To get value from data, data stewards must understand and apply business requirements. When business ambiguity arises about how to best serve data stakeholders, data stewards need to know how to find out this information and with whom to speak. By doing so, stewards can align fit-for-purpose data with business needs and improve data quality. Data stewards understand business standards’ frameworks when taking good care of data assets.
Data governance, either formal or a non-invasive approach, reflects these structures and provides context and direction to these frameworks. When a data steward misunderstands the business framework and misapplies data governance, data quality suffers. Just as a martial arts practitioner in either Kung Fu, Karate, Capoeira, or Neo-Bartitsu needs to understand its concepts and context to best an opponent, data stewards should follow the rules and concepts making data fit for purpose.
In a rapidly changing marketplace, a company’s business specifications vary, especially around data quality. Objective data quality measurements and thresholds differ based on opportunities and threats. Within this reality, data stewards need to have flexibility with the business standards, while still keeping the data assets generated integral to the way the business proceeds.
How can a company find and retain data stewards who grasp its business requirements and make data fit for purpose while pushing for change as data quality needs to evolve? Four activities provide the basis for success:
- Describe data steward roles
- Find people best fit data stewardship roles
- Foster a collaborative culture respectful of thoughtful questions and expertise
- Further collective data literacy
Describe Data Steward Roles
A company needs to know what data stewards’ roles will do to meet business requirements and enhance data quality. In general, data stewards maintain data throughout its life cycle, according to business requirements and work with data owners. However, how they vary depends on a company’s data strategy and data governance implementations. One data stewardship role in one company does not necessarily translate to another.
For example, the Earth Science Information Partners (ESIP) wanted to foster collaboration between geoscientists using a FAIR framework (making data findable accessible, interoperable, and usable). Towards that end, ESIP needed data stewards to create data citation guidelines and uniform metrics that could be used, by scientists, to find information across multiple earth science data repositories. The data stewards need to have expertise in geoscience; they require technical and librarian-like knowledge to format and organize data and metadata.
Freddie Mac, on the other hand, aimed to achieve data stewardship credibility within a Ready (laying the groundwork), Set (identifying and working with stewards and stakeholders, and Go (do the data governance program as agreed) framework. Data stewards at Freddie Mac needed to establish and maintain effective relationships and connections with data owners, to consult on metrics making data usable and meeting these key performance indicators (KPI). A data steward at Freddie Mac needed to be an empathetic, business subject matter expert who understood the company culture and had excellent interpersonal skills.
ESIP and Freddie Mac’s data stewards overlapped in getting high data quality value to their users by cleaning that data and defining rules and policies, as part of the team. However, their specific goals and frameworks required very different data stewardship roles to meet stakeholder needs.
Find People That Best Fit Data Stewardship Roles
Given that data stewardship roles cover a few overarching abilities, within specific business frameworks, companies need to invest in finding the best people to steward the data. At a high level, a data steward needs to know the available data assets, how to define them, and how well they serve the business purpose. Often stewards multi-task between defensive data management (ensuring compliance with regulations) and competitive data management (leveraging business opportunities). But the depth and application of subject and information technology expertise differ in stewardship roles.
Donna Burbank from Global Data Strategy breaks data stewardship into two types: business and technical. Business data stewards manage data, of which they have subject matter expertise. They do daily data cleansing tasks and maintain data quality to align with data owners’ requirements. Technical data stewards have expertise in the data systems used. They collaborate with business data stewards to give technical support and have the digital and IT knowledge to automate some data quality tasks. Many data steward roles fit a spectrum of business and data engineering needs.
So, companies need to probe their people about their range of business and technical capabilities. Following up with previous examples, the data steward at ESIP leaned heavily towards having IT knowledge, especially around semantics and handling different geoscience ontologies. In contrast, data stewards at Freddie Mac took on a more substantial business role. A data steward who understands what makes good data quality assets lines up their skill sets as much as possible with the data stewardship roles.
Foster a Collaborative Culture
Setting up the right people for a company’s data steward role impacts data quality positively. But credibility towards the data steward’s understanding, definition, and work with business assets and demands need to be fostered through collaboration. The stewards’ expertise needs to be tapped and trusted by data owners, stakeholders, business analysts, customers, and others to provide metrics and validation to business requirements of high data quality.
For example, a bank’s business needs to upgrade international check data processing at an enterprise. Executives delegate IT to develop and update a data management system.
Months into the project, the system integrates currency exchange rates incorrectly, causing data quality issues to emerge in calculating deposits to and withdrawals from accounts. The executives want the data stewards, who report to IT, to identify and fix data quality for the business.
Would the accountants, and the business people, outside of IT, and who need to collaborate with the data stewards, trust their data quality recommendations? The answer, in many cases, would be no. Some IT programmers may not see the need to carefully understand business requirements because the data stewards work for the IT department, making the final data quality judgment. Even excellent data stewards could find a data quality misstep by IT would not be appreciated and be less likely to question adherence to business requirements.
Such an environment lends itself to mistrust between IT and business, pointing fingers at one another over data quality issues and misreading business requirements. Instead, data stewardship needs to function between both business and IT, fostering collaboration between the two, as stewards follow the data inputted and traveling downstream.
In a culture that fosters collaboration among stewards, IT, and business, processes start to become more productive by reducing a pain point or opening new revenue streams. Then the IT business relationship gains credibility. Both give and receive objective feedback through an executive mediator or authority. Data stewards see that leadership and teamwork are valued. They can adapt or clarify business requirements and push for better data quality, as business specifications change.
Further Collective Data Literacy Across the Enterprise
Good data stewards understand and apply data quality well when organizations further data literacy. Data stewards who have a high level of data literacy can read, work with, analyze, and argue with data. While fundamental graph reading and statistical skills help data stewards improve data quality performance, organizations need to go past these basics toward becoming more collectively data literate, to communicate and apply contextual knowledge to achieve the business and data strategy.
Enhancing collective data literacy means advancing the data steward’s understanding of the data‘s context, where the gaps lie in grasping and applying it to business needs. This collective data literacy cements understanding about needed stewardship roles, the best fit for the data stewardship work, and the collaboration required between business and IT departments. Data stewards demonstrate this collective data literacy by following through on data governance policies and procedures.
Collective data literacy provides, for example, a steward’s understanding of where a foreign exchange rate gets passed, who owns it, what to ask, and how to communicate it. They can efficiently find out from the department that sets the inter-bank rates and exchange rates, where their data goes, and how it gets maintained, instead of getting stuck within particular IT and business departments.
Collective data literacy provides data stewards resources to listen and respond well to business requirements, establish credibility within the enterprise, and foster collaboration among stakeholders. It does this through dialog that clarifies business standards and responds to changes in the marketplace. That way, data quality improves through data stewards as the business evolves.