In part one of this series, we defined data governance and looked at the missteps that lead to massive cleanup projects. In part two, we examined common data governance models and reviewed which work best for different types of organizations. In this post, we will walk through seven key steps to data governance.
Even if you are knowledgeable on the subject of data governance, knowing where to begin can still be a challenge. These steps will help set you on the right path towards an effective data governance framework:
- Establish Data Governance Organization
The first step is to evaluate various governance models and pick the one that best suits your organization. The role of data governance organization varies from one model to the other. However establishing ownership, establishing processes and procedures are some of the common to all the models. Here are some of the common responsibilities of Data Governance organizations
- Develop master data maintenance procedures
- Clarify rules, issues etc. with the business functions such as sales, purchase, finance
- Specify and develop tools for supporting the master data maintenance
- Support the daily business process execution in managing master data objects.
- The tasks of the master data governance organization can be completely operational or a mixture of operational and project based tasks with defined objectives.
- Identify Strategic Master Data Objects
Data governance certainly helps in improving the consistency of the data and keeps it in sync with the design of the system. However, it is not a good idea to govern each and every piece of data that is maintained. It is imperative to identify the data objects that need to be governed. Some of the key consideration for selection of data objects for governance
- Strategic importance to the company.
- The master data object is used globally across the organization.
- Large impact on the business if data objects not maintained correctly.
- Data complexity
- Maintenance of the master data object is not a core activity for any of the users of the master data object.
- Allocate Ownership
One of the primary reasons that result in bad data over a period of time is not having a defined ownership of specific data elements. One of the primary objectives of data governance is to eliminate this confusion by defining ownership of various aspects of governance.
The first step is to identify the ownership of various data elements at a global or local level. The strategic data objects and fields need to be owned by a global team and rest can be handled at a local level.
The next step is to identify ownership for
- Data fields – Ownership of data entry at a field level and
- User guide – Document the purpose and meaning of individual field values to avoid misinterpretation.
- Governance – Ownership to define and modify current field values
- Technical – Ownership to add/remove and update field values
- Identify Master Data Maintenance Rules
This is an essential step and probably takes the longest time. Data migration mapping rules if documented during the implementation can be an excellent starting point. Typically you need to document the following
- Field values – Rules for maintenance of field values spanning across various business scenarios and business units
- Organizational dependencies – When there are multiple business units, organizational units are involved, need to document which field values apply to which business unit and which do not apply.
- Data dependencies – Cross dependencies of data fields
- Use of Profiles (If automated tool is leveraged) – When an automated tool comes into picture, grouping a bunch of rules and making them as profiles can simplify the data maintenance and can drive consistency
- Establish Master Data Maintenance Procedures
Once the rules are documented, the next step is to build procedures that act as guides to the people who actually maintain the data. It is very important to build the procedures and to keep them updated based on current situation. The data governance team should own these procedures and keep them updated based on the inputs from the business. Typically procedure documents
- Who maintains data?
- When/How often?
- Based on what?
- Special requirements?
- Organizational differences?
- Functional differences?
- Field selection?
- Field values?
- Establish Tools for Master Data Maintenance
Building tools for maintenance and audit of data goes a long way in making sure the processed and procedures are being followed. The more difficult the maintenance process is, the higher the chance of not following it. It makes a lot of sense leveraging various tools for
- Maintenance of data
- Maintaining workflows for approvals and handoffs from one to another
- Mass changes and Mass uploads
- Periodic audits for the health check
There are various tools available in the market that can perform all these functions. SAP MDG, Itelligence it.mds, SAP Information steward that have built-in capabilities to automate various governance processes and ensure compliance.
- Establish Rules and Jobs for Master Data Archiving
While it is important to maintain the data correctly and catch the errors quickly, the governance strategy is not complete without defining an archiving strategy. This completes information life cycle and provides guidelines on when certain data elements need to retire. Various benefits of archiving include
- This helps in maintaining the system performance at an optimal level
- Reduce the database size and reduce maintenance costs while hosting and using in-memory database devices
- Simplify searches and lookups
Some of the key aspects that need to be defined are
- Which records to archive?
- Records marked for deletion
- Records, not used for xx months
- When and how often to archive?
- Where to save the archive files?
- For how long?
Once you have completed these steps, you digital core is ready for next level of data consistency which is foundation to the augmented intelligence.