Chief data officers zero in on governance challenges

As organizations seek to increase their use of analytics, they often find they have to get their data houses in order first. If data is siloed or structured in ways that make it unusable to others, the opportunities for analytics are limited. Information governance work groups are created to form policy, security and usage agreements. But even those work groups often report to senior management that what is needed is an executive-level position focused full time on governance. Thus, the chief data officer position was born.

Of course, what the CDO (not to be confused with a chief digital officer) does day to day depends on the industry and the size of the organization. There is not one CDO job description, just as there is no single information governance framework. But among the most significant challenges new CDOs face is establishing a new position and office and gaining credibility in the organization.

Brett Goldstein came to be the city of Chicago’s first chief data officer without having given a lot of thought to data governance earlier in his career.

He had worked at several technology startups before joining the Chicago Police Department, where he spent five years focusing on how to use the data police had to target areas to prevent violence.

He was just about to return to the private sector when Mayor Rahm Emanuel was elected in 2011 on a platform of greater transparency. “Mayor Emanuel asked me to come down to the transition headquarters because he wanted to create this new role,” Goldstein recalls. “He wanted to have this focus on open data, but he also said, ‘At the police department, you were able to use data to start doing business smarter. What if we did that on a broader city level and create this position?’”

Goldstein didn’t have much experience with the open data concept, but the mandate from Mayor Emanuel was not just to get into open data, but also to become the leader in the space. Previously, the city had dipped its toes in very basic data sharing with the public. “There was a platform and they had a few open data sets, but one of our 100-day objectives was to have a significant move in open data,” Goldstein says.

Big crime data release

He decided to do something big with crime data, which he already knew well. Historically, the city would post data on the police department website that would show 90 days of data. Companies like EveryBlock would scrape it off the site. “We decided to release all the incident-level crime data going back 10 years,” he says. “This was the biggest incident-level crime data release that had ever happened. It became international news. We went from being Chicago with a history of a lack of transparency to being the most transparent open data site as it pertains to crime data in the world. That was consistent with the mayor’s vision. I wanted to disrupt the status quo and show we had nothing to be afraid of.”

Of course, the recent controversies around the release of police video involving a shooting in Chicago suggests there’s more to transparency than open data initiatives.

In addition to opening up data to the public, Goldstein focused on predictive analytics and modeling within city departments. One thing that surprised him was that when he mentioned analytics, most people assumed they needed a consultant or a vendor, which would be too expensive. “My response was heck, no. Why not sit down and do it now? I have no fear of diving into code. That was new and different in city hall,” he says. They used an open source platform called ‘R’ and held classes in a city hall conference room. “There are lots of people who want to do business smarter. They just feel like they don’t have the tools,” says Goldstein, who left the city in 2013 to become a senior fellow in urban science at the University of Chicago.

Goldstein says he had to learn about the city’s departmental needs and how to talk about change. “Until you understand the business environment and politics, you don’t know what you need to change and there is an embedded status quo,” he says. “When you look to change a process, you need to do so in a thoughtful way. I don’t mean slowly. But you can’t say we need to do everything differently. You have just thrown a word bomb out there. You have to figure out how you are going to do it and support people. It is not sufficient to do smart analytics. You need to explain to people what you are doing. You can’t talk data science to them.”

Champion for the user

Nicholas Marko, M.D., is chief data officer for Geisinger Health System, a large integrated health system based in Pennsylvania and known in the healthcare field for the sophistication of its technology deployments.

Speaking about the creation of his position at the 2015 Big Data & Healthcare Analytics Forum in New York City, Marko said healthcare is realizing it can improve the quality and efficiency of care if it uses the information it has to the maximum possible extent. Geisinger is developing structures around how it stores the information, how it is accessed and federated out to various users in the enterprise, and that includes physicians, researchers and patients, according to Marko. “These are all areas that need someone focusing on that part of the data domain. The CIO is often focused on infrastructure and IT policies. The CDO is focused on the content of the information and how it is used and is often a champion for the user,” he added.

There are still relatively few CDOs in healthcare, Marko said. So far, the healthcare systems that are deriving the most value from having CDOs are large, information-rich enterprises. They tend to be clinical organizations that have been collecting data for 15 years and have a variety of data domains they have to integrate or organizations that have a track record of using their data to focus on patient care or improving the processes of care.

Emphasis on policy

Two main issues led to the creation of the chief data officer position at the University of Wisconsin-Madison, says Jason Fishbain, the CDO there. One was that faculty members who wanted to work on improving teaching and learning were having trouble getting access to data. Another was that people outside the institutional research office were creating their own reports whose data sometimes conflicted with things institutional research had created. “Then the question was why don’t they match,” he says. Although he hadn’t had a CDO title before, Fishbain had been doing similar work for eight years in the healthcare field. “It is a very similar environment when you are talking about data governance, right? Talking about it with physicians is similar to talking about governance with faculty members,” he says.

UW-Madison has started with an emphasis on policy. Who should have access to data and why? If they do have access, what is their responsibility as a user? “That is where we have started here. There will be other pieces of the program to come,” Fishbain adds.

Mike Kelly, chief data officer at the University of South Carolina, says the creation of his position built on several years of data stewardship work around the implementation of an enterprise resource planning (ERP) system. Between2006 and 2008, the University of South Carolina developed the concept of data stewardship and identified who data stewards are by the areas they manage. They started working together through a structure called the Data Administration Advisory Committee (DAAC).

The ERP project was put off due to the recession, but the stewardship work continued. In 2013, members of the DAAC approached the university’s chief information officer and said they needed someone to coordinate their activities and hold them accountable for the decisions they make. “They were asking for a position to coordinate them,” Kelly says. “It was seen as the chair of data stewards. Our CIO was paying attention to this emerging role across business and industry. I had the luxury of having a hand in writing the concepts and reviewing the input. We kept asking the stewards, ‘Does this sound about right?’ Higher education is a very democratic culture. We check back a lot.”