Data Governance
Establishes rules and policies to ensure reliable and effective governance of the organization’s data. These rules define processes and protocols to ensure the data asset’s usability, quality, and policy compliance.
Data Stewardship: Definition
Puts tactical roles and activities into effect to ensure adherence and support of the data governance plan. It includes assigning people to uphold the plan and developing a strategy for monitoring and maintenance of organization’s data.
Data Governance
Step :1
Assess the state of your data
- Before developing a governance and stewardship plan, you must gain a deep understanding of your organization’s current condition of data.
- Some key questions and research will clarify where you’ll need to focus your energy initially, as well as where you can improve in the future.
- It’s also important to understand how data is or is not managed, so you can help eliminate bad behaviors and promote best practices.
- A few simple questions will allow you to assess your organization’s current approach to data.
Try to find out the answer to the following questions,
- Who is using data?
- What are the business needs for data?
- Which data is used the most?
- How is the data being used?
Step: 2
Developing Data Governance Framework
- Data Definitions
- Quality Standards
- Roles & Ownership
- Responsible: owns the data
- Accountable: must sign off on or approve changes
- Consulted: can provide information and support
- Informed: needs to be notified but not consulted
- Security & Permissions
- Quality Control Process
Step : 3
Implementation of Data Governance
- Stewardship assignment
- Data Quality measures
- Data Standardization e.g., Guidelines
- Data enrichment e.g. Third-party system or appexchange
- Duplicate management e.g. Duplicate Rules
- Data cleansing e.g. Third-party system or appexchange
- Monitoring Data Quality e.g. Appexchange, Dashboards
- Security Controls – Profile/Permission Sets etc.
Measuring Data Quality
Review and segment data based on the following attributes that determine quality and usefulness:
- Age: Time elapsed since date last modified Completeness: Percentage of records that include key data fields
- Usage: Time since last used in reports or other applications
- Accuracy: Matches against a trusted source
- Consistency: Field formatting and spelling is the same across records (state or province, phone, country, and so on)
- Duplication: Amount of identical or very similar records that vary in quality
Data Policies
- Score data based upon data measuring quality
- Define must have and optional fields per object and implement validation rules.
- Use consistent formatting of the data, e.g., phone numbers, address, etc.
- Apply naming std. for fields and other metadata types
- Built-in security
Roles and Responsibilities
- Enterprise Level: Senior Leadership Team
- Support, Sponsorship, and Understanding of Data Governance
- Strategic Level: Data Governance Council :
- Approve things that need to be approved – i.e., Data Policy, Data Role Framework, methods, priorities, tools, etc.
- Tactical Level: Data Domain Stewards
- Act as the point communications person for distributing rules and regulations per domain of data to the operational stewards in their business unit
- Operational Level: Operational Data Stewards
- Producing, creating, updating, deleting, retiring, archiving the data that will be managed. – Data Producer
Track, Report & Learn
- Deploy appexchange products to measure data quality
- Use data enrichment appexchange products
- Create reports and dashboards within Salesforce to measure data quality
- Seek feedback from the various levels in the organization and refine the policies if needed
Data Governance Operating Models
Centralized Operating Model
- It relies on a single individual to make decisions and provide direction for the data governance program.
- Suitable for a large organization
Decentralized Operating Model
- There is no single Data Governance owner as everything is committee-based.
- Suitable for small-medium businesses
Hybrid\Federated Operating Model
- There is still a centralized structure that oversees the enterprise data level for which it has bottom-up input wide participation from the business units.
- The centralized structure provides a framework, tools, and best practices for the business units to follow. Still, in theory, it also provides the units with enough autonomy to manage business unit-specific data. It offers channels of influence to gather input for data sets impacting enterprise data or the other way around.
Data Stewardship in the context of Salesforce
Metadata Management
- Implement identification of redundant fields regularly.
- Remove unnecessary fields as part of technical debt
- Also, consider unused profiles, permission rules, or any other metadata which is not required
Data Management
- Use appexchange products to identity data quality levels
- Identify the source of incorrect data
- Use matching rules, duplicate rules, and duplicate jobs
- Use merge feature offered by Salesforce or use appexchange product for auto merging
Security
- Review data-sharing model in salesforce
References
- http://tdan.com/data-governance-roles-and-responsibilities/24774
- https://www.whitepapers.em360tech.com/wp-content/files_mf/white_paper/wp_iway_7steps.pdf
- https://a.sfdcstatic.com/content/dam/www/ocms-backup/assets/pdf/misc/data_Governance_Stewardship_ebook.pdf
- https://www.lightsondata.com/data-governance-operating-models-exposed/
