 Live from the Fairmont Hotel in San Jose, California, it's The Cube at Big Data SV 2015. Okay, welcome back everyone. You are watching The Cube live here in Silicon Valley for Big Data SV. This is our second event in Silicon Valley with fourth overall. We had two events in Big Data, New York City NYC, Big Data NYC and now Big Data SV. We go out to the events, we start to see if we can always have our own event in conjunction with Stratoconference and Hadoop world going on this week. Of course, we're bringing all the cube magic to you, all the data, extracting the signal and the noise. I'm John Furrier with Jeff Kelly. Our next guest is the esteemed Bill Schmarzo, author of The Big Data Bible, we're calling it, but a friend, Cube alumni, congratulations on your success with your book. Thank you. Thank you. And of course, I wrote the inside cover quote. That's why it's popular. Otherwise, it's a winner again. No, seriously, you wrote this book, book on the plane, on a whim. And it's just taking off. People would like this book because you just were like practical advice. Let's talk about the success of the book and just give an update and some of the anecdotal things you've heard. So the book has, the book's boring, right? It's a very pragmatic how to guide and how to sort of head down the Big Data path, right? And so it's full of exercises and techniques and the kind of things we do on consulting projects that really helps somebody get started. In fact, it's actually being used by a couple of universities at the textbook now. Seton Hall is using it and I'm actually teaching a class now at University of San Francisco called the Big Data NBA. And this is used as kind of the textbook for that. So when we called you the Dean of Big Data, it was more of a terms of endearment. We love you, you're like, hey, you're the guru and we know each other, this was kind of a fun term. But now you actually are the Dean of Big Data. Your book is being used in the university. I think I found a second career, right? When I get tired of flying on airplanes all the time, maybe I can become a professor somewhere, the professor of Big Data. Yeah, we'll throw in some speaking gigs too. You know, they got great retainers. You know how they just write it on the book. Necessarily, so it's been fun. You have any good experiences so far with customers with this book? Yeah, I was in a, and I had a client meeting recently and I walk into the meeting, sit down and the CIO walks in and he hands his book and he goes, this book is the Bible. Every major organization is reading this book. It is how we're attacking things. And I was like, what book is that? Oh, it's my book. Oh yeah, very great. And he said, because it lays down in a very pragmatic fashion, how to start. And it doesn't start by saying what technology you need to go after, because to be honest with you, the minute this book was written and published on the technology side, it was outdated, right, it was outdated. So it frames a problem from a business perspective, which gives you the guardrails. The guardrails that allow the organization to make decisions about what data should I go after? What sort of analytics am I going to need? What's the technology foundation for this thing? And it allows organizations to kind of step at the same one step at a time as they build out their big data strategy. This is the book, Understanding Big Data, published by an amazing job by Wiley, great publisher of books. You got to check it out. Wiley's amazing. Great, great books. And Bill, so what's now, I mean, what I'm really, first of all, I'm impressed that you got the book done because one, it's one, you travel a lot. And you told the story on theCUBE earlier a couple of years ago. You just started banging out chapters because you're on a flight and you're just bored. I actually did two chapters in sitting in the airport in Kansas City. We were had like an eight hour delay and I was right and I was just like, okay, just start cranking these things out, right? So you're sitting in airports and in hotel lobbies. I've seen every airplane movie, right? So what else are you going to do? You're right, exactly, you're right. So what's going on now? Customers, what's changed? Now that the book's out there, obviously it's a great book. People are using this reference point for a big data journey. What's now chapter two of your journey? What are you up to now? Where are your customers? And what are some of the things you're seeing? So the book, I think, from a how-to framework perspective still works. What we're learning is we're learning new techniques. There's three exercises we end up developing as part of this class I'm teaching at San Francisco and they're exercises we now incorporate into our approach we do with customers. One of them is around how do you do buy analysis, B-Y analysis to help uncover new sources of data that you may want to go after. And so we're using the San Francisco as my Petri dish as my class there as a way to tease out. You could do big data yoga. Yoga. Yoga and taking these exercises, like stretch the data, move the data around. There you go, we have a lot of mm, mm, mm, yeah. What have you learned? So what have you learned now? What's the big aha for you this time? Because obviously a lot of stuff's going on in the industry. You have the pivotal thing happening with this big data platform, open to platform alliance, all this stuff happening. What's going on with the technology solutions? Well, from a technology perspective, Hadoop is one. All right, no one talks about if Hadoop, it's how I use Hadoop. And something that we had talked about years ago and my experience at Yahoo, we knew the technology was solid and too many folks were spending time doing proof of concepts on the technology instead of proof of value on how to get value on the technology. And I tell organizations don't do a proof of value, a proof of concept on Hadoop. I mean, Google, Yahoo, the NSA, they're all using the technology. Trust me, the technology works. The challenge organizations have is not with the technology. The challenge organizations have is, where and how do I start? How do I ensure that I've got organizational buy-in? How do I ensure that I'm focused on area where I can drive a positive ROI and get a payback in nine to 12 months? The opportunities inside of accounts, and to every account we've talked to, the problem is never lack of opportunities. It's always too many. And so the part of the process is, how do you figure out where to start to build that roadmap so you basically take and knock off one use case at a time and as you're knocking off those use cases or building out your analytic foundation, you're building out your intellectual assets? That is, it still works, it still works and Hadoop is a foundation for that. So how do you get the business involved? Not just from a, you want to find the business use case but then you got to get the rest of the organization on board and get that buy-in. How do you go about doing that? So we run this thing called a vision workshop and it's not a one-day exercise, it's actually a two-week engagement where we come in and we work with a client to figure out what problem they're trying to solve. So we're doing this project for a hospital chain in Denver. They're key business initiatives around hospital-acquired infections, staff infections. People come in for some kind of treatment, they get staff infections, they spend more time in the hospital, they cost the hospital money, sometimes they even die. So how do we reduce hospital-acquired infections? So if that's the problem we're going after, what we do is we work with the key business stakeholders who are involved with that. We bring them in and the two things we do with them that just, it's so simple. For the problem they're going after, we want to understand what decisions you're trying to make and what question you're trying to answer to support that decision. Now the reason why decisions are really important because we're going to actually build analytics around each of the decisions. Not just predictive analytics but prescriptive. We want to figure out when we talk to them about not only what are you trying to uncover but what actions are you trying to take from the decisions and how do we deliver recommendations or prescriptive analytics to help make the human more effective in the process. And so we go through that process, we bring them in early, they realize we're there to support the decision-making process and then we do some good things. We build a simple mock-up to show them how this might render itself within their mobile app or in their web app or in their call center or whatever the right environment might be. And that really helps them to see how the, begin with an end in mind, it really shows them that here's where we're going to go. Everything we're going to do along the path to get there as far as the decisions, the questions, the data, the analytic models, that the technology eventually is going to come in here and they don't care about the technology. But at the end you're going to get a smart app that tells you that when there's a snowstorm, the first snowstorm of the year in Denver, if you get between three to four inches, the number of ER visits increases by 27%. What's interesting, they say, oh yeah, we knew they had increased it. We knew it increased, right? We didn't know 27%. Now I know exactly how many more nurses to hire, how many more doctors I have on hand. It's in some cases all we're doing is quantifying the hunches that they have. And that's why it's important to bring them in early because they know that decision they're trying to make and they have those hunches. We're going to use analytics to either prove or disprove them. But nowhere in that conversation did the word Hadoop come up. Not once. And they could give a holy hoot about the technology. Hadoop, Spark, MapReduce, Yarn, it's alphabet souped in there. What they care about is, how do I make that better decision? Yeah, so we'll talk about how this compares to kind of the more traditional way of doing quote unquote analytics in the enterprise, which is kind of the data warehouse approach, which is more kind of looking back versus what do I do next? How does it compare? Is this complimentary, do you think? Or is this disruptive or some combination of both? Well, it's complimentary in the sense that BI is about descriptive analytics. Measuring what happened. Big data is about predictive analytics, what is going to happen, and it's about prescriptive analytics, what should I do? And so at the high level, they are very complimentary. The challenge we have is that most BI organizations think they're doing predictive, right? And yeah, they're building trend lines that show some lines going up and things, but they really aren't figuring out what's going to happen with any level of confidence and they have no idea how to deliver recommendations. So while at the high level, the data warehouse, the EDW is not going to go anywhere, it's still very important, you still need to have end of day, end of week, end of month, those reports come out on Monday morning when the executives show up, the dashboards are all updated, but it only tells you what happened, how many beds I filled, how many patients I have, what my sales were last week, right? The data science is doing the predictive about what is going to happen, snowstorm comes, you're going to have 27% more, and then the recommendations. You need to bring in two extra nurses, you're bringing in one extra doctor. By the way, the kinds of incidences you're going to have are going to be lacerations at head wounds, you're going to need to have following medication, get to the prescriptive, get to those recommendations, and it sort of requires, one of the things we find in the process is that IT organizations are really good at thinking better, cheaper, faster, better, they're not good at thinking different. And for most organizations, it's not the technology gets in the way, it's the mindset of saying, okay, I got to enterprise data warehouse, I'm thinking retrospectively what's happening, I need to change my mentality. And some people can't make that switch. And do you find the business is more likely to take new approaches, they're easier to talk to, and think about doing things a different way rather than IT, which might be stuck in a certain way of doing things? I think that's a good point, Jeff. I think it is exactly right. I think the business folks are seen for the first time an opportunity for something going forward. Well, they're the ones feeling the pain. Oh my gosh, yeah, they're the ones who are, today they're getting these charts and dashboards and reports, big, huge reports, and what the heck am I gonna do with this? We had an engagement with a grocery chain and we were developing a, we created this mock-up and the store manager of the store we were dealing with had this BI report, you know, green arrow, yellow arrow, red arrow, and she goes, what the heck am I supposed to do with this? I didn't tell them anything I can act on, right? And so the users are frustrated by the BI tools telling you that you're suck without giving any guidance about how to unsuck. So Bill, I got to, I want to chill down on that because one of the things we hear all the time first of all, you know, we do hear that Hadoop is a boardroom conversation that's do not Hadoop or at least what Hadoop will do and there's a solution, that's one thing. But also there's pressure to move fast and change, obviously, you know, it's happening all around us in the industry, open data platforms, one example, much others. But I want to talk about the customer. We've heard from practitioners on theCUBE and also privately, it's like, hey, if someone walks in my door and says, I got a platform for you, I'm going to shoot myself. I mean, literally, you know. Meaning they're like so oversold on platforms, just, hey, buy my platform and your problems are solved. When they just want some tooling and then maybe back into a platform, are you seeing fatigue on the sales side, customers getting somebody with platform this, platform that, a lot of rip and replace, a lot of promises, not a lot of delivery. So do you see that? And if you do or don't, talk about that dynamic, what's going on. So John, I think you're spot on. I mean, I think the customers out there are really sick of hearing yet another platform. They don't know what it means and they don't know how it helps them. I think last time I was on it, we jokingly said, but it's not the three V's of big data. It's the four M's, right? Make me more money. So they don't understand how the platform helps them make more money. And so they push back on the vendors. And we see a lot of product sales flatlining because they're just tired of buying on someone's promise. What's happening, especially inside of EMC is we're seeing that our go to market approach has to change. It has to be focused on our customer success. And I'm not talking about customer success in stand up distinct in technology. I'm talking about customers and success and helping to drive their top line, helping to sell more products, helping to be more effective here. Help your customers, customers. Help your customers, customers. That's what you mean. Yeah. Well, in some case, yeah. And here's a good example. We're doing this project for a casino. And how do we... Hard to write a book on a casino, isn't it? Oh, God. Yeah, it is. I'm fortunate the money's going the wrong direction. But I can tell you, I know what slot machines to play. So, but you know, they're trying to figure out how they get more interaction, right? So, and yeah, they've got ways to interact with the actual player themselves, what kind of incentives they send them and what kind of comps they give them. But how do you arm all those frontline employees, the ambassadors, the hostesses, the pit boss, the floor manager, the valet, the waitress, the bartender, right? All these people who are touching the players throughout the day, how do I make them more effective in conveying a better player experience? And so, well, yes, it is ultimately about how do I get the consumer to do more, the player to do more, the patient to be more effective, the teacher to be more effective, the student to be more effective, right? It's how do I basically figure out how to empower the human in between to make them more effective. So, I know you're coming into the casino and I know this is your favorite slot machine, and so I make sure that slot machine's available and I take you right to it, I rush it through, checkouts, you're not spending any time in the hotel, I got your drink waiting for you there, start shoving, well, you're probably a nickel player, but actually no one knew, you're probably a big dollar player, but you're shoving your dollars in there. No, I don't do slots, I'm not a slot person, not a big slot guy. But that's, you know, the idea is how do I, how does the hostess then get you to your game quicker and make sure you have a great experience? Well, it's about personalizing the experience, and that's what it's all about, whether it's in casinos or anything, anything, any industry where you're a consumer facing, it's about, you know, people become, come to expect that kind of personalization you get in the, in, you know, when you open your Facebook account and you expect that kind of personalization, this is all about me, and they expect that now in the real world as well. Yes, they do, it's about personalization, not only to your customers, but also personalization to your employees. That's a good point too. You got to think about the, you know, the single most important person in the grocery business is a store manager, not the senior vice president of marketing, depending on what he or she may say, it's a store manager who every day is making price and decision and placement decisions and markdown decisions, how do you make them more effective? Right, how do I make sure that I'm building the right kinds of insights, recommendations that helps them to know that, hey, on this particular day, you know, you're, hey, you're a block, you're a mile and a half from Stanford football stadium and they're playing Cal this weekend. You're going to need to have more beer, and by the way, Stanford fans don't drink Budweiser. And they can't bring beer in Stanford Stadium, I know that because we try to sneak it in, but that's a whole different story. As you know, so let's bring up this interesting concept, immersive experiences in the moment. You see things like Twitter, certainly on the crowd, conversations with crowd chat, what we do, you're talking about crossing the street, mobile device, you know, games, you're talking about in the moment, real time, right? So what are you seeing with real time? Obviously that's becoming more and more of the value proposition because people are in real time with mobile devices, whether they're things or people. And let me kind of dissect the term real time for a second. So what we find is when we look at a business process, a business problem we're going after, we try to decompose that into the data events that occur. And there's almost always in there some data event where timeliness is important. Let me give you an example. Hospital acquired infections. When you check into the hospital, we figure you got between, that the person that admissions has between four to five minutes to figure out the likelihood of you capturing staff infection, right? We've already pre-built a score that looks at your likelihood, but the nurse also looks for, do you have any open wounds, right? And we're also going to take into account what kind of treatment are you in for? You're going to need a catheter, right? Those three variables tell us about the probability you're going to probably capture staff infection. And if your score is too high, we actually can push you different part of the hospital, right, to receive more care. You might spend an extra day there, but you're not going to get a staff infection, right? So in that case, I don't, I'm not making a sub-segment, a sub-second based decision, right? You're walking in front of a Starbucks. There you are, boom, here's your favorite drink. Come in and get it, right? We got, sometimes the timeless might be measured in minutes, maybe even hours, but it's not batch, right? We're trying to be more time. It's near real time. It's near real time. And so the only way that you can really think through how real-time it needs to be is to take that business process and decompose it. So we've got one minute left. I want to ask you a question. I want you to respond to this. Wait, I only got one minute left? No, we'll go over it. We'll always go over with you. I got to ask you guys a question. Technically, it's technically, you know. I have to ask you guys a question yet. I got a tough question for you. Well, let's ask me a tough question after I get this one out to you. I want you to react to the following two concepts. Systems of record and systems of engagement. What's the difference and where's the action? Oh, that's a great question. It's the debate I have a lot of organizations is creating a 360 degree view of the customer, which I think is BS, right? Why, right? Versus creating a profile on a customer where you've quantified all their behaviors. To me, that's a difference, right? I have a system of record, which is my historical 360 degree view of them, but I'm actually more interested in building this system of engagement. What's there? We're working with a financial services firm and we're creating a retirement readiness score for all their customers, right? It's like a FICA score where you take all the other aspects and you create these scores. The hospital chain wants to create a health score and a stress score. The stress score, by the way, is going to feed in lumosity data so we can help measure what's going on. So to me, it's a difference between 360 degree view of the customer, which is nice but not actionable, but it's important data to have versus this system of engagement where I'm actually got a bunch of metrics, a bunch of measures, a bunch of scores in place that actually allow me to help me make better decisions. So more real-time, more active data? It's definitely, yes, it's active. I'm not sure. Are they decoupled or they integrated? They're decoupled. They're decoupled. To me, you're going to feed that system of record data and you're going to go through it. Because you need to have historical perspective. Which is checking, just checking where you're at here. But you still have to understand who that individual is, right? So you don't want to have, if you've got three profiles for the same person, that's not good. You do need to have that 360 degree view, but then it's about adding in the components around what should I do to engage the people. And you may not capture everything about that customer. There's maybe things that you don't need to capture. Maybe you don't need the 360 degree view. Right, maybe 180 is good enough. And it depends on your business and what the use case is. In some cases, it's like data quality. You don't need 100% data quality for every use case. Good point. And if you don't have the history, I mean, what's real-time data without historical perspective? That's noise, right? So you got to have that historical perspective because then you build with profiles. And I know what kind of products John buys and what kind of food you're interested in and what kind of games you like to play, right? If I know that historically, and I notice a change in that, that may be something I want to act on. Right, you're getting ready to leave. You're going through a life of change, a vendor, something like that. I need to have that history so that as a real data comes in, I have something to compare it to. So we know you've got your finger on the pulse. As always, Dean, a big data. We think, well, we think, we talked on this a lot last night heavily and all day yesterday, we think that we'll keep on and this look at an angle, Dean, that systems of engagement is a holy grail. And because undefined, no one has that, it's not mature, but it's real-time, it's in the moment, it's immersive, it's active. And we call those analytic profiles. We're building analytic profiles on the strategic nouns of their business. And it might be customers, it might be stores, it might be wind turbines, it might be jet engines. Okay, so now ask us the question. What do you want to do? Yeah, this is a reversal role here. The Dean is going to ask the students. Big data is all about openness, right? Open technology, in many cases, open business models and open data. Here you are at the Strata event, run by Riley who is an advocate at Open and yet you guys are over here at the Fairmont. It seems to be something contrary here that maybe they say open out of one side of their mouth but don't mean it out of the other side of their mouth. Is there a question? Yes. Is there a specific question? Yes, so what do you think about their definition of open? I think O'Reilly is a closed organization. I mean, it's a closed mindset, right? A closed is about a walled garden. AOL was a closed system. It did great and then died. Walled gardens don't work in the media business. That's proven. O'Reilly is a walled garden and they pretend to be open, but they're really not. And so theCUBE has been part of the open community. We help people. You wrote a book. It's helped people. We were last night talking to people. There's some big data stuff. We heard people's lives are being saved. Someone was doing this big data project based on theCUBE. Was doing some suicide prevention technology that changed some lives there. So people changed their business. So we're open. So we're open source media. So O'Reilly is closed. We're open and that's just the way it is. And hey, that's their business model. Ours is different. So we don't necessarily have any feelings that way. It's just a bad decision on their part as far as we're concerned. But we support that. They're putting out some content. We'll point to it. In fact, we're linking a lot of their stuff. The reason why I ask the question is that one of the things that I'm observing almost bubbling up right now is organizations moving to more open business models. Models that allow third parties to make money on their platform. And one of the ways that becomes important is because you think about these, they become marketplaces and you're capturing data about players, the products and you're trying to help match and merge things together. The whole idea of how do I build a model that no company can innovate as fast as a market place? You're like a management consultant as well as a technology leader, which is awesome to see customers get that value from your expertise at EMC. And I agree. So our philosophy, and talk about the O'Reilly question, it's just the difference of philosophy. It's like they have their business model, they're closed, they'll ultimately go the path of what walled gardens become highly profitable and then they milk it and have to reinvent themselves again. But we're seeing this inside out organization where value in the community models are all about sharing and collaboration, some called the sharing economy, what not. That's what we're part of. We're part of this new revolution. But Dan Hutchins at CSC was on the cube at Oracle Open World, where it's now open because the queue is there. And they're open, Open World. He pointed out what he's doing at CSC is building an inside out organization, meaning they believe that they have to take their inside of their company and make it outside. So for instance, they're doing crowd chats internally, having all hands meetings on crowd chat publicly before the open. So doing things in the open really changes the dynamic. So you're seeing corporations going inside out, you're seeing paradigms like perma-less security. You know, a good friend of the cube, Steve Herrod over there at the General Land. He's got investments where it's not about the mode anymore in the walled gardens about, you know, notification economy, sharing economy, inside out seems to be the model. So I would conclude that this part of the conversation by saying that there was a quote made recently that cold Trump's cash. I would tell you that business models trump code, that an open business model is the key to success. And what's happening is organizations, not only in vendors in the big data space, but now corporations are starting to realize that in order to be successful long term, they need to become more open. And big data plays a very important role because the way to become open is to create this marketplace and provide insights to all the participants that allows them all to be successful. Bill, great to have you. And thanks for opening up the Pandora's box on our openness. And you know, we love, Tim O'Reilly's been on theCUBE many times, we love O'Reilly, we love you open. For us, content is everything, right? Content is what creates opportunity in the social web. It's about social, thanks for coming on. As always, being a great friend. This is theCUBE, we're live in Silicon Valley for big data SV in conjunction with Hadoop World and Strata Conference. We are here live in Silicon Valley at the Fairmont. We'll be right back after this short break.