 from Cambridge, Massachusetts, extracting the signal from the noise. It's theCUBE, covering the MIT Chief Data Officer and Information Quality Symposium. Welcome back to MIT in Cambridge, Massachusetts, everybody. This is Dave Vellante with my co-host, Sam Cahane, and it's our pleasure to have Gopal Subramanian, who is the Senior Vice President of Fidelity. Welcome. Thanks for coming on theCUBE. Thank you. So this is your first time at the MIT IQ funding day MIT Information Quality Chief Data Officer Symposium. It's a mouthful. Of course, a lot of guys from the Army and the Navy here, so they like mouthfuls. So what's your take on the event? Sort of what brought you here? So I work very closely with Dr. Rapa, who runs the data science practice in North Carolina. He's been on the Chief Data Officer. Yeah, he was in MIT, he was a part of theCUBE. So we work a lot with his office, and we recently hired folks from NC campus. So he invited me to join the panel today, and I'm here. So Michael, actually he was one of the first to actually institute that sort of data science program, and they've done a great success with that. So talk about your role at Fidelity. So I have two roles I play at Fidelity. The primary role is I manage data for asset management. Basically, that supports all the asset classes, the fixed income, equity, high income, and so on. And the second role I play is I actually chair the technology advisory group for data, where we try to bring data together as a single practice for Fidelity, even though we have each division have a data practice. We have built a community of practice. So there are a lot of ideas that come out from our engineering community across the board. How do you bubble that up? Something that we very closely work with our community. So the two roles, one is on the asset management side and the other one is across Fidelity. Are you a de facto chief data officer? We don't have a chief data officer at Fidelity. Maybe the role that I play is one, but we don't have it. We have a enterprise CIO, but we don't have a CDO. You don't report to the CIO? No, I don't report to the CIO. I am part of the asset management still. I report into a CTO of asset management. The CTO of asset management, not the CIO. But that's what I say, two roles. One is reporting into CTO, focus on asset management, but I do the second part, which is basically across Fidelity. We always have these discussions about organizations. Organizations are important and you've certainly seen the CTO role emerge in certain industries, financial services, being one of them. Why, you know, Fidelity is a leader, an innovator, and a lot of people will follow, it's like the NFL, what's Fidelity doing? We'll try to figure that out and do what they do, it was number one. Why no CDO, is it? See, I think the CDO, even in the conference today, we're talking about it, for a CDO role to be successful, and if you look at it, there's only a handful of CDOs across, they mentioned like 300 CDOs or something, the role has to be clear for a CDO to be successful. Otherwise you'll start that role in 18 months, you will fade away. So I think Fidelity is taking a cautious look at this and see if it makes sense for us to actually create a CDO role, and if we create one, what does it mean to the organization and to the individual? So I think there's a lot goes beyond that before we actually create it and we heard it multiple times today. So I think even though we're ahead of technology, we do a lot of good stuff, we are very innovative in a lot of places, this is something that it is going to take its course. Well, in Fidelity, from my understanding of the organization, you got a lot of entrepreneurial spirit in the different divisions, and the heads of those divisions have a lot of energy and speed, and I can see it'd be very difficult for an organization like that that's going in so many different directions and so focused within their different areas to say, okay now, here's the data czar, make it all work together that you almost set up that person for failure. So we do have multiple data czars, if you will, right? Like I said, each division, we have leads there who actually, the one good thing that we have done creating this data technology advisory group, which I chair, is actually bringing those czars together and actually creating a practice that we can share from each other, if you notice, right? So there's, asset management might get a tool for a particular use cases that might be useful for another division, which we were not talking so much. Now we are bringing all of those together and saying, okay, that's a use case, that applies to what we are trying to do. Let's learn from each other. So we are doing a lot of those, not necessarily having one role from the organization perspective. Yeah, and that's a, so you're sharing best practice. I mean, a lot of people can relate to this. Framework, best practices, all of those good stuff. You know, sometimes the whole data thing is amorphous, it's hard to get your hands on a lot of people in our audience so you can understand marketing, right? Marketing's easy. Marketing's hard, by the way, but understanding this example is easy. You got a marketing manager in the division, the division head wants control of that marketing manager. Of course, the corporate marketing wants corporate messaging. They don't want to talk about your products. And there's always that tension. So it sounds like you're taking a practical approach and taking it slowly so you don't fail. That's sensible. Now I want to talk more about this technology advisory group that you've created. So what was the catalyst for that? What was the timeframe for creating that? So we started this, I want to say, when I was 18 months, our enterprise CIO started one of the big programs called Compute Strategy where we focus on infrastructure as a service, platform as a service, and data as a service. So we actually put a strategy together for all those three different tiers and I was leading the data as a service piece of it. That data as a service, that particular strategy led into one of the recommendations is to actually create a data tag because we found that it's not like we do the strategy and we are done with it, we have to start influencing our division to execute on the strategy. And the best way for you to do that is to actually continue on the data as a service that counts until the execution happens. So that's how it was created by our enterprise CIO which I want to say about 18 months back. So essentially, again, I know a little bit about Fidelity, so you're providing services throughout the organization, and infrastructure, platform, and data. Interesting, you think about the framework in the technology industry. It's typically infrastructure as a service, platform as a service, and software as a service. You've got data as a service. Is that a high level transport layer? So there is also software as a service. What we're trying to do with the data as a service, the philosophy behind data as a service is very simple. If you look at the foundation, you build the foundation, data, structures, underlying model, quality, everything is done. That's great. How can we be nimble? If you're not creating something data as a service, then each project might take six months, eight months, 10 months. How can we be more nimble to our business partners and at the underlying insulate your structures in such a way that you can actually provide the service on top of it? So how do you create that layer is what we're talking about data as a service. And now we obviously have to work with the platform as a service, in some cases. You have to work with infrastructure. Tomorrow you go cloud. You have to work with infrastructure as a service. So these are interrelated in one sense, but the idea is how can we three groups work together be more nimble to our business partners. That's the overarching goal. Well, and of course it's tracking what's happening in the industry, technology industry in general. Infrastructure as servers over here, storage over here, networks over here, softwares in stovepipes, the platform as a service. It's sort of, I guess it's SOA has sort of evolved into platform as a service. And those stovepipes have proved, they've sort of outlived their useful life, I take. And so now you're starting to see these things come together. So it sounds like you're taking a holistic approach obviously. So that's a multi-year program that continues, but then the council or the tag was created out of it. And we continue to make sure that from the data perspective, we can provide those kind of services to our divisions. And then of course the whole Hadoop big data theme has been somewhat of a catalyst as well. How has that changed your business? Well, at least now if you look at it, there are some questions that was difficult to answer in the past. We have so much amount of data right now, be it structured, unstructured, whatnot. The goal for our organization is can you get those data so that some of the questions that has been unanswered in the past can we answer those questions, right? So that's what we're trying to do. So we're trying to create a framework where we call a piece of it as a discovery warehouse. So can we get the data in and put the data in a format that we can actually start exploring ASAP so that now the business gets value immediately. And then once they hit gold, we convert the gold into our enterprise warehouse. So you actually move it to enterprise warehouse. Now you actually create the data, the service there through your portfolio applications or whatnot. So we're trying to differentiate between discovery warehouse and your enterprise warehouse. And the big data plays a huge role in the discovery warehouse piece of it. And I'm not looking at specific technology. I'm not saying Hadoop, it could be Hadoop, it could be NoSQL, it could be Cassandra, it could be anything. But the point there is the whole big data ecosystem can provide a discovery warehouse where we can be more discovering stuff and then be enterprise at later. Yeah, I mean, independent of the technologies, there's some things that are pretty dramatically different. It's shipping data, the code to the data as opposed to data to the code, doing things in a distributed manner, leveraging more sort of off-the-shelf so-called commodity components. Things that in the past you couldn't do because you didn't have tools to do that, they didn't work. So you'd go out and you'd buy a box. And if you look at the technology, it's more scale out rather than scale up, right? So with the technology that we have, you can add more processes and you can scale out linear. So my question related to that is in most organizations, I'm sure Fidelity Assembly, you've got your application portfolio, infrastructure supports that, and you've got business processes that tie into that application portfolio. And as the technology infrastructure evolves and becomes more flexible, which is one of the goals, speed to market, time to value, all those things, do the business processes change? Or in other words, if you're scaling out, as euphemistically, do your business processes scale out concomitant to that infrastructure? That's a good question. And we have to do that. The thought process needs to change from the business perspective. There's a lot more automation right now that we can do. So one of the things that we are doing that in asset management right now is as a part of the data modeling exercise from the technology perspective, be it conceptual, model, logical, to physical, we are doing the same thing on the business side as a business process model. We need to think about our business process model, how it has evolved and how it has changed. And that's a constant exercise that business needs to start working with the technology and vice versa. So how does that work? Does the business process guru sort of inform and then the infrastructure and the software molds to that or is it sort of the other way around? No, it's actually, yeah, there are, I mean, we are very fortunate that some of our business partners are really technologies having some case. So we have been very lucky that way. And the big thing about that is they also own some of the stuff and say, hey, I know there's a gap here, let's work together. So we have some leaders who actually step in and say, okay, I want to see this. So they push back on our CTO and others and say, let's work together. I want to create a business process. I know you have the other technique, but let's work on it. So I think we are lucky enough to have that kind of a partners who actually, they embrace technology. They know where the gaps are and they push us. Well, it's true. Fidelity in a lot of ways is a big tech shop. I mean, your products that you develop or their technology that enables all these products and you guys spend a lot of money on technology. Phenomenal job of that. So, well, sort of off topic of the data. So let's come back to the data a little bit. There's a lot of discussion at this event around the sort of, you know, governance. It's the sort of role and this to us, we've been talking about it with Paul and we've talked about theCUBE all week here, last two days, this sort of balance between, you know, governance and compliance and all the edicts versus innovation. I wonder if you could talk about that balance from a data perspective. So that's where I was alluding to some, if you look at it, the whole discovery versus enterprise. So enterprise, you need to have governance. When you're discovering some stuff, maybe you are a little bit loose on that particular piece of the governance because you're trying to find out what are the patterns I have and how can this help eventually with the business. But from the governance perspective, we actually, the way in which we think through the governance is we define by subject areas. For each subject area, we actually have three roles that we actively play. And one is basically a business owner, which is basically coming from the business. Who basically defines the policies? Who should use, who should not use? There are certain fees that we may get, only certain users can use it, so other divisions might not use it. What are the policies? So it's a big role that a business owner plays for that particular subject area. When I say subject area, it could be holdings, assets, it could be portfolio reference, it could be security reference, it could be anyone of those. Then the second role is business steward. Business steward is basically saying if there's a quality problem, data quality problem, or an SLA problem, who should I work with technology to solve that particular problem? So there's somebody who works very closely from the business perspective, so that's a business steward. The third one is basically we call subject area owner, which is basically from the technology perspective. And this person is responsible to understand the data structure, who's using it, how they're using it, how much of reuse is happening. So these three folks come together and help in terms of governing a particular subject area. So that has worked very well for us in creating the structure. Business is part of this particular thing, which makes it easy for us. So that's a strategy that we have used. And then wherever we think that we are innovating, where we need to make sure that we are ahead of the curve on some of that, that's where they do on the discovery warehouse piece of it. And once they figure out how to enterprise it, then they go through this governance structure. So speaking of innovation, you know, Dave was mentioning before, you're kind of like the NFL, you know, always creating what's new. What's your next project? What's your next challenge? Well, there's a lot. As I said, the compute strategy, we're using the compute strategy terminology and the digital space into this infrastructure platform and the data. A lot of innovation goes on in that space. If you notice what we are trying to do I'll give you an example of what TAG does. I'm looking for the technology perspective. We do what is called as an innovation trip. We look at startup companies. We have some problems that we have to solve. And when we go to the startup companies, we try to match these problems to a technology solution. The data tag did one in San Francisco. We looked at 18 startup companies in three days. And in San Francisco, California, we, Palo Alto and San Francisco, we looked at 18 companies. And the goal was to see, okay, we know that in the whole big data space there are some issues that we have and how do we marry that with the proper technology? So we looked at it and then the next step of that is basically come up with a pilot that's going to help that particular solution. Now we are probably looking at two, three years and say there are some companies which probably are doing some work on artificial intelligence and maybe we need to think about it right now so that three years and four years from now we can start leveraging this company. So there are a lot of pockets of this innovation that happens. We do innovation campaigns. And then we do a 48 hour, what we call as an innovation stuff where the first 24 hours is basically defining the problem. The next 24 hours is basically solving that particular problem. So it also helps in kind of a motivation factor for us. So there's a lot of pockets of innovation where people do to look forward in three to five years from the technology perspective based on some business problems that the business are asking us and then see if we marry that. Does that effort oftentimes fidelity will make investments? There's a lot of investment that happens. So you inform that process? Yes, yes. So when we go for these kind of trips we inform saying that these are certain things that you might want to invest and certain things that might be useful for what we are trying to do to solve the specific business problems too. So it's a combination. It's a strategic value, not just an IRR. It's a combination. So that's fascinating that you're going to San Francisco and meeting with all these small startups to learn. Do you think other companies your size are doing that or do you think that's... I'm sure they are doing in some kind of form but we are very active. My boss who is the CTO, he does a lot of these innovation trips actually. Not just in US. I mean the last time he was in, I think last year he was in Europe looking at some small companies and how they do data as a service. So there's a lot of ideas that you can get out of these and then in some cases we work with their products. So it's a combination. It's got to be really easy for you to get a meeting. Without the brand is extremely good. What you find, I mean obviously Silicon Valley is a unique vortex of innovation but you know you're in Boston, New York seems to be bubbling up. New York, yeah. Where do you see? Washington DC. Washington DC. Overseas, I don't know if you guys spend any time in Israel but maybe it's more hardware oriented but where else do you spend time? So you got it. Been to New York, San Francisco, the Bay Area a lot. As I said, I didn't go but some of my colleagues went to Europe. They spent some time in Germany and other places. So they found some good companies out there. As I said they were focusing on data as a service there. Washington DC has a lot of good companies on the security side of it. So if there is an interest on that space there's a lot of, because there's government agency and others where security is a big piece of it. There's a lot of very innovative way that they look at security and that could be an area that if somebody's focusing on a lot of small innovative companies coming out from that area. And we're in a small hub right here in Cambridge as well. There is, so we do a lot here too in Cambridge too. So yeah. What do you think about data as a service? We talked earlier about how it's, we're talking internal service to fidelity is, does fidelity sell data in any way, shape, or form? No, at this point of time we don't sell it. It's more of services rather than data. Yeah. Do you think that's an opportunity for financial services companies in general or is it just too funky and regulated? I think it is very regulated. There's a lot of sensitive information that we have. So it's, see when we are trying to go into cloud and it is more when rather than if, but we work very closely with our information security to make sure that whatever we are trying to do from the cloud perspective is acceptable because there's a lot of, see the biggest thing is fidelity runs on a trust, right? So people have trusted us, the brand and their money. So we have to be very cautious about it before we step in. We are very innovative from the technology perspective and all those things at the same time. We still, we have to face the customers and we have to be very clear about it. So it's a balance that you have to strike. One of the key, the keynote for the conference just with Sandy Pentland talked about this enigma, sort of a new framework, potential framework for internet security inspired by the Bitcoin blockchain. Bitcoin, yeah. And privacy too. And the general topic of security with all this cloud and mobile and big data and social media, is security a do-over for our industry? I think it is. We have a new security officer and he comes with a different perspective. Like I said, it is not if, it is when. So he's looking at it in an open mind and trying to put all the right controls. So we have a right boundary. So we don't cross the boundary, but at the same time, we are secure. So there's an active conversation going on in our community to see what's the best way to move forward. Sometimes it might look like it's paranoid, but I think it is for the right reason, but we have to strike the balance. Well, there's a lot of innovation going on, obviously. It has to be, because the bad guys are really smart. It is, it is. And they're really... It's a matter of time. Yeah, so, interesting. So just sort of closing thoughts on this event, things you've heard, takeaways. Well, this is my first time. I actually enjoyed it. The speakers were great. I get to meet some of them out after the event. It's actually, you can learn a lot from each other. The ideas have been great. Last month, I was with my CIO in one of the CIO conference in Newport. I mean, you get to, when you go to these kind of conferences, there are certain things that you do and you actually learn from others. There are different techniques of doing the certain thing. It is always useful. I really love it. And it's intimate here, you know? It is. We spent a lot of time in Las Vegas and a lot of walking. Too much walking. Go Paul, well, thanks very much for coming. Thank you so much for having me. Thank you. Really appreciate it. A pleasure meeting you. All right, keep right there, everybody. Sam Cahane and I will be right back after this. This is Silicon Angles, The Cube. We're live at MIT in Cambridge, right back.