The demand for data continues to increase, and some higher education institutions have invested in executive-level positions focused on a 360 degree view of data management. The Technical and Reporting Advisory Group (TRAG) interviewed Nick Stevens, Chief Data Officer at the University of Kansas (KU) about the nature of his role at KU, tips for those interested in taking on such a role, and what keeps him up at night.
What does a Chief Data Officer or CDO do, exactly?
Well, in the context of the University of Kansas, I’m responsible for the Data Warehouse infrastructure, the Business Intelligence (BI) stack, including the enterprise dashboards we have built as the source of truth for a financial budget report. I also have responsibility for the source environments, like the star schemas, and Institutional Research (IR) studies. This means I also provide information for our executives and also serve as the lead of all data governance initiatives, including serving as Chair of the committee. That’s what I do all day long.
How would you break your day up? Operational vs. strategy, for example?
It depends on the time of year. I would say that probably 60% of my time is spent on strategy, either driving something outward or facilitating strategy sessions. This can include working with academic deans, the Provost, or the Chancellor. Another 20% would be spent on operational things. This could include doing analysis myself or participating in analysis with a team. For instance, I’ve been leading an intensive equity study for faculty, looking at race, gender, and ethnicity, so for this project I’ve been pulling a team together to do this work and participating in the analysis due to the nature of the project. Finally, the remaining 20% would be technology, including data architecture, warehouse design, project prioritization, and working with external vendors.
How is your work different than, say, that of a data steward?
Well, I’ll share my opinion with you though others might have a different interpretation. Data stewards are responsible for the integrity of data in source systems that we ultimately integrate into the warehouse. They’re also responsible for determining the meaning and correct usage of data, so making sure the correct context can be pushed out to the rest of the campus. The Chief Data Officer (CDO) ensures the correct processes are in place for stewards to provide that context, and making sure that is pushed out to the community broadly. Also, the CDO makes sure policies reflect the values of the institution and not just a single business unit. CDO also makes sure all the stewards are coming together to consider the global University perspective. Another way of saying this might be that the stewards are gatekeepers; they are protecting the institution and kind of define security through business logic, and they ensure everyone has context. The CDO is building the capability to do data decision making, and making sure people have access to that data.
Explain different kinds of governance (technical and people)
To have a mature governance model, you have to focus on both. A challenge can be that you do data governance on the backs of people who are at least 100% resourced. When you make the scope both technical and people, it’s hard to make progress. With our data council, we’ve focused on defining the three strategic things we can actually get done in a year, and we’ve defined them. But, another piece of our council has been from a technology perspective, we’ve engaged them a lot in a change control process. We have a Change Advisory Board (CAB), and if we’re making any changes to the data warehouse, we publish our change requests out to our data stewards and have our 15 minute weekly CAB to provide opportunity for comment. Typically, if we're just adding a column to the database and surfacing it to the BI layer, we’d just notify a steward that this was happening and wouldn’t seek approval. If we are changing any logic or calculation behind the column, that would go through the change control process. The only time this isn’t the case is if we have, for example, so many requests to add columns that it’s not reasonable to go through all of them in CAB, we get the steward involved to help with prioritization. So we use the data council to determine what should move into the data warehouse, but also for change control and trying to get out of them changes that are going to occur in source systems so we know what the downstream impact is.
Is your job technical or business focused?
It’s both. An analytics leader needs to be strategic, but I think it’s important that they never lose touch with their craft. I still analyze data myself, I do mapping, I do Quality Assurance and look at data sets. I enjoy doing these things and it keeps me in touch with what my team is having to go through. But, I have a strong investment in what our technology stack looks like, and it’s the primary lens through which our campus sees the team because it’s what they interact with on a daily basis, so their perception of how they interact with the data matters. Their trust of the analysis we perform in the future can be viewed through the accuracy of the data in the systems that we’re all accessing to do our work. I have analysts and developers who report to me, and they’re different types of analysts doing different work. Having context and experience in both of those helps to have credibility with those teams.
If someone is interested in this role, what would their individual development plan include?
Finding someone who has both a technology and business background can be challenging, but both are important. I think I’ve been lucky to have roles in both on my campus. I was responsible at one point for research administration infrastructure--systems that support research proposals post-award--but in that role I was also responsible for external reporting. I made a career change and moved to implementing shared services on campus, and part of that work was building models where we could decide how many staff should move from units to shared services and how we could make sure we moved the right number of people. Then, I ultimately became responsible for Data Warehouse infrastructure and BI platforms, and later moved to the CDO role. I think you have to have some time of experience in a University line of business, like research, finance, registrar, or admissions...somewhere that uses data on a daily basis. But, you need to have some type of technology background. With almost every occupation in higher ed, you really need to have a master’s degree in something. One might choose social sciences, applied statistics, or something where you get a blend of technology from a tool set perspective, but also analytic experience.
Prior to this, I didn’t interact with faculty on pure technology issues. Now, I interact a lot directly with them. Sometimes when you build a regression model, you may go toe-to-toe with them, and I need to be able to back-up my method.
What’s been the biggest surprise for you since you took this role?
The scope. Being responsible for technology, research studies, and IPEDS is a really broad scope, and it’s required recruiting strong talent in those areas to lead each of those functions. The level of attention to detail it takes to do it effectively is daunting. If we change this “one thing,” what are all the downstream implications? Probably one of the hardest components of my job is having the appropriate context and responsibility for how everything gets used, and getting up-to-speed on that quickly.
What keeps a CDO up at night?
Who is putting together some type of analysis and giving it to someone to make a decision and doing a poor job of it? Just because we give access to data doesn't mean everyone knows how to handle that great responsibility. There’s a cost to poor analysis or a breach, but there’s also a cost to silos, and people not having access to what they need to do their jobs. It’s not necessarily a measurable cost. Also, big data. In this world of big data and having access to social media platforms, what data are we providing to third party platforms with good intentions that poses a risk to the institutions? There’s a lot we could provide that from our perspective is anonymous, but a lot of harm can be done in the world of big data and social media to identify people. We can know whether or not students are in class through wi-fi access points...do we want or need to know?
Even if someone fills out a survey and talks about an occupation or job they had even in general terms, you can find that out with high confidence through LinkedIn data who that person is. The challenge with all of those is that at the end of the day, it’s the institution's brand that takes the hit, even if we’re passing along data responsibility and we have legal means to do that. In the event of a breach, it’s always the institution that takes the hit, and the perception is affected. That’s the rub.
Final thoughts?
Being a CDO as the University of Kansas has defined it has a daunting scope and is challenging, but it is extremely rewarding. I don’t know many jobs where you are in charge of your own destiny from start to finish. Being responsible for everything that goes into the warehouse, documenting context, and being able to apply that information when a need comes up is rewarding if you can handle the scope of responsibility. To handle that scope of responsibility, I’ve been fortunate to hire talented people that are responsible and will take care of a lot of complicated tasks. Sometimes it’s my responsibility to remind them to take care of themselves in addition to taking care of the data.