Data stewardship is one of the toughest components of data governance to get right. What is a data steward? What is it that stewards really do? What makes someone a good data steward? When these questions go unanswered, the result is often disenchantment with data stewardship. To help you keep stewardship’s (and the steward’s) good name intact, we’ve debunked seven of the most common misconceptions.

1. Data stewardship requires a data governance committee or council
   Stewardship can, in fact, exist without data governance councils or committees. The approaches, standards and rules established and identified by stewards are data governance and can thrive absent a committee or council. Remember, data stewardship is the formalized oversight and accountability of enterprise data assets. It establishes points of contact for data-related issues and aligns data changes with business strategy and objectives. As the data matures, so does data governance.
2. Data stewardship is a responsibility of IT
   The proximity of IT to the data sources, systems and management applications makes it an easy starting point for stewardship. However, this promotes stewardship as the sole responsibility of IT, when it is really a function of both business and IT. Knowledge and subject matter expertise about the business processes matter as much as systems expertise, which requires an ongoing and active dialog between business and IT. For instance, stewards sometimes have the direct authority to cleanse and update data, but more often, they work with their counterparts in IT to automate data correction.
   
3. Stewardship is synonymous with ownership
   Stewards, while responsible for overseeing and caring for the data on behalf of the enterprise, do not own the data. Ownership implies individual responsibility and little data sharing. In reality, data should be an enterprise-owned asset. Many of today’s data predicaments stem from rampant individual ownership and siloed data; both of which make accountability and oversight extremely difficult. This also muddles business processes that can diminish data quality and affect data interpretation.
4. Everyone is a data steward
   Serving as a steward in a clearly defined role or with a distinct set of responsibilities is different than stewarding the data on behalf of the entire organization. While everyone should take pride and responsibility in having reliable, accurate enterprise data, not everyone needs to be measured on specific improvements in data quality, reductions in defects or asked to enforce governance policies and standards. It's the same reason why not everyone is a manager although most activities are managed.
5. Data stewards are only responsible for fixing bad data
   Resigning data stewardship to data quality is self-limiting. Yes, data stewards identify potential data quality issues and help cleanse data, but larger scale problems may require changes to business processes or application modifications, which involves business and IT. To avoid downstream issues, data stewards should proactively work with business process and application development teams to identify data requirements.
6. Data stewards must be subject matter experts
   Data stewards should be knowledgeable about business objectives, vocabulary and key processes. Individuals with strong analytical skills, who are naturally inquisitive, can learn domain or subject area expertise over time. Stewards become experts in business areas by learning terminology and understanding the business rules in the underlying data. A steward may be a subject matter expert, but not every subject matter expert makes a good data steward. Stewards must be able to collaborate and build consensus. Communication skills are critical for anyone who wants to be a successful data steward.
7. Steward must be part of the job title
   When data stewardship works best, the role has been formalized. When the role is assumed, stewardship becomes a secondary or “as time permits” activity. This makes it difficult for stewards to have any real authority, creating dependency on existing relationships and limiting their ability to enforce standards and guidelines.

The real question is not if you need data stewardship. You do. It’s about clearly articulating the roles and responsibilities and tying stewardship to existing organizational functions and processes. Want to learn how stewardship teams are organized? Check out SAS Best Practices’ white paper, "The Five Models of Data Stewardship."

Video: Who are data stewards?