Getting started: Demonstrate the ROI and benefits to both tech/biz communities, develop a best-practice approach, implement data governance/stewardship, sustain a managed metadata environment
- Costs and efficiency can benefit from re-use using centralized environment and good data governance
- Properly managed metadata and data quality require a corporate data management strategy
- Key danger points are attempting to deliver too much too soon, focusing primarily on technology and not using the necessary skills to implement best-practices
- Tech requirements for metadata management: Repository (meta data database), access tools, data modeling tools (and the data models), ETL tools
- A metadata strategy helps data quality by identifying poor, incorrect or missing data plus a commitment to good data management (necessitated by metadata management) always results in better data quality
Next in the series:
In the next interview, Anne Marie talks about the technology aspects of metadata management, in particular she will discuss how organisations can implement their metadata repository.
Anne Marie Smith, Ph.D.
Anne Marie Smith is an Information Management professional with broad experience across a range of industries.
She has exceptional, demonstrated skills in enterprise information management (EIM) strategic consulting, business requirements gathering and analysis, data governance and stewardship, data architecture, data and process modeling, enterprise data management program development, master and reference data management, meta data management, data warehouse design and development, project management, and EIM methodology development and implementation.
She has strategic consulting experience with master data management (MDM) and data warehousing / business intelligence (DW/BI) engagements. She has developed data governance programs for several organizations and led the development of enterprise information management programs for corporations and consultancies.
- 25 years consulting and data management implementation experience
- PhD 2006, MIS, Northcentral University, MBA 1987, MIS, LaSalle University: BA 1977, Foreign Languages/Political Science, LaSalle University
- Published author on enterprise information management strategy development with regular online column at http://www.tdan.com
- Keynote speaker and frequent lecturer at industry conferences and events, including Enterprise Data World, MIT, and the Data Governance Conference
- Expert strategist and EIM thought leader