Do different types of data require different levels of governance?
The need is growing for enterprises to leverage disparate internal and external data sources to satisfy growth strategies, reduce risk, and improve service effectiveness. Analytics power many of these prospects. At the same time, organizations appreciate the
fact that all sources have varying dimensions of quality, sourcing and integration characteristics, metadata, and security and privacy standards. The transaction data is usually governed the most as it directly impacts the business operational levers. But,
governing master and reference data reduces efforts and redundancies in the long run.
The answer to the question also depends on the definition of data and the view of value within the organization. Multiple levels from the below aspects can be used -
– Data View, Process view, Application view
- Data View - Data Domain, Dataset
- Process view – Process, activity
- Application view - Platform, application, system
– Master, Reference, Transaction data
As you progress, the problems put before the data governance council may become more routine, and some of the members may feel that they could be delegating the work or attending fewer meetings.
This is a common phenomenon that every organization faces with cross-functional communities that provide decisions on the opportunities or problems faced by projects, programs or Line of business. They feel most of it can be delegated by up skilling their
staff. The data governance head enforcing participation, tying accountabilities to rewards and performance management, having workshops that emphasize the importance of monetizing the value of data and the greater benefits to the organization should help in
these instances. A consistent performance management plan of the capabilities orchestrated across the enterprise along with the involvement of the board should assist in such scenarios.
How should you classify data to better govern it in the enterprise?
In order to understand your data better, the processes, people and technology (systems) that effect data need to be well stated across the information lifecycle. With a formal data governance program, you would want to define data in hierarchies of data
domains and datasets at a conceptual level.
The classification is a vital milestone to keep your organization talking in data terms, with the business and technology likewise. Data Governance is a bridge between business and IT. To have synergies flowing among divisions, one should rightly classify
data, with the perfect involvement along with accountability defined.
- Classify the data as enterprise and non-enterprise or layer the data into buckets based on risk and value impact.
- In order to assess if a data element is to be governed on priority, assess if the processes/activities leveraging the data element impacts the risk and value proposition of the enterprise. A data element can be leveraged by multiple processes that
span divisions. You would want to get your cross-functional community or stewards representing business divisions prioritize data elements in such instances.
If you want to trickle governance into every vein of the organization, it is suggestible to refer to all analysis activities and reporting in the sense of data. If you were to create a dashboard to showcase the progress of managing data quality within the
enterprise, your scorecard would show measurement against the integrity or validity of data. First, I suggest the creation of dashboard views that showcase data quality at a dataset or domain level and not from the views of applications or processes. But the
views of data across processes, applications, Lines of Business augment your efforts in better assessment and communication of the progress. One of the obstacles of finding resistance to change can dwindle as you progress along with these definitions of data.
What are the roadblocks that organizations face while attaining maturity in governance?
Most of the enterprises do not face roadblocks in kick-starting a formal data governance program as the board and stakeholders would be energized by the nuances of governance. Most pilot small with a newly appointed governance head leveraging technology
partners to orchestrate governance activities. This happens with a pace and often hits a grid lock, once the organization attains maturity in some of the dimensions say data quality, the governance division.
The next level of maturity demands business ownership of data, accountability defined within the enterprise, risk and rewards linked programs, skills development, and new solutions that offer required capabilities. A trade-off needs be made on spends vs
benefits on the required capabilities. To name a few next steps - stewards are recruited, partners are up skilled, data councils are formed and representation is shifted to business, solutions are acquired, the value of data is assessed.
Where should you focus most of the governance efforts?
You can’t direct your spends and efforts on all the dimensions of governance across all data leveraged by the organization.
| Always start with a strong business case while defining needs, problem statements along with benefits|.
It is ideal to assess where the firm stands in terms of maturity across dimensions including data quality, information privacy and security, metadata management, stewardship and risk management. Facilitate an organization-wide survey of perception along
with quantitative measures to assess governance maturity across each line of business. Conduct workshops with key stakeholders in business & IT and then facilitate a maturity assessment. Once the gaps are understood, prioritize the same, use industry models
DAMA, CMMI, IBM, data flux, Kalido and others to come up with the final maturity assessment. The data governance head or the CDO would have to ease this in the council or board discussions to rank the dimensions, data domains on which the organization would
want to act. Further, define data that is of high value/high risk to the enterprise and direct your efforts so.
Most of the firms today are focusing their efforts on data that directly impacts the risk than growth strategies that relate to the business value proposition.
How should you orchestrate the governance activities in the organization?
An ideal approach would be to define the strategic goals, perform a current capability assessment, audit resources, and define policies, processes, procedures and standards. Data quality can be established as a service, with service promotion, service usage,
and service improvement as major components. Orchestration of the processes would require a clear definition of the operating model. This would often necessitate having variants of self-service, full-service and assisted service models based on the availability
of capabilities within a line of business. One can orchestrate quality activities through business changes (projects) by “ensuring the data that is sourced by the program is valid follows integrity and consistency standards”. A big bang approach to quality
assessment and measurement is possible by leveraging the cross-functional community, stewards, and data owners to embark on data profiling and assessing the impact of quality while applying information in context.
What aspects does a data governance framework encompass?
A customized data governance framework covers
- An operating model that identifies data stewards, knowledge workers, Data owners, technology partners, council members and other affected stakeholders.
- Discovering and standardizing the processes and procedures
- Change control with defined stakeholder communication strategy
- Implementation roadmap with operational aspects and work breakdown structure
- Performance management plan to measure progress and report
* The views expressed in this blog are of my own and do not represent my company's in any nature.