 Hi, everybody, we're back. This is Dave Vellante of Wikibon.org. My co-host today, Paul Gillan, who I'd like to thank, had to leave, but he was fantastic, so I really appreciate his attendance here on theCUBE. He was a CUBE newbie, but you never would've known it. This is theCUBE, and we're here at the MIT Information Quality Symposium. We've been here for two days. This is, we're at the tail end now, and, but we've got a very high energy guest, Mission Lean Casey is here. She's the Chief Data Officer of the Federal Reserve. Mission Lean, great to have you on theCUBE. Thanks for coming by. Thanks, Dave, appreciate it. Yeah, so you were the first Chief Data Officer at the state level in the state of Colorado, the first Chief Data Officer at the Federal Reserve. What makes a good Chief Data Officer? What are the characteristics of a Chief Data Officer that we should look for? Well, I think there's a combination of skills. I think strategic leadership is certainly one of those things. Someone who's been in the data world and the high tech world, that's important as well, but also people who are used to dealing with political situations. Traditionally, data's been siloed off, and we're at this point where Chief Data Officers are really silo busters in a number of ways, and Chief Diplomatic Officers, and having to span and cross lines of business and bring cross-functional people together to really leverage and optimize data in ways that it hasn't been done before, so I spend a lot of my time actually on the communication side, on marketing and strategy and figuring out what the service portfolio of our group really needs to be relative to the board's mission and strategy and how we leverage data and optimize data in support of board mission, but even at the state government level, being able to blend the data strategy and tie that and align that to mission strategy is really important, so someone who's a good strategic thinker and can work with the C-level executives across the organization to really understand what the goals and missions and objectives of the organization are. So organizational dynamics, obviously, communications, strategy, did you have a background in organizational dynamics in any way, shape, or form, or did you just sort of natural at it? No, it's something I've been observing over a number of years. I've worked in a lot of different types of organizations, actually. Small companies, force startups, very large organizations like IBM Global Services, and so I think to me as sort of an observational person, the structure of an organization can really support or impact and hold back work that needs to be done and I think it's an important topic that is under discussed in the conversation around CDO. It's so true. I mean, organizations can either be an inhibitor or accelerator and it's sometimes hard to figure out how to approach that. Change, culture change and the culture of an organization is also in a similar model and organizations really being able to embrace a different mindset around managing data. So I spend a lot of time thinking about that and executing, trying to execute against that. So let's talk a little practical experience. So when did you join the State of Colorado as a CDO? Well, I joined the State in 2007 and I actually was the Director of Identity Management for the State and then in 2008, the legislature passed a law. The first one that we did around better improved data management. They were having challenges and the governor was having challenges in understanding and getting answers about key business data sets around number of employees we had, number of citizens we were serving, number of vehicles in our fleet as examples and they wanted to have better data to improve policy making, to improve financial decisions and they recognized that we could be doing a much better job of managing data, sharing data, exchanging data to improve outcomes for not only our citizens but the businesses that we were supporting in the State of Colorado. Which year? 2008. That was eight. Okay, so it went from sort of identity management right into the fire of- Well, I actually held both roles for the remainder of my tenure, yeah. Okay, so I'm sure you got a big raise for that. No. What were your first 100 days like? It was the first time you were a CDO, right? At the State of Colorado? Yeah, right, is that correct? You weren't a CDO before that. I was not a Chief Data Officer prior to that. No, so yeah. Your first role, first time you were a CDO, you're like, okay, this is new. Yeah. There weren't a lot of, there still aren't a lot of CDOs around. So what were your first 100 days like? Not a lot of role models. Well, we were actually executing against a piece of legislation that was around information sharing and information exchange in the State. So one of the first things I did was actually pull together a cross-functional team that was representative of the majority of the State agencies that we had and we started building a plan that tracked along three angles. First of all, trying to inventory the data that we did have, trying to understand how information was and wasn't being shared across the organization and getting our arms around policies and processes that may or may not be in place to support that. So a baseline? We're baselining and also gathering information about what could be possible if we were managing our information. So trying to understand your constraints, the out-of-scope expectations, where we wanted to get to and what the future State could look like in having those conversations as part of, we're here, we need to go there, what does that roadmap look like? So that was part of step one? That was part of step one. Okay. And we subsequently passed three laws after that around data and information management. We actually, the second law we passed put our Enterprise Governance Council in legislation. We wanted to make sure that it would span administration changes. So instead of doing executive order, we chose to go through the legislative route and had phenomenal bipartisan support from our partners in the State Legislature and then pass a couple of other laws after that. So I want to make sure I didn't miss it. I think I heard you say that when you started out, you really focused on three areas, which I called baselining and you said that part of that was the future State and what you could actually achieve. What else was part of that three? Organizational understanding who we needed to have in order to operate and implement. So our approach was generally understanding where we needed to go, but what the current state needed to look like. Operationally, who we needed to have as part of an implementational team. And then thirdly, getting our arms around the policies and processes, we would need to support that. Okay, and then just for others who are thinking about doing something like this, we heard that trend say it's not a project and I know I'm speaking almost like in project terms, but you gotta get started. You gotta get started somewhere. And these three things, the baselining, understanding the organizational requirements and the policies that you need to put in places, sounds like a reasonable thing to do. How long did that take just to get your arms around that before you could actually start the journey? We took the better part of a good year. Six months actually to do the beginning of the baselining. Out of that work, we identified some target priority areas that fed into the bigger picture roadmap that we put together about how we needed to execute and what the priority areas were because we hadn't managed data commonly as an organizational asset. If you picture a state government, we're a $19 billion organization with 30,000 employees and every agency prior to Governor Ritter's administration had its own chief information officer, its own way of capturing data, storing data, maintaining data at some set of policies and processes. So certainly there was a lot of work to do, but we needed to be very surgical about how we went about accomplishing mission. And so we put together a series of priority needs for the state organization and hide it to what I called four anchor projects which were high priority statewide initiatives, education, juvenile justice, healthcare, and transportation and brought common groups together to actually start tackling the common challenges that they were each facing, but in a common way. Okay, so now it's 2009, right? So this started in 2008, if I heard that right. It started in 2008, 2009. Of course, the world was dominated, the entire discourse in every corner of the world was dominated by the financial crisis. What effect did that have on your ability to execute? Well, so we had to cut 20% of state budget at the same time and we were also undergoing an IT consolidation. So sort of the backdrop to me becoming the chief data officer was one of the first things the governor did coming in was to centralize IT under the governor's office of information technology. So state IT had never, state IT as a corporate service had not existed before. Again, each agency had its own CIO, its own budget, its own IT personnel, et cetera. But we were going to centralize service model. So that was going on and there were a lot of changes that were happening in the midst of financial crisis overlay and the fact that we also had to lay off 20% of staff but still supporting all the services needed to be done. So for example, unemployment rates just shot through the roof and there were major, major pieces of work that needed to be tackled around the state's unemployment system which was actually implemented in 1969. Right? So Neil Armstrong walked on the moon when this system got in and the state, written in lines of cobalt and the state legislators couldn't understand why we couldn't just make some gooey change. And it all of a sudden be able to handle the 600,000 more new applicants that we were dealing with. So there were a lot of dynamics going on at the same time that we were trying to do data governance and improve data management and pull together a cross-functional governance body and actually implement again specific project. So the IT folks were migrating the unemployment system. They weren't migrating it, by the way. They built a web front and on top of it but it didn't get migrated. Good because it never would have got done. For all I know it's still operating in the 1970s. I mean it, it probably would have never got done if they had to freeze the code and migrate it. Exactly. They couldn't have frozen the code and taken the whole thing under. It was one of those situations where you all of a sudden had to deal with this tremendous spike in service but still support constituent needs. Now you, where did you fit in the organization? Within IT or? So I reported directly to the chief information officer which I generally don't recommend. It's been a consensus here that that is not the best. I want to talk about, go ahead. So I'll talk about the plus and minuses and generally speaking because the data needs of an organization are really business driven. The decision rights, the accountabilities, the business processes, all of those things really are business owned decisions and so fundamentally I really believe the chief data officer should be on the business side of the house somewhere. In the case of state governments though from a structural perspective, typically, you know, you don't have a CEO and you don't have a chief operating officer typically. So it's a different sort of organization structure. Because the governor was consolidating IT and moved us into the governor's office, we were sort of neutral Switzerland, right? The only other places to put a chief data officer would have been into one of the executive branch agencies which would have pissed off every other executive branch agency, right? So we were only the only neutral place. Inherent advantage here. Exactly, exactly. So in that case, because of the organization structure, it worked out fine but generally speaking, I truly support a really tight alliance between the chief data officer and the chief information officer or chief technology officer because truly the technology needs to be supporting the business and can support the business, whether it's through analytics or enterprise architecture, there's so many ways but the data requirements need to be different from the business side. Okay, so talk about the outcomes of all that effort. So now you're slowly coming out of the recession in 2010. You've got, you know... Well, the rest of the country was slowly coming out of the recession. State governments are now, just now coming out. Yeah, I was just saying, the tech boom was 2010 but that didn't ripple through. Yeah, I mean, if you think about government organizations, they tend to lag behind. They're just now actually starting to get the revenue increases. Yeah, 2010 was maybe, okay, this might actually end at some point. Right. But it looks like it's not getting worse. That was sort of the state of 2010. Exactly, so it was starting to stabilize but certainly state coffers weren't increasing and we weren't adding headcount. It's what we're doing. But what we had done was set up a very small dynamic, what I considered a highly functional team. It was myself. We had a chief enterprise data architect, a business analyst and a project manager and we had, again, identified these four major cross-functional initiatives that were actually getting funding from the federal government. So we had money coming in to do these things and so we were able to actually continue our work with funding streams that were continuing. So that was actually extremely valuable for the state. And so what we did was partner with each of these initiatives and had them build money in to support the work that our office was overseeing and doing and that included everything from hiring, business analysts for the projects to inventory, the systems and applications and data sets relevant to those projects because we didn't have an enterprise data inventory for the state. We didn't have an enterprise application inventory. So there were some baseline foundational things that needed to be done that we started with these four initiatives and started to fill in the gaps. And over time, hopefully that those gaps have been closed and they've got a better sense from an organizational view what data they have, how that ties to business process, business function, et cetera. Now, when did you leave the state? We had a gubernatorial transition at the beginning of 2011, Governor Ritter did not run for reelection. And so I left at the end of January of 2011 and just went into private consulting practice. So what impacts were you able to see or were you able to measure at that point in time? Started in 2008, exited in, you said the beginning of 2011, end of 2011. The beginning of 2011, January 2011. What impacts did you? There were a number of things. So first of all, we did end up with four laws around enterprise information management and information sharing. We ended up with quite a number of enterprise policies and processes to begin our sort of repository of artifacts around data governance and started to align the way the various divisions were actually managing their data. One of the biggest wins was we ended up getting a $17 million grant from US Department of Education to build our state longitudinal data system in the education space, which never would have been built without that grant funding. But one of the things that stood out to the US Department of Ed was how our data governance process was being managed within the state that made us one of the early winners in the race to the top grants. Okay, now you had a legislative edict in the state of Colorado. We did, highly supported by the legislature and the governor. So that was the catalyst for the role of the chief data officer. Now you joined the Federal Reserve. What was the driver there for them to say, hey, we need a CDO? So with the backdrop of the financial crisis in mind, a couple of years ago, the Fed engaged McKinsey consulting to come in and start doing some helping support strategic planning for the organization. I think there was a realization that while the Fed is obviously an extremely effective organization, they've traditionally managed the research and the supervisory regulation the size of the business very separately and hadn't until that point quite seen the interconnectedness of where the two intersect which may or may not have caused the financial crisis. I'm not gonna get into that, certainly I don't know that, but out of the planning was an identification of the need to share more information across organizational silos and to be able to better manage data as an organizational asset. And so if you look at the strategic framework that the Federal Reserve Board has, it's number one priority is financial stability and it's number two priority is better data management and data governance. And part of the series of objectives that fits into strategic element number two is the creation of the Office of the Chief Data Officer. So I joined the Fed on May the 20th and that's when officially the Office of the Chief Data Officer got launched. And that was an independent decision obviously that the Fed made, right? And said, okay, this is a good idea, we need to do this, the executives of the Fed said, all right, we want to do this, we approve it, let's do it. So that there wasn't obviously the same kind of legislative edict. So the question I have is, because we heard today only, let's say one in 10 organizations have a board approved data strategy, probably less than that have a Chief Data Officer. It's gonna be low single digits maybe, I don't know. And maybe there's some kind of de facto data officer, but generally speaking, if there's not an edict, there's maybe less of an incentive or a less likely level of probability that you'll see a Chief Data Officer. Or do you subscribe to the scenario that because organizations are trying to be more data driven that data are a Chief Data Officer is going to be fundamental and organizations are going to see this? How do you think that'll shake out initially? I think over time organizations that really are information heavy to their business will begin to have the Chief Data Officer role more and more. I certainly think that the best way to set it up is to have the executive sponsor and authority of your Chief Executive Officer, your board of directors, the C-suite. They absolutely need to be behind that role in order for that role to succeed. Part and parcel of that is a change in organization culture that enables folks to get out of the mindset of the data's mine, the data's mine, and really view data as an enterprise asset and be incentivized to manage it that way. Yeah, so what should the Chief Data Officer, what should be the scope? I don't like the word own, but you know what I mean by that, not own, but what should be the sphere of responsibility for this Chief Data Officer? I don't know if you saw Mario's discussion today. He's on one of the panels with Derek Strauss. Yeah, the first one this morning, yeah. I was amazed at the number of disciplines that he had up on that slide. There were 11 areas. It was like M&A, ops, data governance, policy, security. Mario's very talented. He can do a lot of stuff, so. Yeah, I mean, I almost felt as though, wow, that's almost too ambitious. Now, of course, he's got a much smaller organization, but what areas of the organization should be under that umbrella of the Chief Data Officer? I think part of that answer is that it's gonna depend on the organization, right, and what's needed, and I've been in organizations where they have owned as much as what Mario showed, which is really sort of the Dimmbach model, right? Everything from data governance through data architecture to business intelligence and analytics. In some organizations, data acquisition is part of that, depending on whether or not organizations are looking for content externally or not looking for content externally. In some organizations, the BIC is a really big part of that, right, the Business Intelligence Competency Center. Some organizations don't have that. I guess my view of the world is that, generally speaking, the Chief Data Officer, first of all, owns the Enterprise Data Strategy, right, and works with the rest of the C-suite and key executives and stakeholders across the organization to understand how the data strategy is gonna be in support of the Enterprise Business Strategy and goals and missions, whether it's we wanna increase revenue opportunities and be more innovative, or if we're trying to reduce operational costs and improve operational efficiencies, right? And there's a range of things that can be done in there. Enterprise Data Governance and Enterprise Data Management, I think, are really important core pieces of the Chief Data Officer. In particular, I think, certainly, organizations that are really gonna undertake this need to set up an Enterprise Data Governance Council, whatever that's called, that has executive support and it's very cross-functional in nature and is looking at the systemic issues around data and really tackling policy process standards in a common way for the organization. Lastly, I think the really sort of big other role the Chief Data Officer plays is as change agent for organizations and again, that's everything from breaking down the traditional silos and getting people to think about the bigger enterprise needs while still managing the what's in it for me conversation to be in the storyteller about why improved data management is the right thing to do in support of business goals and objectives. So there's this change agent role and evangelism role that I think the Chief Data Officer has a really big voice in playing to the organization. Transformation opportunity. Transformation and innovation, if you can get to that point, right? I think Chief Data Officers today are doing a lot of the block and tackling work because it's been under data hat while lots of people and organizations say data is the most important asset of the organization. The time and money hasn't been invested in managing data. Yeah, and the data role's been just a back office role in many organizations. In many organizations and so I think for Chief Data Officers coming in, it's very much a startup shop and a lot of foundational work needs to be done. We heard it this morning, just even having an inventory of all your data assets. Most organizations don't have that, right? So getting that baseline is really foundational in terms of understanding how your information flows are happening, what systems are connected to what systems, dealing with data quality issues, dealing with metadata issues, the whole realm of things. You can't manage your data if you don't know what data you have. You can't secure it if it's invisible to you from a security officer perspective and you can't optimize it and leverage it if you don't know it's there. So we've talked sort of indirectly about this but I want to ask specifically how social data fits and how you see CDOs leveraging social data. I know, again, as a good consultant, good former consultant, it does depend. It does depend. But we haven't talked directly about that today. Social data specifically, what do you see there? I think there are really specific, interesting use cases that certainly are happening today, right? The NSA is a perfect example of, they're following our phone calls, they're following us on social media. They're connecting those dots, right? Interesting to follow the opinions on that too. They fall on both sides of the aisle, don't they? They fall on both sides of the aisle. And certainly the Intel community is doing a much better job of connecting the dots and they are following social media and looking for patterns and analytics. On the flip side, I think healthcare organizations are even looking at social media feeds and how to incorporate that as part of their 360 member profiling or member consumer engagement and outreach and how can we reach out and connect with our customers and our members and our potential clients in a more effective way than we've been able to in the past. Retailers are a great example of this, right? Of how they're leveraging social media to send you a coupon or let you know that there's been a product announcement. But I think there's a lot of big opportunities, again, to pull that into other big data analytics, to understanding where things are trending and be able to tie that into a broader suite of analytical services that an organization has. All right, Mr. Lee, my last question is advice for young people. I'm always telling my kids, get into some kind of data field which means they probably won't because I'm advising them that. But what advice would you give to kids in college? They're interested in data. They're interested in this field. Maybe even aspire to be a data czar at some point in time. What would you advise them? A couple of things. And actually, if you heard Dr. Michael Rappa yesterday, I think he's got a really nice approach to how he runs his program down at Etsy State. We didn't hear him, but he was on theCUBE and we're actually gonna do some stuff with him. He should, he's great. But he takes a really holistic approach because if you don't have the strategic thinking skills and you don't have the communication and marketing skills, then what you've got is a really great data analyst or data programmer, but someone who can't communicate the value and tell the story to the organization about how the data's improving X, Y, or Z, right? And at the end of the day, that data's gotta have some impact and value to the business. And so having folks who have some technical skills but more importantly are able to think strategically, understand the business, understand what's driving the business, and have those excellent communication skills so they can sell the story and even know how to tell a story to sell, right? Why does this data matter to us as an organization? Why is it important? Those are the kinds of future data leaders. I think we're gonna need in the future and hopefully the business schools and masters in science programs and engineering programs and mathematics programs are beginning to get that message. You hear that way. So I was gonna say, what should my major be? Should I be a marketing major, a psychology major? I was a marketing major undergrad and then got my MBA. So I'm not a hardcore programmer but I am someone that spent a lot of time on the business side of the house and leveraging technology to improve business outcome and leveraging data to improve business outcome and be able to tell that story in a successful way to an organization. I said the last question. I wanna ask one more. Thinking back on your experience, the state of Colorado, if you had to do it over again in thinking about maybe some of the mistakes you made, what would you do differently? One of the things I say is the biggest thing I did not do well in Colorado and I partially blame it on budget cuts and partially on being understaffed was I did not communicate the value as well to the middle and lower levels of the organization. Legislators, governors, executives across all of the division areas all bought in. They all got it, they all understood it but where things started to fall apart and when things did stall out in certain areas it was always at that middle layer and we didn't have a great communication strategy and plan because we were just trying to execute and get a lot done. And so in this role, we will have a dedicated communications person and dedicated training education person and a dedicated change management person as part of our program management office so that as we are developing enterprise standards, policies, procedures, et cetera, we've got dedicated people who can work across the various functional areas to communicate, to build awareness, to build the knowledge and skills to actually do the implementation to sell the what's in it for me story that needs to be told over and over and over and over again. Top to bottom as we heard. Top to bottom. Top to bottoms up. Michelin, thanks very much. You're cool and smart and really appreciate you coming on and sharing your great knowledge about the CDO role. A great cap to an awesome two days here so really appreciate your time. Thank you Dave. Keep it right there, everybody. I was talking about smart. I've been at MIT for two days. I just feel smarter hanging out here. All right, keep it right there. I'll be back to wrap. This is theCUBE, Silicon Angles coverage from MIT's Information Quality Symposium. We'll be right back.