A recent article in Forbes (Trends of Note for CIOs) identified the top trends as follows:
- Establishing a culture of innovation
- Maturing data governance and maintenance
- Getting real value out of the cloud
- Staying on top of the evolving threats related to cybersecurity
- Containers and microservices as the driver to IT agility
Data governance took the second-place spot for a reason. The obvious reason is all of the attention around data privacy regulation such as GDPR and CCPA, and rightfully so. Data governance can help facilitate the inventorying and tagging of Personal Information through metadata management, and can help define risk, priorities and business requirements through the operating model.
Another reason data governance has become a ‘need to have’ rather than a ‘nice to have’ is that data proliferation continues to be on the rise. Data is created in so many ways and is no longer always consistently ingested, standardized and documented (if it ever was!). Individuals take advantage of self-service capabilities to pull data sets into their own silos as well. As a result, organizations find it difficult to know what data exists, where it resides, what it means, and how it’s being modified and leveraged.
The article also states: “Some people refer to data as the new oil of business. It is really the new crude oil. The question is how you turn it from a commodity into something refined that you can get value out of.” This is so true!
The three key learnings I’ve had about actually implementing data governance are:
- The initial assessment and roadmap is a necessity.
- Implementations fail all too often.
- If you’re in data governance ‘maintenance’ mode, then most likely you haven’t fully taken advantage of all data governance can provide.
The initial assessment and roadmap is a necessity
A data governance capability has to be right-sized for an organization, and ideally it is an enterprise wide approach. Understanding the current state of the various aspects that impact a successful data governance capability are important.
- Is there an overall data management strategy that can be aligned to?
- What are the key business imperatives a data governance capability needs to support?
- Has a governance structure been established, with stakeholders, charters, roles and responsibilities?
- Are there defined processes for establishing and resolving prioritization issues among stakeholders?
- Are there defined processes for issue escalation and resolution?
- Have data management and data privacy policies been defined, developed and validated?
- Is there a methodology to ensure compliance with data related policies, processes and standards across the data lifecycle?
- Do metrics exist to monitor data governance activities based on stakeholder criteria?
- If metadata is a part of the data governance program, are the people, processes and tools defined, understood and followed?
- If data quality is a part of the data governance program, are the people, processes and tools defined, understood and followed?
Implementations fail all too often
When I’ve been called in to consult on data governance, it’s frequently after 2, 3 or more attempts have failed. There is no single, right way to do this, and that’s one of the reasons it can be so hard to be successful and to make it stick. Data Governance needs to be put in place to solve specific data management needs, and it needs to be right sized for the organization. Just like agile, there is an iterative aspect to Data Governance that stakeholders need to realize and accept. One typical reason for failure is timing for Data Governance activities. For example:
- Getting stakeholders involved too soon without anything for them to actually do. A better approach is to ensure the operating model and initial processes have been defined, then start engaging stakeholders. It may be beneficial to start with a subset of stakeholders based on initial focus areas, then expand.
- Providing stakeholders with a role without any training for that role
- Installing data governance tools without processes and workflows designed
- Moving ahead with data governance without a charter, scope, objectives, or a supporting data governance office to ensure progress
- Not publishing training and communications to give the program substance
If you’re in data governance ‘maintenance’ mode, then most likely you haven’t fully taken advantage of all data governance can provide
There are so many ways to leverage a solid data governance operating model once it’s been developed and is working. Typically, an organization may initiate a data governance capability to ensure a business glossary and other metadata is developed and managed. Then, they may move into a data quality management program, leveraging the business data domain owners, business data stewards and technical data stewards that have already been identified. Subsequently, if processes were initially managed manually, tools may be implemented and configured to support metadata and/or data quality. After that, an organization may continue to leverage the data governance capability to ensure that policies are in place and monitored to support data lifecycle management or regulatory compliance. Overall, scope areas may include:
- Metadata management
- Data quality management
- Reference and master data management
- Data lifecycle management
- Data warehouse/data lake management
- Regulatory compliance
- BI & reporting management
- Data policy management
- Data architecture management
Each one of these areas can leverage a robust data governance operating model and especially the data governance office as well as data stewards to ensure requirements are defined, processes and standards are in place, best practices are followed, and the right people are involved at the right times.
Even though data governance can be somewhat fluid and iterative in its development, there are best practices and an overall, thoughtfully positioned roadmap that still need to be designed and followed to be successful, and to continue to add value to an organization.