This Q&A is part of a GovLoop series called “CDO Conversations.” We’ll feature conversational interviews with federal, state and local chief data officers to get to know the role and the people behind the titles.
Chief data officers (CDOs) are increasingly instated at all levels of government. But due to how new the formal position is, job descriptions may not be quite clear on how CDOs actually go about performing their roles. They fill an executive post tasked with leveraging data as a strategic asset, but what does that actually mean and how is that work taking shape in government? And how can rank-and-file employees understand the nuances of using data as an asset?
Tyler Kleykamp, former CDO for the state of Connecticut and current director of the newly minted State CDO Network at the Beeck Center (pronounced bek), sees data as the inverse of traditional assets that depreciate in value.
“If you think about a car, it loses value the more it gets used,” Kleykamp said in an interview with GovLoop. “But data really becomes more valuable, both the more it’s used and the more abundant it becomes.”
The uniqueness of data leads to a unique set of priorities and approaches that a chief data officer should take, especially as the position is beginning to form in government. Kleykamp spoke about how the State CDO Network, a nationwide community of state CDOs, supports existing and yet-to-exist CDOs, how the position has evolved and what non-tech leaders should know about data.
The interview below has been lightly edited for brevity and clarity.
GOVLOOP: Can you briefly tell me how you became the first chief data officer for the state of Connecticut?
KLEYKAMP: What’s unique for me compared to some other people that fill a state or city CDO role is that I was essentially a career government employee for the state of Connecticut. I’m a geographer by background, so I came to work in state government on what I thought was going to be a two-year project, and I ended up being there quite a bit longer.
A lot of my career was spent gathering data from other parts of the agency I worked in or, when I began working in our Office of Policy and Management, other state agencies or departments to do some form of policy analysis. I also spent two years working out of the governor’s office during the American Recovery and Reinvestment Act, and that was a huge data-gathering exercise.
When the administration began to consider creating the role, I became engaged in the drafting of the executive order. And essentially, as it was taking shape, the Secretary’s Office of Policy and Management decided that a lot of what they were asking the role to do was something that I was already doing, in somewhat of an informal capacity. So, they offered me the job, and the rest is history.
How did you become the director of the State CDO Network at the Beeck Center?
I tend to get a lot of credit for setting up the network, but around 2016, the CDOs for Texas and New Jersey — Ed Kelly and, at the time, Liz Rowe — began reaching out to some of the other CDOs. It was the two of them that started reaching out to people to coordinate, set up an email list, set up a conference call periodically. [But] the reality is, CDOs are pretty busy, so things kind of died down a bit. Then, probably about a year later, I realized I got a lot out of talking to them on a periodic basis, so I set up a monthly conference call.
I had seen what was going on with the Civic Analytics Network at Harvard and really thought we could use some support. I started writing up blogs and other things to put out there with the hopes that philanthropy would step forward and want to support the work that we were doing.
This past spring, somebody from the Beeck Center reached out to me. They potentially had some funding lined up and asked if I would be interested in moving over there to take it on. I think it was critical to some of the people that were interested in funding this that somebody who had been in the role [would] come do it, versus somebody who had never done the job. That’s how I wound up there.
What kind of activities or support do you provide as a network?
We have three areas of focus that we’re working on right now. One is supporting the states that are looking to create a CDO position. We’ll serve as a resource for them, help them with getting the job description together, planning what a CDO should be doing in their state, maybe even help them identify candidates for the role depending on their hiring process.
It’s literally my job to help states establish a Chief Data Officer. Even if you don’t work in state government, reach out if you’d like a CDO in your state. https://t.co/pyEkO0JJrP
— Tyler Kleykamp (@TKleykamp) December 13, 2019
The second part is supporting the existing chief data officers. We do that in a few ways. One is providing that communication infrastructure for them — an email list, a shared document repository, monthly video conference calls and other things in between. If they’re interested in a particular state project, we might host a one-off webinar. We did one initial in-person gathering in November when we launched, and we’ll be doing three more over the next 18 months.
We’re also building out a toolkit for them around some common things. Right now, we’re working on model legislation, but it could be things like playbooks around conducting data inventory or data-sharing agreement templates.
The third component is identifying some more discrete public policy or service delivery areas, where we could maybe organize a smaller group of states, say four or five, to really work through some of the more nuanced issues around leveraging data, whether it’s child welfare or criminal justice. We have a focus on economic mobility issues [too].
What do you believe is the value of a CDO to an organization? That is to say, what kinds of things can agencies do with a CDO that they couldn’t do before?
I think the thing that’s unique about the CDO in terms of asset management is thinking strategically about how data can be leveraged beyond the purpose that it’s initially collected. Data is different from traditional assets. If you think about a car, it loses value the more it gets used. Or people say data is the new oil — well, the more oil there is, the less valuable it becomes. But data really becomes more valuable, both the more it’s used and the more abundant it becomes. When we can combine two sets of data, it can give us insight into something. And that’s really the unique role that CDOs play. How do we get the most value out of data, thinking strategically about it, coming up with ways to integrate data across different programs and services?
How is the role distinct from a chief information officer (CIO)?
Especially as the way we deliver services has become increasingly digital, a CIO’s job has gotten big. They range from cybersecurity to just keeping applications up and running, even making sure phone systems are working. The platforms that are enabling technology are really where [state CIOs] are focused on. And so, I think the CDO role becomes different in this idea of the strategic use of data, being able to understand what policy priorities may be and identifying opportunities to better leverage data in a more proactive way.
How has the chief data officer position evolved since it first came on the scene?
I think a lot of the early CDO roles were heavily focused on open data. And that’s true for myself and some of the other early states, New York being one of them.
I think states are realizing that, as valuable as open data is, in terms of raw data, a lot of the state data can’t be made open data legally. But it can be shared responsibly and legally within government. So I think that’s where we’re seeing the evolution of the CDO role, this shift from open data to a more broad emphasis on integrating data across different agencies, either from a service delivery perspective or from the analytics-type perspective.
The more emerging technologies around machine learning or artificial intelligence — those are the technologies that if you don’t have good, high-quality data powering the algorithms, things can go really poorly with them. There’s a whole slew of ethical constraints and bias that can exist within data that you have to be in tune with, to ensure that those technologies are not having unintended consequences.
Is that something the Beeck Center is working on or involved in now?
We are looking at it. There’s a lot of interest in it. Right now, we’re still very much focused on the more foundational aspects of a CDO role, trying to put some of the tools in place to better support [CDOs] in their ability to share and integrate data across agencies. We would be happy to be a resource to do some more in-depth research or advising for them.
How can CDOs ensure they have proper empowerment as an executive making decisions at their agency?
I think the critical issue for CDOs is that they themselves need to understand what the priorities of the leadership are in their states and think about how they can add value to those as CDOs. When governors think child welfare, they don’t automatically think to call their CDO and say, “How can we use data to make our child welfare system work better?” CDOs need to start thinking about that in terms of what the governor’s priorities are, and how things like data governance, sharing data and analytics support the people that are working on these priority issues.
And then I think that leadership in states need to be clear in articulating how they would like to better use data when they’re creating CDO positions. It’s unlikely that a state is going to create a CDO position that is going to have a multimillion-dollar budget, right? We’re seeing they’re mostly one-person offices. There’s a handful of states that are a little more mature and well-resourced, but most of them are fairly limited in what they’re able to do. So governors and top-level leadership in their offices need to think strategically about what their priorities are for leveraging data and ensure that they’re communicating those to their chief data officer.
Can you speak about the main challenges you see CDOs facing?
Overall, I think we see in the state government there’s a lack of data literacy. So, frontline employees who may not understand the value of [data]. A lot of data just isn’t ready and standardized in a condition to be easily put to use.
There’s also a series of legal challenges. What I mean by that is there are a lot of laws in place to protect data, and they’re there for good reason. But there are a lot of them. There can be any number of state or federal laws that protect the data from a given program, and they all have their own nuance in terms of what is or is not acceptable to use. So it can be challenging to navigate the myriad legal issues that exist to protect data.
How can CDOs help make data more accessible in the workplace for frontline employees?
I think, like anything, it’s the education and training of employees. When you think about government, there are all these training requirements that we have, whether it’s violence in the workplace or ethics training. Every year you have to take something and refresh it, and they’re somewhat mandatory. So I think we’re getting to a point where, as we become more and more digital overall, there’s a need to have some form of basic data literacy training for government employees. I would say that at all levels of government, even if it’s a one-hour [training] where you give employees the basics, that can go a long way, in the same way that CIOs are implementing cybersecurity training for all their state employees.
What do you want people in leadership positions who are not necessarily tech-related to know about data?
I’ll use the juvenile justice system as an example. People might say we don’t know what happens to individuals once they leave the juvenile justice system, because they don’t track that. The reality is, we can answer questions about that, and we can do it in a way that protects those individuals’ identities. [We can] know whether once they leave the juvenile justice system if they are employed, if they’re in high school or if they go to college. We can know whether they end up back in the adult criminal justice system. We have that data — [and] we can securely integrate it and de-identify it so that we’re not invading their privacy.
We can answer some of these really challenging questions with the data that we already have. Even [if] we’re not tracking something, there may be data that can inform a lot of these critical functions. That’s the topline goal — we want to empower and enable policymakers to answer these critical questions in easy and repeatable ways.
If you know or if you are a chief data officer interested in being part of this series, please reach out to pearl.kim@govloop.com!