 It's theCUBE, covering IBM Chief Data Officer Strategy Summit. Brought to you by IBM, and now here are your hosts, Dave Vellante and Stu Miniman. Welcome back to Boston, everybody. This is theCUBE, the worldwide leader in live tech coverage. We're here at the Chief Data Officer Summit that IBM is hosting in Boston. I'm joined by Courtney Abercrombie. Courtney, your title is too long. I'm just going to call you a Cognitive Rockstar. And Alan Crane is here from USAA's assistant vice president at that firm. Welcome to theCUBE, great to see you guys. Thank you. So this event, I love it. I mean, we first met at the MIT Chief Data Officer Conference. You were all over that, networking with the CDOs, helping them out and just really, I think identified early on the importance of this constituency. Why, how did you sort of realize and where have you taken it? It's more important than it's ever been. And we are so grateful every time that we see a new Chief Data Officer coming in because you just can't govern and do data by committee. If you really hope to be transformational in your company, all these huge different technologies that are out there, all this amazing, rich data like weather data and the ability to leverage, you know, social media information, bringing that all together and really establishing an innovation platform for your company, you can't do that by committee. You really have to have a leader in charge of it and that's what Chief Data Officers are here to do. And so every time we see one, we're so grateful. So we just heard from Inderpal Bhandari on his recommendation for how you get started. It was pretty precise and prescriptive, but I wonder, Alan, so tell us about the Chief Data Officer role at USAA. Has it been around for a while? Of course it's a regulated business, so probably more data-oriented or cognizant than most businesses, but tell us about your journey. We started probably about four or five years ago and it was a combination of trying to consolidate data and analytics operations and then decentralize them. And we found that there was advantages and pros and cons of doing both. You'd get the efficiencies, but once you got the efficiencies you'd lose the business expertise and then we'd have to decentralize. So we ended up landing a couple of years ago on what we call a hub and spoke system, where we have centralized governance and management of key data assets, data modeling, data science type work, and then we still allow the various lines of business to have their own data offices. And the one I run for USAA is our Distribution Channels Office for all the data and analytics. And we take about 100 million phone calls a year, about 2 billion web interactions, mobile interactions. We take about 18,000 hours, that's roughly two years of phone conversation data in per day. We take about 50 million lines of web analytic traffic per day as well. So trying to make sense of that to nurture member relationships, reinforce trust and remove obstacles for our customers. So you're supporting the agent systems, is that right? I support the agent systems as well as the digital systems. Okay, and so the objective is obviously to grow the business, keep it running, keep the customers happy, very operational focused, efficient. Okay, and so when you, that's really interesting, this sort of hub and spoke, decentralization gets you speed and closer to the business, centralization gets you that efficiency. Do you feel like you found that right balance? I mean, have you struck it? I think so. I think early on, it was more, we had more cerebral alignment, meaning that it seemed logical to us, but actually once the last couple of years, we've had some growing pains with roles, responsibilities, overlap, some redundancy, those types of things, but I think we've landed in a good place. And that's what I'm pretty proud of because we've been able to balance the agility with the governance necessary to have good governance in place, but then also be able to move at the speed of the business point. So Courtney, one of the things we heard, one of the themes this morning, within IBM, it's the role of the Chief Data Officer's office is to really empower the lines of business with data so that you can empower your customers, is what Bob Pigeano was telling us with data. So how are you doing that? Do you have new services? Do you have processes or how is that all working? Right, we do, we have a lot of things, actually because we've been working so much with people like Allen's group who have been leaders quite frankly and in establishing best practices on even how to set up these hub and spokes. A lot of people want to talk to the CDO and they've spun off even a lot of CDOs into other organizations in fact, but I mean, they're really a leader in this area. So one of the things that we've noticed is the thing that gives everybody the biggest grief is trying to figure out how to work with unstructured data and all this volume of data. It's just insane. And just like I was saying in the panel earlier, only about 5% of your actual internal data is enough to actually create a context around your customers. You really have to be able to go with all this exogenous data to understand what were the bigger ramifications that were going on in any customer event, whether it's a call in or whether it's a, I'm not happy today with something that you tried to sell me or something that you didn't respond to fast enough, which I'm sure Allen could equate to. So we have this new data as a service that we've put together based on the way the weather company has put their platform together. We're using a lot of those same kind of like micro services that you saw Bob put on the screen, everything from, I mean, open source, as much open source as we can get. And it's all cloud based. So, and it's ways to digest and mix up both that internal data with all of that big voluminous external data. So Allen, I'm interested in, so you got the organizational part down, at least you've settled on an approach. What are some of the other big challenges that you face in terms of analytics and cognitive projects in your organization and how are you dealing with those? Well, to take a step back, USAA were a financial services company that supports the military and their families. We now have 12 million members and we're known for our service. And most of the time, those moments of truth, if you will, where our service really shines, has been when someone talks to us on the phone, when those member service reps are giving that incredible service that they're known for. And the reason being is that the MSR is the aggregator of all that data. When you call in, it's all about you. There's two screens full of your information and the MSR is not interested in anything else but just serving you. Our digital experience is more transactional in orientation and it's more utilitarian. And we're trying to make it more personal, trying to make it more how do we know about you? And so one of the cues that we're taking from the MSR community through cognitive learning is we like to say the only way to get into the call is to get into the call. And that is to truly get into the speech to text, then do the text mining on that to see what are the other topics that are coming out that could surface that we're not actually capturing. And then how do we use those topics at a member level to then help inform the digital experience to make it more personal? How do I detect life events? Our MSRs are actually trained to listen for things like words like fiance, marriage, moving, maybe even a baby crying in the background. How do we take that knowledge and turn that into something that machine learning can give us insights that can feed back into our digital transactions? So, this is what our group is doing. It's a big task. So how are you doing that? I mean, it's obviously we always talk about people process and technology. Break that down for us. I mean, how are you approaching that massive opportunity? Part of it is, I look at it as like a set of those Russian nested dolls. Every time you solve one problem, there's another problem inside of it. The first problem is getting the access to the data. And where do you store? We're taking in two years of data per day of phone call data into a system. Where do you put all that? And then where do you put a week's worth, a month's worth, a quarter's worth of data like that? Then once you solve that problem, how do you redact all that personal information so that that private information that you really don't need, that data exhaust that would actually create a liability for you in our world so that you can really stay focused on what are the key themes that the member needs. And then the third thing is now that you've got the access to the data, it's transcribed for you. It's been redacted from its PII type work. Well now you need the horsepower of analysts. And we're exploring partnerships with IBM, both locally and in the States as well as internationally to look at data science as a service and try to understand how can we tap into this huge volume of data that we've got to explore those types of themes that are coming up. The biggest challenge is in typical transaction logging systems, you have to know what you're logging, you have to know what you're looking for before you know where to put the data. And so it's almost like you kind of have to already know that it's there to know how much you're querying for it. And what we need to do more is we pivot more towards machine learning is that we need the data to tell us what's important to look at. And that's really the value of working with these folks. So obviously the data is increasingly unstructured. We heard this morning, whatever 80, 90% is unstructured. So you're, I don't know, whatever, you're putting it into whatever, data lakes, swamp, ocean, et cetera. Everything, everywhere. And you're using sort of machine learning to both find signal, but also protect yourself from risk. So you've got to, you said, you got to redact private information. So much of that information could be in no schema. Absolutely. Okay, so you're, where are you in terms of solving that problem? You're in the first inning? Are you deeper than that? We're probably, I would say beyond the first inning, but we, so we've kind of figured out what that process is to get the data and all the piece parts working together. We've made some incredible insights already, things that people, you know, I had no idea that was there. But I'd say we still have a long way to go. It's particularly in terms of scaling, scaling the process, scaling the analytics, scaling the partnerships, figuring out how do we get the most throughput. I would say it's one of those things where we're measuring it on maybe having a couple of good wins this year, a couple of really good projects that have come across. We want to kind of take that to about 10 projects next year in this space. And that's how we're kind of measuring the velocity and the success. Data divas. Yeah, data divas, okay. I walked in a little late and there was one of the, so breakfast this morning, data divas, you hold this every year at this event. Yes, we do. It's growing. Now we got data dudes. I was one of the few data dudes there. Yeah, we were great. Walked in with one of the women chief data officers and said, I got no problem with people calling me a BI. I was like, whoa, I just, I'll sit down. Real quick, real fast. So what's the intent of that? What are the learnings that you take out of those? I think it's more, it's, you know, you could honestly say this isn't just a data diva problem. This is also, you know, anybody who feels like they're not being heard, it's real easy to get drowned out in a lot of voices when it comes to data and analytics. Everybody has an opinion. I think, remember Ursula is always saying all's fair and love more than data. And it feels like, you know, sometimes you all come to the table and whoever has the loudest voice and whoever bangs their chest the loudest kind of wins the game. But I think in this case, you know, a lot of women are taking these roles. In fact, we saw, you know, a while back from Gartner, that number about 25% of chief data officers are actually women because the role is evolving out of the business lines as opposed to more IT lines. And so, I mean, it makes sense that, you know, we're natural collaborators. I mean, like the biggest struggle in data governance isn't setting up frameworks. It's getting people to actually cooperate and bring data to the table and talk about their business processes that support that. And that's something that women do really well but we've got to find our voice and our strength and our resolve and we've got to support each other in trying to bring more diverse thinking to the table. So it's all those kinds of issues and how do you balance family? I mean, we're seeing more and more, you know, I don't know if you know this but there's actual statistics around millennials and that males are actually starting to take on more and more role of being the caregiver in the family. So, I mean, as we see that, it's an interesting turnabout because now all of a sudden it's no longer, you know, women having that traditional role of, you know, I got to always be home. Now we're actually starting to see a flip of that, which is, you know, I think it's kind of welcome. My husband's definitely, I say he's a better parent than me. It's honest. He'll watch this and he can thank me later. It was a great discussion this morning. Alan, want to get your feedback on this event and also you participated in a couple of sessions yesterday, maybe you could share with our audience some of the key takeaways, the event in general and the specific ones that you worked on yesterday. Well, I've been fortunate to come to the event for a couple of years now and when we were just, what, 50 or so of us that were showing up. So, you know, I see that the evolution just in a couple of years time, conversations have really changed. First meeting that we had, people were saying, where do you report in the organization? How many people do you have? What do you do for your job? There were very different answers to any of that. Everywhere from, I'm an independent contributor that's a data evangelist to, I run legions of data analysts and reporting shops, you know, and so forth and everything in between. And so what I saw first year was really kind of a coalescing of what it really means to be a data officer in the company. That actually happened pretty quickly in my mind when by seeing it through the lens of my peers here. The other thing was when you think about the topics, the topics are getting a lot more pointed. They're getting more pointed around the monetization of data, communicating data through visualization, storytelling, key insights that you, you know, using different technologies. And we talked a lot yesterday about storytelling and storytelling is not through visualization and storytelling is not just about like who has the most, you know, colors on a slide or animation of your bubble charts and things like that. But sometimes the best stories are told with the most simple charts because they resonate with your customers. And so what I think is it, it's almost like kind of getting a back to the basics when it comes to taking data and making it meaningful. We're only gonna grow our organizations and data and data scientists and analysts if we can communicate to the rest of the organization our value and the key to creating that value is they can see themselves in our data. Yeah, the Viz as we like to call it sometimes is critical to that storytelling. Sometimes I worry and we go on to these conferences and you go into a booth and say, look what we can do with machine learning and you'll just be looking at just this data and say, what do I do with all this? What do I do with all this, yeah. I don't know how it would make sense of it. So is there a special storyteller role within your organization or are you all storytellers? Do you cross train on that or? It's funny you'd ask that one of the gentlemen of my team, he actually came to me about six months ago and he says, I'm really good at the analysis part but I really have a passion for things like Photoshop, things like the various video and video editing type software. He says, I want to be your storyteller. I wanna be creating a team of data and analytics storytellers for the rest of the organization. So we pitched the idea to our central hub and spoke leadership group. They loved it, they loved the idea and he is now oversubscribed you would say in terms of demand for how do you tell the data, how do you tell the data story and how it's moving the business forward? And that takes the form kind of everything from infographics to also about how do you make it personal when now 7,000 MSRs have access to their own data? Really telling that at a very personal level almost like a vignette of an MSR who's now able to manage themselves using the data that they were not able to have before where in the past only managers had access to their performance results. This video actually pulls on the heartstrings but it not only does that but it really tells the story of how doing these types of things and creating these different data assets for the rest of your organization can actually have a very meaningful benefit to how they view work and how they view autonomy and how they view their own personal growth. That's critical, especially in a decentralized organization at least the quasi decentralized organization getting everybody on the same page and understand what the vision is and what the direction is. So often if you don't have that storytelling capability you have thousands of stories and a lot of times there's dissonance. I mean I'm not saying there's not in your organization but have you seen the organization because of that storytelling capability become more agile, at least more sort of effective and efficient at moving forward to the objectives? Well, as a data person I'm always biased that data can win an argument if presented the right way. It's the challenges when you're trying to overcome or go into a direction and in this case it was we wanted to give more autonomy to the MSR community. Well the management of that call center we're a 94 year old company and so the management of that call center has been doing things a certain way for many, many, many, many years and the managers having access to the data, the reps not, that was how we did things. And so when you make a change like that there's a lot of hesitation of what is this gonna do to us? How is this gonna change? And what we were able to show with data and through these visualizations is you really don't have anything to worry about. You're only gonna have upside in this conversation because at the end of the day what's gonna empower people is having access and power of their own destiny. Yeah, access is really the key, isn't it? Because we've all been in the meetings where somebody stands up and they've got some data point and they're pounding the table. Oftentimes it's a man. And he's a powerful P and L leader and jamming data down your throats and you don't necessarily, the poor sap that he's beating up doesn't have, the target doesn't have access to the data. This concept of citizen data scientist begins to level that playing field, doesn't it, are you seeing that? It does and I wanna actually come back to what you were saying because there's a larger thought there which is that we don't often address and that's this change management concept. I mean we look at all these, I mean everybody looks at all these technologies and all this information and how much data can you possibly get your hands on. But at the end of the day it's all about trying to create an outcome, some joint outcome for the business and it can be threatening, it can be threatening to the C-suite people who are actually deploying the use of these data-driven tools because it may go against their gut. And oftentimes the poor messenger of that when you have to be the one that stands up and go against that senior vice president's gut, the one who's pounding and saying, no but I know better, that can be a tough position to be in without having some sort of change management philosophy going on with the introduction of data and analytics and with the introduction of tools because there's a whole reframing that, hey my gut instinct that got me here all the way to the top doesn't necessarily mean that it's gonna continue to scale in this new world with all of our competitors and all these massive changes going on in the marketplace right now, my gut's not gonna get me there anymore. So it's hard, it's hard. And I think a lot of executives don't really know to invest in that change management that goes with it, that you need to change philosophies and mindsets and slowly introduce visualizations and things that get people slowly on board as opposed to just throwing it at them and saying here, believe it, mm. Well such a thing, I mean, it wasn't that long ago, certainly this millennium where publications like the Harvard Business Review had cover stories on why gut feel beats analysis by paralysis. That seems to be changing. And the data purist would say, the data doesn't lie. As long as you can interpret it correctly, let the data tell us what to do as opposed to trying to push an agenda, but there's still politics involved. There's just things out there that you can't even perceive of that are coming your way. I mean, like Blockbuster Netflix, Alibaba versus Standard Retailers. I mean, there's just things out there without the use of things like machine learning and being comfortable with the use of things like machine learning. A lot of people think of that kind of stuff as, ooh, don't get your hoodoo voodoo into my business. I don't know what that algorithm stuff does. Bitcoin. Yeah, I mean Bitcoin. What the hell is this? And now look what's happening. Yeah, it's coming and you need to get ready. There's an important role, though, I think, of instinct. You don't want to dismiss a 20-year leader in a particular operations because they've gotten themselves where they're at because in large part, maybe they didn't have all the data, but they learned through a lot of those things. And I think it's when you marry those things up and if you can bring in a kind of humble way to that kind of leader and win them over and show how it may be validating some of their points or maybe how it explains it in a different way, maybe it's not exactly what they want to see, but it's helping to inform their business. And you come into them as a partner as opposed to a gotcha, you know. Then you can really change the business that way. And what is it? The limbic brain just doesn't feel right? Is that the part of the brain that tells you that and informs you that? And so it's hard to sometimes put, but you're right, Alan. There is a component of this, which is gut feel, instinct, and it probably relates to experience. So it's like when Deep Blue beat Gary Kasparoff, we talk about this all the time. It turns out that the best chess player in the world isn't a machine, it's a human and a machine. That's right, that's exactly right. It's always the people training these things. That's where it gets its information. So at the end of the day, you're right. It's always still instinct to some level. All right, cool, we gotta go. All right. Last word on the event, you know, what's next? I love my Chief Data Officer, I miss you guys. It's good to be here. It's good to be here, yeah. It's good to be here, we appreciate it. All right, we'll leave it there. Thanks, you guys. All right, thank you. All right, keep it right there, everybody. This is theCUBE, we're live from IBM's Chief Data Officer Summit in Boston, right back. My name is Dave Vellante.