If you're looking to launch a data governance program in your organization, don't forget to include a for creating...
data quality metrics that can quantify the initiative's business value. That was one of the top and implementation priorities cited by data governance managers who spoke during a virtual conference held this week.
Other to-do items detailed by the speakers included getting deep levels of business buy-in and participation, setting up a solid data governance structure for managing the process, and devising a communications strategy for spreading the governance word internally. But, they said, a lot of making it all work -- and getting continued support and funding -- hinges on being able to show that a governance program is paying tangible business dividends. And that typically comes down to documenting revenue gains or cost savings from governance-generated data quality improvements.
"What [business managers] really care about is data quality," said Jennifer Ippoliti, firmwide data governance lead for the chief data office at financial services firm JPMorgan Chase & Co. "That's how the business sees it -- not as a data governance problem but as a data quality problem. And sometimes, that's how you have to spin it to make it real for people."
In a presentation as part of Dataversity's Enterprise Data Governance Online conference, Ippoliti said data governance teams need to establish data quality improvement goals and then track and report on the progress being made in meeting them. And they should be "business-meaningful goals," not data-centric ones, she said. "If you have a goal of improving 500 pieces of data by the end of the year, so what?"
No hiding on data quality issues
Jennifer Ippolitifirmwide data governance lead, JPMorgan Chase & Co.
Outcomes that fall short of the desired mark on the data quality metrics being tracked shouldn't be swept under the rug, Ippoliti added. At New York-based JPMorgan, data governance team members point out both good and not-so-good results and quality measurements during meetings with business representatives. "We use the term 'blame and shame' at times," she said.
Michele Koch, director of enterprise data management and the data governance office at Navient Corp., also spoke during the virtual event. Koch said she learned an early lesson about the importance of data quality metrics when Navient, a student loan management and servicing company in Wilmington, Del., was still part of loan provider Sallie Mae.
At the time, Koch led a pilot data governance project to address quality issues in seven data fields that were being used to trigger new marketing emails to prospective customers. All went well -- until corporate management asked her for some quantifiable metrics on the project. "I realized that I didn't have the information," she said. "And I had to go back and spend quite a bit of time gathering it."
Navient, which was split off from Sallie Mae in 2014, now plans and monitors data governance projects based on three business drivers: increasing revenue, lowering costs and reducing corporate risks and regulatory compliance issues. The data governance team assesses the business impact of data quality problems through questionnaires and interviews with business managers who are members of the company's data governance council, then uses a worksheet to calculate the business value of potential fixes.
Coloring in the data quality lines
The governance office also maintains a master spreadsheet of data quality metrics that's updated weekly, with green, amber and red color coding to indicate whether particular data sets have acceptable error rates or have fallen below predefined accuracy thresholds. Barbara Deemer, Navient's chief data steward, joined Koch in the presentation and said that when something goes red, the governance team does research on the cause of the data errors and what it would cost to fix them versus the expected benefits so the governance council can decide which issues should be addressed.
In addition, an email is sent weekly to council members with the latest information on data quality, so they can see what's being monitored and where things stand on error rates. Deemer said the data governance team also distributes monthly reports to people associated with the governance program via a business intelligence (BI) dashboard. The dashboard, built with SAP's BusinessObjects BI software, provides details on three types of metrics: the state of data quality in Navient's systems, the business value derived from quality improvement efforts and statistics on data fixes.
Other technologies used as part of the governance program include SAS Institute's data quality management tools and an Oracle database, which stores the results on data errors.
But the overarching goal is to ensure that the governance efforts are seen in business terms, not as an IT exercise. Deemer, who has a background in accounting, noted that before interviewing executives about the business value of data quality issues, she maps them to Navient's financial chart of accounts. Doing so, she said, lets the execs "instantly see where these issues might be costing their divisions money, and how fixing them could improve operations."
Read a Q&A with Navient's Michele Koch on best practices for data governance
Consultant Andy Hayler on why companies need to do better on data quality metrics
The problems that data inconsistencies can cause in companies -- and how to avoid them
Find out how SAS Data Governance helps support enterprise data tasks