 Welcome back to Wikibon Weekly, presented by theCUBE. This week, once again, we've got Bill Schmarzo, who's the EMC CTO of Global Services in their big data practice, talking to us about some of the trends and changes that are taking place in big data as businesses try to do a better job of applying data to better business insights and models that actually can take action. Now, Bill, as we sit here in Silicon Valley, not too far from Stanford and not too far from so many businesses that are in many respects at the vanguard, because the tech industry is very, very big in using a lot of these technologies. Increasingly, we're discovering that people are trying to leverage the value of their data by creating models. So it's not, we're trying to move away from just a single bespoke approach to gathering data, generating an insight, and then starting fresh. We want to start creating models that are capable of supporting not only this action consistently over time, but can be used to support other derivative actions that are tied to that particular decision. How are your customers today starting to think about transitioning from data just as an input to modeled assets comprised of data through analytics that are capable of driving new business behaviors consistently? So we're starting to see more of our customer base starting to build out analytic models. The challenge that is that many of these organizations don't have an overarching vision of what that, what the analytics could do for them. They still treat them as one-off activities instead of as an asset to be utilized and reused, which is resolving in a lot of fairly mature, big data type organizations having what I call orphaned analytics. There was some business need. The analytics were successful. They addressed that business need, but it only was used once and it was never sort of operationalized. And while you get sort of an immediate high from that, it's sort of, you've got some value. You leave a lot of value on the floor because you don't reuse that. So there's, I think organizations who are, some of the more mature organizations are realizing that we've created a bunch of orphaned analytics in lack and overarching vision because they've, while they may have thought about data as an asset, they haven't thought about the analytics as an asset that can be captured, shared, and reused across the organization. Now, this is crucially important because the historical norm of thinking about it as an asset, is an asset has attributes of scarcity. I'm going to take this money and apply it to this use. I'm going to take this machine and I'm going to apply it to this use. I'm going to take this piece of real estate and apply it to this use. And I can apply it to multiple uses at once. Now maybe I can apply a building to accounting and HRR and other types of things, but ultimately that asset has high asset specificities and economists would say. It is specialized to a particular asset. Now the weird thing about data is that it doesn't follow those rules of scarcity. Now it's interesting that historically we have taken, we've made efforts to try to make data scarce so that it looks more like an asset, either by how- That happens within organizations, by the way. What way is he data silos, right? That's exactly right. We use copyright law to protect data, which is okay, we understand that. But also, we don't share data within the business. That this is my data, this is my asset. My source of power. I'm going to use it differentially against you when we start thinking about the politics of the organization. So clearly to make full use of some of these models we're building, we have to start breaking down the barriers organizationally, institutionally, of how not only people think about data, but how they work together to share and move and apply data. Amen. So you've hit on a really key point. When I walk into organization and I see data silos, I know I'm in trouble. Because data silos are the anti-data science, right? It's the data science is about leveraging as much data as possible to get a more clear view on the operation so I can make better decisions. I can build better models. And so when I walk into organization and I see data silos, I know the chances of success there from a broad perspective are very limited. Yeah, they're going to have pockets of success. They'll develop some analytics within their space, but they'll never get reused. And you hit on another key point, Peter, is that data as a currency is very different than dollars and people, right? Think about, you know, dollar. If I have a dollar bill, I go into Starbucks to buy a couple. Well, if I go to Starbucks with a buck, I can't buy a buck. So let's say I got $5. And I go to tip somebody. Yeah, I can tip somebody, but I won't be able to buy anything. So I got $5 to go to Starbucks to buy a cup of coffee. That $5 is used once and then it's gone, right? And the same way with people, right? You can do this, I can do this, but I'm not both here doing this and then, you know, in Peoria, Illinois doing something else, right? I'm only, I've got a transaction limitation. I can do one thing at a time. Data doesn't have that, right? As you've talked many times, data has this network effect that the more data that I can share, the more value it becomes. So it's really, the value of data is knocking down those data silos, which is a huge cultural issue, right? So source of power, the highest paid person's opinion in the organization is gonna hold onto that data. But if I get organizations, we get organizations to realize that data has value, more value when it's used across more parts of the organization, then you start to crack down those silos and you started to really help organizations realize that shared value, that currency value of data. So it suggests, Bill, when you walk in, because you spend a lot of time with customers and generating significant returns on these big data disciplines, it suggests that when you walk in and you encounter data silos, that you personally, and I think this reflects what the businesses make us do, you have a choice. You could either say, accept the constraints and operate within those, within that silo, recognizing that the returns are gonna be commensurately limited, or you can turn to someone and say, you know, there's something we could do here to break down the silo and move. How often do you end up in that conversation to try to get them to see data differently and something to be shared across the business? So I'm not in a missionary cells. So when I walk into an organization where I see the data silos, I don't have enough time in my life to spend trying to help them do that. And we'll walk away. I mean, I think you have to walk away from opportunities that you know you can't be successful. We have very scarce resources on my team, scarce number of data engineers and data architects and data scientists and user experience and visualization people, right? We have a limited set. And so it's my responsibility to our business at EMC, soon to be EMC slash Dell, to make certain that our resources are being focused on those kind of clients where we know we can have the biggest success and drive the biggest opportunities for growth. And it's not like you're hurting for opportunities. You know, they're everywhere. So, but still, I will say that I am more in the business of selling the concepts of moving folks down. I'm sure that when you have gone into some of these accounts and you start going through the process of generating the returns that the client expects, the client themselves periodically go bang. Yes. So tell us about when you've seen those moments and what you've learned from those moments. And if you were in a missionary selling, how you would translate that into going out and trying to tell the world there's a different way of doing things. The best way to get the bing moment is to have them visualize it within their own environment. So we run these vision workshop projects where we typically start with a somebody CIO or somebody in the IT side who sees the chance to knock down these silos, who sees an opportunity to expand their own business wealth by helping organizations to their own organization to get value from data. And so we find somebody like that that we've got a really good chance because they're on a mission to help either themselves or the company to get more value. And so what we do in these vision workshops is we basically show them what they could do. We get a sample of small sliver of their data, our data site to spend a couple of weeks and we show them what they could do with the data. And the idea around that is to make sure that not only does the IT folks know what the realm of the possible is, but more importantly, the business people all start and start realizing if I have access to that data and analytics, I can make magnitude better decisions across the organization. So you've got to, you can tell stories about other companies and what they've done. You can pray to a whole bunch of really smart people in front of them, but it's not real and tell us with their own data. And then they go, oh my gosh, I didn't know I could even do that. And that's when that ping moment happens and we will be in sessions and you will almost literally hear the pings going off. You see, you see somebody in the room do this. I get it. It happened yesterday, I was in this meeting yesterday in Denver and it's a very smart account. They're really good. And we're meeting with, we've had our first meeting which was a CIO and this team and now they brought in the business users. We had supply chain and manufacturing and procurement and logistics are all in the room and we're going through our process and a woman sitting right in front of me, she goes, wait a second and she goes through the scenario of how she would use this data to help them identify their most important young engineers, identify how to keep them, identify, she went through this entire process to identify the data she'd want to do and all the decisions she'd want to make. And I stopped and I was like, you, she got- Would you like to teach the course? Yeah, you got it. And everybody in the room was like, when she said it, I mean, for me, it's like, okay, that's interesting. But when she said it and she put it in context of what they're trying to do, there was like the pings in the room were like ding, ding, ding, ding, ding, ding. It was like, there was like, you know, Vegas and the slot machines going off. So there will be times when somebody, that moment happens, when somebody in the business sort of all of a sudden realizes it, oh my gosh, I can do this to change the way I run my business. Then you've got to form that for success. But it's hard to get people sort of to do that envisioning process to really to invest in themselves, to do that. So- So I'll tell you Bill, the times that when I have had that experience, and spending a lot of time over the years with CIOs and some big data people, but the times where I've seen it is if you can, if you walk into an account, if you walk into a situation and someone's focused on an operational element of their function, so that they have naturally circumscribed the decision, the data that they need and everything else to their silo, you're gonna have a problem. When you can get them to say, well, let's connect this back to the customer. Let's connect this back to the brand. Let's connect this back to the business outcome in the marketplace. Now they may say, ah, that's too big for me and you can understand why they may share that, but when you can start tying them back to those broader questions, the binging starts happening. Or you immediately realize you're talking to the wrong guy. So we like to focus on what we call key business initiatives. So organizations are trying, on the business side are trying to accomplish something. And we like to focus on this nine to 12 month window. And the reason why we like business initiatives is because they tend to be cross functional. So customer acquisition isn't just sales and marketing. It's probably got customer service and might have finance in it. So it allows you to have a broader view of what they need, which is important from an analytic perspective, but from a cultural perspective, it's also important because now you've brought in three or four different parts of the business to help them sort of figure out what they need to do. And so when you target the business initiatives, when we make it all about improving that business initiative, there's a couple of things. First off, get the organizational support, but that probably the most important thing is if somebody in the organization decided that reducing customer term by 10% is what we're gonna go after, somebody in the organization probably associated and put some value around that. That's worth this much to us. And when we can target something like that that's got some sort of known range of value, then all the decisions we gotta make regarding data and analytics and some of the organizational changes become much easier to bear. Yeah, absolutely because there is that concrete statement of what the value is and everything, and you can pull that thread through the entire process. Which, by the way, is the exact opposite way which most vendors sell and most IT organizations approach, right? They go out and they buy some technology, they go get some version of Hadoop, they throw some data on there, they hire some data scientists and they hope that something happens. And even if they find something, it's no guarantee that the business people are gonna respect what they find. Well, that's a great point, Bill, because I do wanna talk about this notion of models and some of the technologies. But at the end of the day, these are just tools. I mean, we could do everything we're talking about on with abacuses and paper. It might take an infinite number of monkeys and an infinite number of times to do it, but you could probably do it. But there are tooling that makes it easier or not. But the two elements of this have to be not only being able to bring the data in, build the models and apply to the decisions, but also from a technology person perspective, understand how all of that is gonna flow to the right person at the right time so the action actually takes place. Yes, yes, you've got to think, even in the early phases of envisioning, you have to think about how do you operationalize it? How, in most cases, we find that the people who are trying to make more effective are humans. It might be a teacher, a physician, a nurse, a parole officer, a engineer, a technician, and we need to think early in the process about how are we gonna deliver the insights to them in order for them to do their jobs better? So the operationalization of the entire process isn't only you wait until the end to do. In fact, you'd like to bring those people into the process early because they probably know things about their work environment that no one else does because they live it every day. And so bringing those people into the process, even in the envisioning part, ensures you that when you get to the operational phase, which is where really the rubber meets the road and all the money is value is realized, you can successfully actually get in there. You have incorporated what is really important to the overall design of the solution, which is someone doing something different. So let's talk a little bit about technology. I know you're not a guy that goes out there and comments on Spark or this or that, any other thing, but you are very, very familiar with the technologies that people are using and the complexity that those technologies are engendering I was not too long ago having a conversation with the head of analytics at the large bank who basically said, I am now in a position where I cannot differentiate all these open source tools because they all look like they're doing the same thing to me, but they're saying they do something different. How is complexity in the tool set starting to constrain or put a cost on this natural flow of moving from decision out to action? So you've raised the problem that we're seeing across a lot of organizations, especially organizations that take a technology first approach. If you start off with a technology first, so you go out and you spend a number of months figuring out which version of Purdue to get, then as soon as you get it in and you get it installed and get it operational, somebody says, oh, we should be using Spark. Oh, okay, so off they go chasing their tail on Spark and then somebody comes up and says, no, you need Flink. Oh, okay, let's go this way. So the challenge is if you think the value in the big data conversation is in the technology, you will forever be chasing your tail because like you said, there's a plethora of new tools. Heck, while we've been sitting here, it's probably been some startup garage two blocks away has created a whole new tool set that we don't even know about yet, right? So what we advocate to our customers, this sounds kind of weird from a technology company, is that look at your technology as being disposable. That if long term, the value isn't in the technology. We talked in a previous segment about what's the intellectual capital of the big data world? Well, it's the data, the analytic models and the decisions you're making. I didn't say technology anywhere. And so as an organization, those are the three things you're gonna keep and grow long term. The technology might come and go. You may have to throw out Hadoop in two or three years because it's replaced by Spark, it gets replaced by Flintstone, it gets replaced by Barney Rubble, right? Who knows, right? So if you think about the technology as being disposable, it'll allow you to embrace the fact that it's not the key asset, but the key assets are elsewhere and hopefully it'll help organizations to focus their resources on the places where the data sustains forever, right? We've got data sets out there that were probably created originally using CICS systems, right? Data sustains, right? The technology doesn't. So the sooner I think organizations realize that technology is only a means to an end and could be treated as disposable, I think the better off organizations are gonna be. Yeah, we totally agree with you. At the end of the day, from our perspective, what customers need to do is emphasize the impacts and then take a look at technology as either facilitating or impeding that impact from a cost and effectiveness, learning, training, but especially time to get the impact. And I think organizations will start focusing more on building out technology frameworks than they are building out technology. And I think the classic example is a data lake, right? Data lake is not a server with a dupe on it. That's not a data lake, right? There's a bunch of really important stuff around it for cataloging and metadata management and data governance and all kinds of things that surround that. And if you think about it as a framework, then when those underlying technologies become obsolete to replace with something else, you've kept that framework and replaced those aspects with things that are better. Right, so you've got a pipeline that you're managing which becomes, quite frankly, is becomes a central element in many respects to the model. That's not to say that the technology is bound to the model or the model is bound to the technology, but understanding how the model is generated from sources, how transformations take place, where the data gets stored, where the data gets located, how the data gets applied is a crucial element of the model. Yes, what's interesting, Peter, is as we think about what's going on in this space here, some of the really leading edge organizations are realizing from a technology perspective, they're more concerned about two really important aspects. Agility, right, the ability to change and de-risking the solutions. They wanna de-risk, if you have a chance to deliver something that's got to have $100 million impact annually because of advanced analytics, the technology you build it on, you wanna make sure that stuff works. And the cost of that technology is a lot less concerned as much as the risk of being successful. So we're gonna see organizations, I think, over time, so realize that the technology, agility becomes everything, this framework becomes everything, but I've gotta be able to de-risk that so I can make sure I can deliver on the potential of these analytics and analytic models. Well, a mission that I've been on, going back to mission selling, a mission that I've been on, especially since the cloud's been introduced, is to not think about elastic technology. Elastic technology says doing the same thing at different scales, but plastic technology that says you can do the same thing at different scale, but as you accrete experience and gain insight, you also can very flexibly adjust what the technology does in a way that it reshapes in a permanent way so that you can carry on in the future and that becomes your new baseline. I think we need to talk about plastic cloud, plastic big data, plastic technology implementations because it provides not only scale, but also reshaping and restructuring. I think about plastic clouds and plastic data lakes, things like that. The first thing that comes to mind is a TV series called Botched, with, I don't know if you've ever watched it, my daughter unfortunately watches it way too often, but it's about plastic surgery going around. Oh! It's pretty scary. It's pretty scary. There's a scarier than a lot of scary movies I've seen. I'm not admitting to watching this show. And I think about how important it is to make sure you know what the end in mind is. And the end in mind is not the technology, the technology just to support that. So you don't wanna have a botched sort of infrastructure you wanna make sure your infrastructure's in place to support that end goal, whatever the business initiative is, so that you have that plastic kind of approach that can be morphed to meet what that need is. Bill, that's an ugly pun intended metaphor. Thank you. All right, so let's close this segment of Wikibon Weekly on the Cube. Again, Peter Burris with Bill Schmarr's OCTO of the Big Data Practice in EMC's Global Services Group from Silicon Valley. Thank you very much and we'll talk to you soon.