Data Stewardship (2014)

APPENDIX B Sample Training Plan Outlines

Training plans for Technical Data Stewards and project managers are presented.

Keywords

training, Technical Data Steward, project manager, Data Governance

Training Technical Data Stewards

- Data Governance:

Represents the enterprise in all things data and metadata:

– Metadata: definitions, derivations, and creation and usage business rules.

– Data quality: issues, fixes, data quality rules, and appropriate usage.

– Data-related policies and procedures.

– Champions data quality improvement and projects.

– Instigates methodology changes to ensure the capture of metadata and protect data quality.

Participants in Data Governance and Data Stewardship are accountable for the data quality and metadata.

Data Governance Board members are high-ranking individuals with decision-making capabilities.

Data Stewardship Council members are subject-matter experts who know the data well and can make recommendations on how data issues should be remediated.

- Business ownership and IT ownership:

Business owns the data:

– Business is responsible for data definitions and derivations, data quality, funding of data cleanup, and quality improvement.

– Business ownership is through the Data Governance and Data Stewardship effort. All business functions that own data (including IT) are represented.

– Strategy is to expose differences in data definitions and usage.

Business ownership:

– The business function that owns the data element also owns the metadata (definitions, derivations, data quality rules, and creation and usage business rules). These cannot be changed without the permission of the owner.

– Owner is determined by “who would care” if a change was made to the metadata. There is usually one business function of which the business would be significantly impacted if the meaning, allowed values or format, or quality of the data element changed.

IT owns the systems and applications:

– Implements information policies around metadata and data quality.

– Ensures that systems operate in ways that meet business requirements.

– Ensures that systems protect the data and the data quality to the extent they are able.

– Ensures that systems protect data integrity.

- Business Data Stewardship:

Business Data Stewards:

– Are designated by the Data Governance Board member for their business function.

– Have a role that is already in place and being designated a steward formalizes the role. That is, the Business Data Steward is usually someone recognized by their peers as having expert knowledge about the data.

– Are the authorities to whom questions about meaning and rules can be brought.

– Have access to data analysts who can help get questions answered.

– Can enforce their decisions.

– Approve proposed changes to the stewarded data elements. Data elements and metadata cannot be changed without the approval of the Business Data Steward.

– Get input from data analysts who work with the data every day and are directly impacted by the decisions. These analysts often know where there are data quality issues, or where a decision may negatively impact the data element’s quality.

Business Data Stewards are part of the Data Stewardship Council. This council:

– Reviews definitions, derivations, data quality rules, and creation and usage business rules, and determines the official version of each piece of metadata.

– Makes decisions about which business function owns the data elements.

– Works together to identify whether a “new” term is actually new or a duplicate.

– Supports project analysts.

– Reviews and proposes data-related process changes.

- Technical Data Stewardship:

Technical Data Stewards:

– Are responsible to the IT Data Governance Sponsor.

– Are a group of system-knowledgeable IT support personnel.

– Provide insight and expertise into how systems, ETL, storage (e.g., ODS), data warehouses and data marts, business intelligence, and code work.

– Answer technical questions about how the data “got the way it is.”

– Are responsible for the production of data through information systems to support business processes.

– Support impact analysis to understand the scope of proposed changes. This includes ongoing projects as well as ad-hoc inquiries.

– Work together to provide an understanding of the information chain.

– Are responsible for warning the business if proposed changes would cause bad things to happen to the application, system, or process.

There is a Technical Data Steward for each major system, application, and technical process (ETL).

More than one Technical Data Steward may exist for a data element if it passes through multiple systems or applications.

Technical Data Stewardship often formalizes and documents a role that is already in place.

IT owns the systems, applications, and processes that implement the business processes. These cannot be changed without the approval of the Technical Data Steward.

- Data quality, reporting, and business processes:

Impact on systems and applications when there is no Technical Data Stewardship.

Impact on data quality when there is no Technical Data Stewardship:

– Data quality deteriorates due to errors in system design.

– Data quality deteriorates when fields are overloaded.

– Difficult or impossible to determine root causes for data quality issues or to estimate the fix for these issues.

Impact on reporting when there is no Technical Data Stewardship:

– Systems produce data in ways that reporting systems cannot handle.

– Changes in source systems or ETL are not reviewed for reporting impacts.

Impact on business processes when there is no Technical Data Stewardship:

– New processes are proposed and even implemented that cannot be supported by existing systems and applications. This may cause data corruption and misuse of data fields.

– Business process owners must work with Technical Data Stewards to understand:

a. What processes and data are being proposed to be changed.

b. What business issues are being addressed and how important they are.

c. What is the proposed technical solution and what is the impact to downstream systems and processes.

- IT’s role in Data Governance and Data Stewardship:

Supporting Data Governance tools and applications (business glossary, metadata repository, data profiling, automated data cleansing, web portal).

Data custodians (backups, optimization, access permissions, physical implementation, data cleanup execution, data security implementation).

Change management:

– Notification of Business Data Stewards.

– Involvement of Technical Data Stewards.

– Impact analysis performed and communicated.

– Sign-off by Data Stewards on changes that negatively impact data quality.

IT support and sponsorship:

– IT sponsor represents and champions Data Governance and Data Stewardship within IT, as well as provides resources (Technical Data Stewards, tool support, and data custodians).

– The IT sponsor is a leader who is respected and listened to. The sponsor also has an organization that works closely with the data and data structures.

– The IT sponsor is directly responsible for delivering solutions to the business.

Training Project Managers

- Data Governance:

Represents the enterprise in all things data and metadata:

– Metadata: definitions, derivations, and creation and usage business rules.

– Data quality: issues, fixes, data quality rules, and appropriate usage.

– Data-related policies and procedures.

– Champions data quality improvement and projects.

– Instigates methodology changes to ensure the capture of metadata and protect data quality.

Participants in Data Governance and Data Stewardship are accountable for the data quality and metadata.

Data Governance Board members are high-ranking individuals with decision-making capabilities.

Data Stewardship Council members are subject-matter experts who know the data well and can make recommendations on how data issues should be remediated.

- Business ownership:

Business owns the data:

– Business is responsible for data definitions and derivations, data quality, funding of data cleanup, and quality improvement.

– Business ownership is through the Data Governance and Data Stewardship effort. All business functions that own data (including IT) are represented.

– Strategy is to expose differences in data definitions and usage.

Business ownership:

– The business function that owns the data element also owns the metadata (definitions, derivations, data quality rules, and creation and usage business rules). These cannot be changed without the permission of the owner.

– Owner is determined by “who would care” if a change was made to the metadata. There is usually one business function of which the business would be significantly impacted if the meaning, allowed values or format, or quality of the data element changed.

- Business Data Stewardship:

Business Data Stewards:

– Are designated by the Data Governance Board member for their business function.

– Have a role that is already in place and being designated a steward formalizes the role. That is, the Business Data Steward is usually someone recognized by their peers as having expert knowledge about the data.

– Are the authority to whom questions about meaning and rules can be brought.

– Have access to data analysts who can help get questions answered.

– Can enforce their decisions.

– Approves proposed changes to the stewarded data elements. Data elements and metadata cannot be changed without the approval of the Business Data Steward.

– Gets input from data analysts who work with the data every day and are directly impacted by the decisions. These analysts often know where there are data quality issues, or where a decision may negatively impact the data element’s quality.

Business Data Stewards are part of the Data Stewardship Council. This council:

– Reviews definitions, derivations, data quality rules, and creation and usage business rules, and determines the official version of each piece of metadata.

– Makes decisions about which business function owns the data elements.

– Works together to identify whether a “new” term is actually new or a duplicate.

– Supports project analysts.

– Reviews and proposes data-related process changes.

- Project Data Stewardship:

Project Data Stewards represent Data Stewardship on a project.

Responsible for:

– All Data Governance deliverables (definitions, derivations, data quality rules, profiling analysis).

– Documenting results in enterprise-wide business glossary and metadata repository.

– Validating the deliverables with the responsible Business Data Stewards.

Project Data Stewards are supplied and trained by Data Governance, but must be funded by the project.

- Value of having Data Governance and Data Stewardship on a project:

Collection of data definitions. Building a body of stewarded and understood data definitions benefits all who use the data. This is critical with conversions and migrations. The Project Data Steward can deliver the official definition to the project without involving project personnel in long discussions.

Collection of data derivations. This leads to a common way of calculating numbers. The project can deliver results that match the official calculation method. The Project Data Steward can deliver the appropriate derivation to the project without involving project personnel in long discussions.

Identification and resolution of data quality issues. Poor data quality can keep a project from going into production. The risk to a project is lessened by early identification and resolution of data quality issues.

Detection of poor-quality data. Data quality rules (which define what is meant by high-quality data) are identified and documented. The data is then inspected and compared against the data quality rules (a process called profiling) so that poor-quality data can be found early.

- Adjusting project methodology to allow for Data Governance deliverables:

Aligning Data Stewardship activity to the project life cycle.

Working with the Project Management Office funding function to budget for Data Stewardship support.

Initial evaluation of project for Data Governance deliverables:

– Project goals.

– Add, change, or delete data, and extent of that data.

– Potential for the project to reduce or corrupt data quality.

– How many systems will be impacted by the project?

– Evaluation results in an estimate of Data Governance effort required.

Initial tasks for project manager with Enterprise Data Steward:

– Add Data Governance tasks to project timeline (including data profiling).

– Establish timing for Data Governance resources (Project Data Steward, Enterprise Data Steward).

– Help project manager understand benefits of Data Governance.

– Determine project tasks that must involve Data Governance resources (model reviews, interface reviews, conversion mapping, data quality issue review, etc.).

– Get appropriate meeting invites.

– Plan for data profiling where needed (and it almost always is).

Data Governance tasks during requirements:

– Collection and documentation of data definitions and derivations.

– Gap analysis of data against business glossary.

– Definitions and derivations from business glossary provided to project.

– New data definitions and derivations provided by Business Data Stewards.

– Data and metadata approved by Business Data Stewards and recorded in business glossary.

Data Governance tasks during analysis and design:

– Collection and documentation of data quality rules.

– Measuring the extent of data quality issues (analysis of results of data profiling).

– Planning remediation and the impact of poor-quality data.

Data Governance tasks during quality assurance:

– Quality assurance test cases written and executed using data quality rules.

– Quality assurance test cases written and executed using definitions, including whether screens show data expected based on definitions, proper valid value sets, and multiple fields that display the same data.

Data Governance tasks during design:

– Evaluate proposed solution for potential to corrupt data quality. Issues can include repurposing existing fields, overloading fields, repurposing existing code values, lack of data integrity check fields, field contents changed to violate existing data quality rules, and field contents changed to a different granularity.