 Hi everybody, we're back. This is Dave Vellante with Jeff Kelly. We're from Wikibon. We're here live with theCUBE at MIT. We're at the MIT Information Quality Symposium. This is really where practitioners get together and try to figure out, okay, we have all this data that's sprawled all over our organization. Our systems are siloed. How do we put forth some kind of data architecture, data governance? How do we develop an information quality strategy? And how do we evolve that over time? We're learning very quickly that it is a journey. It is not a project with a beginning and an end. It's something that evolves. It's organic. It lives with your organization. And Dat Tran is here. He's the Deputy Assistant Secretary for Data Governance and Analysis at the Department of Veterans Affairs. And we're going to unpack a lot of the information here and the issues that we just discussed. Dat, first of all, welcome to theCUBE. Really appreciate you taking the time here. Thank you. Spend with us. Well, you told a funny story that unfortunately, we really apologize for this in Boston. We couldn't get a hotel room. Our crew was 20 miles away. You got offered a hotel room in New Hampshire and decided, well, I'll be better off staying at the airport and pulling an all-nighter and working and maybe taking a catnap. So how do you feel? I'm feeling great when it comes to data and governance and business intelligence. I'm all like, wrap up, and I can keep going for days. Well, as a person who flies a lot of red eyes, I can tell you, you're on a train right now by three o'clock. You want to be here at bed, so. Yes. So anyway, the VA, you said you gave a great keynote today. You said the VA is the second largest agency behind the DoD. A lot of people know that and you're managing a lot of data. Talk a little bit about the VA and your role there. Okay? VA, like you said, Dave, we are the second largest federal agency, cabinet level agency in the government, US government. We have multiple business line. We basically like conglomerate. If you're looking at this as a private sector, we have a different business line ranging from healthcare to home loan to education service and memorial service. Big data is not something that is new to us. We have been around for since 1930 up until now and over the years we have built many systems and we have many different business line that have different requirement for data. So big data have not been a new thing to us, but what we are really looking at trying to shift to toward VA is to create a customer-centric data environment and that is the effort that currently going on in VA that require us to govern data from people process and technology standpoint and that also will impact how we do business intelligence and also how we improve all information quality. I stole your line dad at my opening. It was big data, big deal. You guys have a lot of data, 11 petabytes, more than 11 petabytes I believe. Yes, yes. Okay, and so the other thing I really liked about your keynote is very transparent. A lot of executives get up, oh yes, we have this data architecture, everything's integrated, it's all wonderful and then when you really peel the onion you find out it's not so integrated. You have a lot of systems, like you said, you've been around since 1930, you've been developing systems in the 70s and the 80s and the 90s, different computing architectures and so you've got literally hundreds if not thousands of applications and systems all with different data sources. Challenging environment, please describe that environment for our audience a bit. We in the VA have a very large information, infrastructure and system that like I said earlier we have a different business line and some of the system have been built since the 70s and the 80s and many of our system from across the business line don't interface or interoperate with each other. So one of the things that we're trying to achieve through VA is this, we are looking at data from a little bit different perspective rather than looking at data from a business, from a business line or system perspective, we are approaching the data from a customer perspective. So currently we have a major effort undergoing in VA starting this year where we are looking at shifting from a fragmented information environment to work toward what we call a customer data integration effort where we will be able to have a 360 degree view of the customer and once we get that information in an integrated fashion, the rest of the information that related to the customer will be able to see who use VA for what benefits and services and understand a lot more about their behavior in term of their usage of various VA business and services. So you said in your keynote, you're not alone, you talk to a lot of CIOs and other leaders in commercial enterprise that have the same problem because as a consumer we can all relate to this. You call up, for instance, your financial institution one day and they transfer you somewhere else and they don't have the information at the fingertip and that's a common challenge. So you talk about this whole customer data integration. Do you see this as a situation where all your systems are actually going to be over time able to share the same customer data, that true 360 degree view and how long will it actually take to get to that vision? And that's a very good question. It is a long term vision that's what we're trying to get to VA. One of the things that we're looking at from a customer data integration standpoint is that we need to get the core common customer data, the identity information, of course, in order for them to use VA, they have to be a veteran or they have to tie to a veteran or a service member. So if we get the core information correct, that's the way we look at it, to build that core piece correctly, what we call the center piece, then we can then link the rest of the information across the enterprise to that core piece. So for example, what we envision is down the road in the next couple of years, we will have a authoritative customer, common customer data, that then information on healthcare, home loan, education benefits and so on can be linked to that core common data so that at any time, any place, anywhere, that if we would like to know how many veterans use in certain VA benefits or services, we will be able to retrieve the same information from the same authoritative source. Excellent, so I want to come back to that and sort of how you applied that, but my colleague, Jeff Kelly, is going to jump in here. He's our lead big data analyst, follows a lot of the technologies that we're talking about, but came from the so-called traditional world. So Jeff, why don't you take it from here? Sure, so yeah, one of the things that struck me in your keynote was you talked about the data silos that you've got in the organization and having covering this world for many years, that's a problem we've been struggling with for quite a while. I wonder if you could delve a little bit more into some of the, not just the technology, but the people in process issues related to breaking down those silos, because, again, having covered this, a lot of the issues revolve around who owns the data and people get very attached to their data sources, sometimes don't want to share them, can lead to meetings half the time being spent arguing over who has the right numbers. How do you address not just the technology, but the people in process issues related to breaking down those silos? The number one thing when we look at how to approach this customer data integration, rather than looking at it from a data and system perspective, we look at it from a business process perspective. So we're looking at what is the mission of the department, we're looking at what it is that we try to deliver in term of business and services, and we work backward from there and identify where do we need to capture the data to support each step of the business processes across VA. And by doing that, we can quickly find out where we have multiple sources of authoritative data that we need to integrate and consolidate, but also at the same time, because it is business driven, we know what data is critical to the business process and what data is basically just not necessarily non-value added, but not critical to the mission, so we can decide the priority of how we approach the integration aspect. The other thing that we look at is that in order to change the mindset of a multi-faceted organization, is the number one thing that must be changed is the culture, and that's the most difficult part, but rather than deal with it from an individual organization perspective, the way we look at it is we bring up the value proposition of what is in it for our clients if we have a customer data integration environment. So when you start talking about what is in it for the veterans and the family and the service member, that is where folks really understand that what we need to do to support the client is more important than anything that have to do with our individual organization, and that's the reason why we approach it from a business process standpoint, but also from a what is in it for the veterans and the family member at the end of the day? You showed a little insert, a little picture in a picture on your slides of the business process map, and you can't really read it because it's obviously the VA's business process map is quite complex, I'm sure. How did you actually document that? It didn't just fall out of the sky, I mean you presumably had to really think through the business processes, the interdependencies, how you actually organized that taxonomy. How did you go about that? What we basically did was we approached it from the business process by reaching out directly to the folks who really understand the benefits and service delivery that we need to do in VA. Like I said during my speech today, that when we look at business people, there are business people with data background, business people with IT background, and then there are the business people who truly understand the business processes. So what we did was we have a team that create a standard interview template where we reach out to over 100 subject matter experts across the VA business line, and we document every single business process, and then we go back and reach out to the subject matter experts that understand the data in the business process and capture the metadata that are beneath the data that are being used to support the business process, and we look at it from the standpoint of in the business process, you have the actors, and then you have the process staff themselves. So we identify who is doing the work or who is responsible for that business process, and in some case, the actor could be the veteran who might be required to provide some information to VA, and so we identify the actors, and then we identify the processes and the information system, and then we link them all together with the metadata supporting the business process that are being used across VA. So you essentially created a business process map and connected that to a data flow? Yes. And so you can now see where the data flows throughout the organization, what business processes are affected by that data, where there's redundant data and potential areas of risk or error, and then what's the next step? How do you act on that information? Okay, what we do is when we're looking at the business process, at the same time, we have a group of SMEs that start looking at what is the future vision for VA to be an integrated service delivery and benefits delivery organization to our veterans? And so from that future vision, then we were able to identify from the business process standpoint where we have the overlaps but where we also have gaps of information that we might currently not capture that we require the veterans to go out and seek. So using the future vision of where we need to be and the current business process, identify the overlap and the gap, what we are ultimately going to be able to do is to reduce or eliminate the overlaps and utilize the resources that are being used today to create all of these silo overlap and address how the gap will be filled in terms of data and information. And you're talking about the overlaps and data. Now, what about the business process transformation? Is that something that, I mean, obviously it's not your primary role, but does that come out of an exercise like this? Actually, it's a natural fit for something like that to come out of this effort of customer data integration. One of the things we learned that when it comes to customer data integration, it is only a symptom. The problem and the root cause is actually the business process and business integration that we need to transform, to apply and transform the department toward a more integrated business approach. So coming out of this CDI effort, what we are going to see is that we will address not only about the data piece, we will also address about the process, the governance and the data management piece in term of the roles and responsibility and the people, the culture and the process. And all of this are being put together in what we call an integrated roadmap on how to achieve a veteran centric or data environment. So I wonder if we could take that to its kind of logical conclusion. Ultimately, you're trying to improve services for veterans. So maybe could you kind of put some context around how are the improvements you're trying to achieve with CDI in a really customer centric view? How is that going to translate to supporting the veterans and the various services you supply in a better way? What will their experience look like when you've maybe five years from now and we're further down this road versus what it looks like today? Okay. Our clients are basically coming out of the military. So that information is currently captured by the Department of Defense. In order for you to become a veteran, you have to serve in the military. So we ultimately, we will know ahead of time before someone potentially will become our client. So the way we envision it is that someday when we get, when we have a CDI, a customer data integration environment in place, is that we will reduce the burden on the veterans or his or her family to have to come up with proof of service, to have to come up with what we call demographic and contact and military service history and come to VA and apply for benefits or services. That when we have a customer data integration environment, it will enable us to have that information early and upfront that will allow us to proactively provide benefits or service or at least will reduce the burden on the veteran to have to come up with that information themselves and provided their proof of service when they come to VA. Number one, number two, our vision is that someday when we have the CDI environment in place, we will enable self-service. Where veterans can come and retrieve his or her information on his or her benefits usage in VA and be able to access that information or will be able to update his or her personal information or contact information online. Today, we are not quite there yet, but what we've seen is the reducing of the burden and also will help improve the business process long term that will enable us to be able to more efficiently effectively deliver benefits and service in a short period of time frame. You had some great one-liners in your discussion today. One I love is don't try to kick in the door in one shot. You might break your foot and you were talking about communication and management by and you can't do what you do and take on such a large initiative without making some mistakes. What would you do differently if you had to do it all over again and what advice would you give your fellow practitioners? The biggest lesson that we have learned is that it's not always about technology. The people factor is the hardest to change and the people factor involved, not just folks that are in senior position, but in order for us to really transform VA, we have to be able to pull ourselves together as a one VA approach and that's included from the senior leadership all the way down to the folks who are doing the work out in the trench. And so one of the things that we push for is to help shape the culture and change the culture by messaging. One of the most important things that we have learned is strategic communication. Oftentimes we take it for granted that we communicate after we come up with a solution or we communicate after we already come up with an idea or a plan. One of the things that we learned that have been very effective for us is we communicate early and often in the process. At the end of the day, although we are data folks and we might not be used to it, but at the end of the day, marketing is very important. And when I say marketing, that's in line with what we call strategic communication. Have an idea, you have to have the value proposition and you have to be able to communicate that at all level from all different perspective so that the technical folks would understand it, the business folks would understand it, what I'd like to call the pedestrian term. If you can communicate your idea and what you're trying to do in pedestrian term when anyone who look at the business plan or the operating plan and say, got it, that's right on target and that is success. That's great advice and certainly we deal with a lot of IT folks and IT folks by their very nature are not marketing folks. So they tend to not communicate early and often and so I think the advice really applies throughout the organization, especially in those roles where you're not used to going out and promoting yourself, so fantastic. Well, Dad, I really appreciate you coming here today and being here at the MIT Information Quality Symposium. It was really a pleasure meeting you. Thank you, Nick. Thank you, Jeff. All right, keep it right there, buddy. We're going to be right back with Michael Nix who is the Director of Analytics at Fletcher Allen Healthcare. This is theCUBE, I'm Dave Vellante here with Jeff Kelly. You can tweet me, I'm at D. Vellante. He's at Jeffrey F. Kelly. Keep it right there. This is Silicon Angles theCUBE. We'll be right back.