The world’s most valuable resource is no longer oil, but data.The Economist, May 2017.
Data is the new commodity that everyone’s betting will drive the next phase of growth across industries. This being the case, organizations are defining new roles and responsibilities within their current setups to ensure effective data governance. One such key role is that of a data steward.
Who is a data steward?
A data steward is the go-to person in an organization for all things data. (S)he knows what and how data is being collected and maintained and is responsible for the following:
The DS is expected to define and implement processes and policies pertaining to the entire lifecycle of datasets, from inflow and storage to processing and transmission. The DS is also accountable for the documentation of these and the maintenance of a glossary of all relevant terminology.
The DS works with teams across the organization to define data quality metrics, and evaluate data based on these. Once gaps or discrepancies are identified, the DS must put in place corrective measures to implement controls and improve data quality. However, the DS does not actually correct the data; this is the responsibility of the data owner.
Data privacy and security
It is the responsibility of the DS to ensure the privacy and security of all data that belongs to the organization and meet compliance requirements, while at the same time, maintain reasonable data transparency. This also includes handling of a data breach or security compromise and defining the steps to be taken in case such a situation arises.
Data stakeholders other than stewards
It is important to note that a data steward is responsible for data whose ownership does not necessarily lie with him. Some other key data stakeholders in any organization include:
- Data producers: those whose activity results in the creation of data.
- Data owners: executives responsible for particular data sets over which they usually have administrative control.
- Data custodians: are responsible for data from a technical standpoint while data stewards are responsible for it from a business standpoint.
- Data users: those who need to access data, typically to analyze it and derive insights.
Different kinds of data stewards
Depending on the structure and needs of the organization and the required skill set and responsibilities, data stewardship roles can be of many kinds.
Domain data steward
This person manages all data and attributes of one data entity. For instance, customer data. By nature, this is a cross-functional role and requires a certain level of subject matter expertise.
Business data steward
This person manages all data (both transactional and reference) pertaining to a single function. For instance, sales and marketing or finance. One advantage of such a role is that setting evaluation metrics and measuring outcomes is easy because it revolves around a specific function.
Process data steward
This person takes care of all data involved in a particular process and is most often part of a process improvement team. It requires him/her to work across different domains and work with multiple domain and business stewards.
Technical data steward
This person is responsible for all the technical aspects of data creation, maintenance, storage, and movement. For instance, an enterprise data architect.
Data Stewardship: All or One?
There are two schools of thought surrounding data stewardship. On one hand, an increasing number of companies are defining Data Steward as an actual, unique position and hiring for it. Common requirements for the role include a strong background in data modeling and architecture, the ability to define data governance policies and educate various stakeholders on them, strong analytical skills, a good balance of business and technical thinking, and excellent communication skills.
At the same time, there are those who argue that data stewardship is not a separate position, but a relationship with data that everyone who handles data must have. This is based on a decentralized system of individual accountability for everyone from the data producer to the end user. Since these stakeholders are already handling and governing the data in many ways, all it takes to build accountability is to formalize it.
The sweet spot is probably somewhere midway: a system in which every data stakeholder comes with individual accountabilities depending on how they use or handle data. There will also be a data steward who has the larger responsibility of building processes, ensuring adherence to them, and continuously improving data quality.