Data Governance is the process of understanding, managing and making the critical data available with the goal to maximize its value and to ensure compliance.
InfoTrellis’ Data Governance Methodology follows a multi-phased iterative approach with 4 stages – Initiate, Define, Deploy and Optimize. This article is the second part of the Data Governance Methodology series by InfoTrellis. The first part of this series – Initiate your Data Governance – listed the essential foundations of successful Data Governance program.
‘Define’ stage primarily deals with defining effective Policies to address Data Governance issues. This article lists the important considerations of this stage.
Understand your Data Governance problem
Detailed investigation to understand the root cause of problem is essential to identify and solve Data Governance issues. For instance, a revenue amount discrepancy in financial report may look like a calculation error in first glance. Upon deep analysis, it could be revealed to be the result of interpreting the same business term, revenue, differently by different users which led to users applying different logic to arrive at the monthly figure.
Once we know the root cause of problem, it is important to categorize it. From our experience, categorizing a business problem into Data Domain Management, Business Process and Data Management Governance areas act as high level guides to understand the nature and scope of Data Governance problem. For example the revenue discrepancy problem mentioned above can be categorized into Finance data domain belonging to Accounting business process and Metadata Management Governance area. This helps to focus on the problem with the correct perspective.
Assemble the team to define Policies
Data Governance is a wide domain and requires varied skillset. For instance, Metadata management skills are different from Data Retention skills. Categorizing the business problem as mentioned above also helps in identifying the required skillset to resolve the issue. From our experience, a dynamic team composition based on the nature of business problem works the best. Typical members of this team are Data owners and Architect of pertinent IT/Business system, Business Data Stewards and Technical Data Stewards who understand the business domain and the mapped Data Governance area.
Define the Policies, Standards and Processes
Policy is generally a high level statement that describes how you would tackle issues or plan actions for the Data Governance area. For the revenue discrepancy problem, you can frame a policy that states – We must define all Business terms in Metadata Repository that can be accessed by all users of Business terms. Metadata Repository must map technical metadata, business rules and data lineage.
A Policy can be broken down into one or more Standards. For the policy mentioned above, you can have following Standards –
1. Business Glossary must be developed to maintain definition of all business terms.
2. Sensitive and Private data must be marked or categorized appropriately in Glossary.
3. Technical Metadata of data attributes in databases must be mapped to Business terms in Glossary.
A Standard could be broken down into one or more Processes. Typically Processes are implemented using an IT tool or program by IT implementation team. For the Standard 1 mentioned above, you can have the following processes –
1. For existing Business terms, import from Excel files into Glossary; for duplicate terms, resolve conflict and retain one instance of each unique term
2. For new Business terms, create the term and its definition in Glossary
3. Create Collections to group associated Business terms.
Select the tool – Some Policies would need an IT tool for implementation. The Enterprise Architect assigned to the Data Governance program can suggest the tool to be used based on Enterprise IT standards, existing tools in enterprise, future usage of tool, Data Governance” maturity of the tool and skillset of team. It is a best practice to select and get approval for the Data Governance Framework and Tools by IT office at enterprise level and make it a standard tool for addressing a specific domain of Data Governance Solutions. This ensures usage of uniform tools in enterprise to address common set of problems.
In conclusion, there are many variations to how teams would be setup and Policies would be defined. Keeping the above points in mind would help the enterprise to formulate the team with skillsets required to define effective Policies.
Stay tuned for Part 3 of this 4 part series on Data Governance from InfoTrellis. In the meanwhile, please send us a note with your queries and feedback.