Business Needs to Get Skin in Data-Quality Game


Bad data may not be the root of all IT evil. But it comes darned close, with its ability to wreck such a diverse array of tech initiatives.


Business intelligence not meeting expectations at your company? Your data might not be clean, timely or centralized enough. When asked how they could derive more value from BI investments, 47 percent of respondents to a recent CIO Insights survey cited improved data quality.


What about your efforts to build a service-oriented architecture? Poor data quality is a problem because SOA involves sharing data as well as other services, Current Analysis analyst James Kobielus told IT Business Edge in an interview last January. He says:

... corporate master reference data is also a critical resource. That's the data that you run the business on, like customer records, HR records, financial information. In an SOA environment, you want to maximize the sharing of the master, cleansed, official data of record throughout your organization.

Bad data can bea drag on your data center. It can get you into trouble with regulators, as another Current Analysis analyst, Robert Lerner, pointed out in an August 2005 interview with IT Business Edge. He says:

Regulatory requirements like Sarbanes-Oxley and HIPAA are bringing (the importance of data quality) home to a lot of organizations. You can establish a nice control for patient privacy, but if you don't know who your patients are, you could be exposing yourself to a whole lot of problems.

OK, we all know bad data is, well, bad. The problem is, many folks find cures like master data management as bad as the disease -- or at least as daunting.


While the problem of bad data often tends to fall to IT, the business should take responsibility for it, says Gartner analyst Ted Friedman in a ComputerWeekly.com article. In the full press release regarding Friedman's research, he recommends that each business unit get some skin in the game by appointing its own data steward. Stewards are responsible for ensuring departmental data is complete, consistent and accurate, and also that it meets broader corporate data standards.


Here's how Friedman envisions this happening:

Successful stewards are placed closest to the point of data capture and maintenance, are intimately knowledgeable about the data and its use in a business context. They have also a stake in improving quality. As such, they are empowered to make business process changes and apply resources to address quality issues. Furthermore, they can influence how their peers execute business processes to achieve further improvements.

The stewards report to a corporate sponsor for data quality, who serves as the ultimate point person by resolving conflicts across stewards and business units, says Friedman. IT Business Edge blogger Loraine Lawson goes Friedman one further by suggesting that companies should employ a C-level executive specifically charged with data governance rather than recruiting a corporate sponsor. She writes:

Data increasingly is the lifeblood for enterprises, whether you're dealing with sales or customers or actual product fulfillment. And when you have a team of high-salary workers grouped together to address an ongoing business challenge, it makes sense that there'd be an executive-level person overseeing that function.