As executive increasingly agree that data is a corporate asset, they are also funding data governance and data quality efforts more willingly. But philosophical alignment is often the easy part. Entrenched organizational behaviors are much more difficult to shift. Many companies have introduced the role of data steward before fully defining the role. In these cases, the beleaguered data stewards are doomed before they even begin.
Case in point: Greg, the director of BI at a health care provider, asked Baseline to participate in some core team meetings as he and his colleagues - a cross-functional group of managers responsible for launching data governance - were beginning a data stewardship effort. I made it to my client's meeting in time to find seven people huddled around Greg's laptop - and they were namedropping! Someone would mention a colleague's name and that would spark a fresh debate:
"Felisha? What does she know about financial data?"
"No, not Ben. He has no credibility with the management team. They'd kill him."
"Matt would be good. He's on the BI steering committee."
In its efforts to put data stewardship in place, the core team had fallen into the trap I like to call "vital signs," as in, "He's breathing. He has a pulse rate ... He could be a data steward!" But assigning the role of data steward is more complex than that - or at least it should be.
Old Habits Die Hard
All too often when planning a new initiative or assigning new roles, teams explore the "who" before understanding the "what" or the "how." In the process of working on the health care provider's data governance initiative, we'd broken the core team of some bad habits, one of which was too many meetings that led to too few decisions. To help them launch data governance, we guided the team through a methodical design process through which they formulated guiding principles, drafted decision rights and selected a high-profile launchpad project. They'd come a long way in setting up the program. But here they were trying to randomly assign warm bodies to a role that not only didn't exist, but needed a lot more clarification.
Many companies that adopted data stewardship early made the mistake of selecting staff members without much rigor or role definition. The newly appointed stewards would travel the corporate corridors, going from one meeting to another to represent data issues. Unfortunately, they didn't always deliver. As time passed, these data stewards became marginalized, and they would be excluded from milestone meetings and important decisions. "You're in this meeting, too?" a disingenuous manager once remarked to a newly minted (and ubiquitous) data steward.
Anytime constituents begin questioning the value of a new job role, its days are probably numbered. This turned out to be true for the new data steward, but we were determined not to let the same fate befall Greg and his core team.
Six Tips That Work
In Baseline's work launching data stewardship, we follow six tried-and-true techniques that have served our clients well. If you practice these, data stewardship will not only be accepted by your stakeholders, it will stick.
1. Be Deliberate About the Organization
There's no such thing as cookie-cutter data stewardship. Indeed, what data stewardship looks like depends a lot on the set of problems it's designed to address. While many insist that data stewardship should be a business function, if the problem is the quality of the data on the operational systems, the business might not have the right skills. Conversely, as IT educates the business on the value of data stewardship, it's important that the role is seen as a bridge between IT and the business, resolving data conflicts and inaccuracies, and deploying information in a structured and consistent way as the business moves forward.
Whether your company introduces data stewardship in the business or in IT, pinpointing the pain will help you identify the best organizational fit.
"Historically, we've done everything by line of business," says Todd Okuley, director of customer data governance for Nationwide Insurance. "In the past, data stewards were tied to individual organizations, and they could lose focus." As part of a new data governance effort targeted to customer data, Okuley and his team assigned data stewards according to the company's customer master data types: profiles, preferences, relationships and interactions. Okuley considers this significant: "Our new model is based on our data, not our organization," he says. "This ensures we get a sustainable stewardship model that provides a comprehensive knowledge of customer across Nationwide."
2. Define Discrete Activities
In their zeal to build an organization, teams that take the "who" approach tend to overlook the work itself. While a data steward's tasks will evolve with the role, it's imperative that the team defines and communicates the work activities that will be performed. This positions the role of data steward as one of delivery and avoids the "roving linebacker" syndrome. The figure describes the set of activities identified as the initial tasks for a new data steward at a brokerage firm.
In this case, our client needed to launch data stewardship by assuring various lines of business that the data - including its definitions, its standards and its business rules - would be thoroughly profiled and documented. Because of this, the data steward was chosen based on his past experience with this type of work. Additional data stewards would be introduced as the initial role showed progress.
3. Expect to Formalize Data Management Processes
In the example, as data was profiled and data deficiencies were tracked, more intelligence would be gained about how to formalize and improve data management processes. As these processes became more formal, collaboration with the business and IT would assume a higher priority. In this way, the team could justify how data stewardship would help the business while at the same time showing a plan for its evolution.
This meant at least one data steward needed to be adept at process documentation. If you're simply introducing a data steward role to support a finite project or, as many do, to "fix the data," you've likely sacrificed an opportunity to entrench broader data management responsibilities. Data stewardship is bigger than data quality. Establishing data stewardship for the long-term means not only communicating individual activities but also putting them into the context of an overall data management process. The client mentioned previously instituted a team of data stewards to enable a larger corporate data management effort including data architecture, data requirements, metadata management, business rule definition, data architecture and data quality. The stewards established themselves as adept at executing new governance policies and consequently, vital to ongoing information management. In fact, they became ambassadors on data's behalf, proselytizing the concept of data as a corporate asset.
4. Communicate the Metrics
It's not enough just to handpick some data-savvy professionals and bequeath them a new job title. Any effective data stewardship program should involve measurement. This means the ability to answer the question, "What has data stewardship contributed to the business that wouldn't have been possible without it?"
At Nationwide, Okuley assigned a data stewardship coordinator for that very purpose. He explains, "The coordinator role is accountable for maintaining the process and mechanisms to ensure progress. That means knowing what success looks like and defining what to measure. Are our issues getting resolved more quickly than they were before? That, to me, is a hard metric."
Of course, there are many ways to measure success. The point is to identify the opportunities and begin making improvements.
5. Don't Ignore the Soft Skills
Everyone talks about soft skills, but in practice, few managers really evaluate them. Strengths like excellent written and verbal communication skills tend to be overshadowed by other skills, most often familiarity with specific data or an incumbent software product. But people can learn data correction processes and tool functionality. The best data stewards I've seen have these three essential skills:
The ability to convey concepts with metaphor and images. This is useful when presenting new processes and concepts to management and business users who may be uncomfortable with data management jargon.
An understanding of "root cause" analysis. Successful data stewards need to peel back the layers of the onion and uncover the problem, be it a business rule glitch or user input error.
A knack for winning others over. The best data stewards actually get to the point where they can ask people to trust them without long-winded explanations of hierarchical models or matching algorithms. And people do.
A company facing multiple data steward candidates should put these criteria at the top of the list. It will make the choice much easier and set a precedent of excellence moving forward.
6. Strategy First, Tools Last
This is an ideal mantra for a data steward, a reminder that data stewardship should provide business value. But that's easy to forget when you're trying to determine which customers belong to which households or why there are 14 different prices for the same product. In their rush to problem solve, many data stewards turn toward technology.
As with most business efforts, automation isn't the answer, just an enabler. To be perceived as worthwhile, data stewardship should tie back to the value of the information that is helping the company compete in a crowded space, create disruptive new products and convert its customers to raving fans.
The ability to support corporate goals and objectives - indeed, to propel them forward - will sustain the role of the data steward. If you are serving a higher business purpose, there will be more than enough time for data profiling and workflow tools.
Moving Toward the Work
The truth is that choosing the right data steward has little to do with who might be available and everything to do with the actual work that needs to be done. The key is to avoid the "vital signs" trap and get down to business, delivering new processes for managing and improving corporate information on behalf of various business programs.
"Our conversations are no longer definitional," explains Okuley. "They're actually about the work we'll be tackling first. And that's been a huge step forward."