Business intelligence without data lineage, quality, and definitions is suspect

Data governance isn’t just about technology. It’s also about establishing business ownership of the data by assigning them official roles and responsibilities.

“It’s impossible to create new enterprise processes and enforce them without executive leadership. The people who believe in it will jump on board. The rest don’t.”

—Terri Mikol, director of data governance at UPMC

UPMC is at the forefront of data-centric, transformative healthcare, with the goal of improving research outcomes for cancer and other diseases. The organization’s $100 million-plus enterprise analytics initiative is aimed at determining the true cost of providing care and analyzing clinical practice variations to create best practices in disease management.

Meeting these goals requires clean, secure, and connected data. But data quality requires more than technology, according to UPMC’s director of data governance, Terri Mikol. She discusses why business education, ownership, and oversight are essential in order transform data into valuable business intelligence.

Data analytics seems like a technology investment, but you’ve made a considerable investment in processes and people. What was your thinking?

Terri Mikol: Assigning data ownership to the business is a key strategy in our efforts to elevate the data analytics abilities throughout UPMC. To become a truly data-driven organization, we want to remove the “secret sauce” from our business intelligence and reports. Once the business understands the process, they can better trust the data.

How did you go about assigning responsibility?

We are involving more people in data-related work through a broad data governance program. Stewards from across UPMC are now engaged in selecting the best sources of data, defining business terms, and working on improving data quality.

Our information owners are from the business, not IT. They govern a domain of data across the organization, regardless of the application and regardless of the business unit. The information owner of registration data, for instance, governs all registration data across our hospitals, clinics, physician offices, emergency centers, and post acute facilities.

Multiple data stewards are named for each data domain as well. Guided by the information owners, these stewards provide local expertise regarding a specific care setting or application. Collectively, they are creating valuable metadata about where all of their data is used and what it all means.

Could you have done this without a governance council?

Mikol: No. I spent 12 years in an analytics role without a council, and it’s impossible to create new enterprise processes and enforce them without executive leadership. The people who believe in it will jump on board. The rest don’t. If you are trying to change the way people work, it needs to come from the top down.

Success really boils down to two things:

  1. Shared stewardship of the data. And that takes ownership and education.
  2. Not quitting. This isn’t new. We all know what it takes to do data warehousing and business intelligence. We just need the stamina and the persistence and the belief to stick with it long enough to truly make it enterprise.

The strength of UPMC’s efforts lies in the recognition that it requires more than technology to succeed at their ultimate goal of moving from predictive to prescriptive analytics. By creating shared stewardship and accountability, the organization has freed data trapped in application silos, resulting in quicker access to better clinical data.  

For more information on the considerable technology requirements involved in Mikol’s data governance initiative at UPMC, read “Converting raw data into business intelligence.”

Related content



UPMC used a collection of Informatica products to improve research outcomes in the quest to cure various diseases


From adversary to ally: Building a CFO-CIO alliance

The CFO and CIO need to foster a relationship based on their shared need of accurate data to advance their reputations as strategic business leaders.


Shifting trends require shifting responsibilities

Developers and analysts must expand their skills if the business is to succeed in its efforts to improve agility.


See big data through a different lens

A successful big data project could be a big win for your organization and your career. Focus on the wrong priorities, however, and your big data project will be over before it even begins. Prioritizing data veracity, visualization and value is the key to success.