There are multiple hurdles to get over along the path to effective data stewardship, but perhaps none is as daunting as the realization that the process involves a healthy dose of plain old hard work -- often without the promise of a payoff that can be easily documented.
The primary benefit of a well-executed data stewardship program is substantial: clean, accurate and consistent data across an organization, thanks to the work of data stewards in developing common data definitions and ensuring that business users adhere to rules on data entry and usage. Compared with the cost savings and competitive advantages offered by other technology and data management initiatives, though, the fruits of a data stewardship framework are more subtle and less widely recognized.
As a result, stewardship efforts are prone to floundering once they get going, according to analysts such as Shawn Rogers, who heads business intelligence (BI) and data warehousing research at consultancy Enterprise Management Associates Inc. in Boulder, Colo.
"There's nothing glamorous about keeping data clean and orderly, and there's no sexy payoff that a lot of other things have in an enterprise," Rogers said. "These are the types of projects that are started and get cut real quick because no one understands their impact."
With those cautionary words in mind, here are five of the most common pitfalls associated with data stewardship programs and advice on how to avoid them:
Lackluster commitment. Oftentimes, formal data stewardship processes are instituted as a result of high-profile data problems or because the stakes for new systems and applications are so high that the risk of failure due to poor data quality isn't an option. Those circumstances can help a program gain lasting support among corporate executives and business users, but that "stickiness" is harder to achieve if stewardship initiatives aren't tied to specific and visible pain points. Enthusiasm can wane when the going gets tough and there isn't a clear sense of the expected value.
"We see a lot of these projects start with the company all riled up, and then there's no real commitment to it -- just lip service," said William McKnight, president of McKnight Consulting Group in Plano, Texas. "People get busy with their own deliverables and don't follow through with data stewardship, and the program spirals downward."
Not recruiting the right people. Identifying good candidates for data steward positions is a key task that trips up many companies. The top priority, analysts say, should be establishing a functional partnership between the IT department and business units that can help companies identify end users who have a full understanding of the data in their domains and the ways in which it is used.
Some of the biggest miscues cited by the analysts include hiring data stewards from the outside or relying too much on executive-level managers who aren't as familiar with the intricacies of an organization's data as more hands-on workers are. It's also essential to make sure that the appointed data stewards are allotted enough time to devote to a program, especially if it isn't their only job responsibility. And the chosen individuals should possess strong communication skills and be effective facilitators and negotiators, said Jonathan G. Geiger, executive vice president at Intelligent Solutions Inc., another Boulder-based consulting company.
No connection to business goals and needs. It's critical to tie a data stewardship framework to an organization's core business priorities. In a company that's pouring millions of dollars into optimizing its supply chain processes, there's a disconnect in proposing a stewardship program aimed at improving the data used in online marketing campaigns, said Jill Dyché, a BI and data management consultant who now works for analytics software vendor SAS Institute Inc. Likewise, data stewards shouldn't get sidetracked on unimportant issues, she added: "If you're looking at how good your product data is but no one is complaining about it, the effort may be for naught because people don't see the benefit."
In addition, giving data stewards general responsibility for customer or product data can be too broad of an approach, according to Dyché. Instead, she advises companies to appoint data stewards at the business process level -- for example, tapping a user who understands every nuance of the quote-to-cash process to oversee all of the data related to sales and invoicing.
Aiming too high. As with most large projects, it's risky to try to facilitate wholesale change through data stewardship as opposed to focusing first on targeted implementations that address specific business needs and data issues. "You need to think globally but act locally," said Barclay Blair, president and founder of ViaLumina Group Ltd., a consultancy in New York that focuses on information governance. "If you try to act on an enterprise level right out of the gate, it's such a large problem that you'll likely fail."
Minimizing the change management issue. Data stewardship programs often face cultural challenges, and companies need to throw proven change management practices at the initiatives to ensure that people understand why they're necessary. For example, make business units aware of the problems caused by poor data quality and show them how sound data stewardship can help resolve those problems -- in language they can understand.
Working closely with the business side can also give proponents of a data stewardship program the kind of anecdotal information that can help build a business case. "Keep your ear to the ground and listen for when people say it takes three days to reconcile numbers or two weeks to find a particular data element," Geiger said. "A lot of people expect [data] problems to be part of the job, but it doesn't have to be that way."
Beth Stackpole is a freelance writer who has been covering the intersection of technology and business for more than 25 years for a variety of publications and websites, including SearchDataManagement.com, SearchBusinessAnalytics.com and other TechTarget sites. Email her at firstname.lastname@example.org.
- Accelerate Big Data Analytics to Capture Data in Motion –DataTorrent
- Implementing Streaming Data Applications –DataTorrent
- An Inside Look at Big Data Architecture –Imanis Data
- Fast Outcomes, Fast Insights, Fast Data –DataTorrent