 Live from the Hilton at Bonnet Creek, Orlando, Florida, extracting the signal from the noise, it's theCUBE, covering Vision 2015. Brought to you by IBM. And now your hosts, Dave Vellante and Jeff Frick. Hi everybody, welcome back to IBM Vision. This is Dave Vellante with Jeff Frick and this is theCUBE. Check out ibmvisiongo.com. It's our digital interactive experience here. Jane Hendricks is with us. She's marketing manager for Predictive Analytics at IBM. Jane, welcome to theCUBE. Good to see you. Great, thank you so much. It's good to be at theCUBE. It's good to be at Vision. So we should have a crystal ball out here. Predictive Analytics, that's your world. What is Predictive Analytics all about? Oh gosh, Predictive Analytics is a little bit of something for everybody. So at its heart, it's technology. It's also a business practice. It's also a user experience, if you will. So Predictive Analytics allows you to tell what the future may hold and to do it with confidence, to do it in a way that you can take this to a boardroom and say, I know what the future will be and this is how I know it and you better listen to me. So that's what Predictive Analytics is under the covers and the rest is kind of how you implement it. So the technology of predicting used to be sort of let's look at what happened and let's assume past this prolog and then sort of put some probabilities around it and it's like Jordan Ellenberg was saying this morning, you can build a model. We can turn some knobs and we can make it say whatever we want. If we don't want to spend that much on mitigating a disaster, we can sort of lower the probability or the impact of a disaster. And so what's different with Predictive Analytics today? I think what's different with Predictive Analytics today is one, people understand that numbers can lie I think people understand numbers in a way that they haven't in the past potentially, right? People are embracing data. And I think Marcus, my previous, he talked about data-driven decisions, how organizations have come to this realization that they need to use the data they have to understand what's going to happen next. So what's really changed is this notion of it being in the hands of more users. It's more of that team sport approach, right? So you have people who are corralling data. You have people who are applying the math to it. You have business users who are taking their business intuition and looking at the math that's being applied. In some cases are applying the math and suggesting ways to make it better. And all of these folks working together are applying this technology that used to be in the back corner, I think, right? It used to be the, how do we say it? It used to be only the very few who did this. Now it's everybody. Now everybody's kind of taking a turn with predictive analytics. And that makes it more pervasive. That makes it more acceptable. And I think it makes the situations of these, let's just turn the dial a little bit. It makes it a lot harder. It's not in the back office. It's so visible. So is this Benjamin Disraeli Nirvana? You know, the highs, dam lies in statistics. What would he say if he were here today? Finally, we have a means of addressing this problem. Absolutely. I think it's the time has come, right? So we have the technology. We have the tools. We have ways of deploying it, right? So it's not just coming up with the answer. It's the means of taking the answer and doing something about it at a deep level. So you have all of these steps, right? All of these pieces, all of these jigsaw puzzles that have taken years and years to build. Now they're here. And now you can actually do something with it. And I think that's what's really exciting. And that's why you see all the people here. So go ahead. That's just to say, and how do you see the propagation move from the back room? Are people asking for it? Are the people coming out and say, hey, you can do this now. You've never done it before. Why don't you try? Is it a top-down directive that says you must have more data behind your suggested decisions? I mean, how is it actually kind of moving through the organization? What do you see in the field? You know, we see flavors of everything you've just described, right? People, it's what I started saying. People come to this. It's kind of, they come to it from a lot of different angles. So I think what we're seeing now is it's kind of this bottom-up approach. It used to be kind of a top-down mandate. You will bring data. You will defend your decisions with data and people would say, okay. And they defend their decisions with data and they try to adjust the dials. They try to defend a decision that maybe they already had with the data at hand, right? But now it's more of this bottom-up. It's this groundswell. The tools have gotten easier to use for folks like us. For folks like me, I'm in marketing, right? What do I have to do with predictive analytics? I shouldn't even touch it. And yet I can't because the tools allow me to play with the data in ways that you never could before, beyond standard dashboarding, beyond reports. To actually find meaning and build models that I can then take to an expert and it's not throwaway work. The expert looks at this and says, you know, you're onto something. Let me refine it. Let me add in some advanced capabilities. Let me really test it out. And then you can turn this over up the chain and say, you know what? Now we've really onto something. Let's deploy it. So I think it starts off at the bottom when people are using these tools. They may use many different tools. They're learning this at university. They may use R, right? Everybody sort of R. Use some open source. They use tools like Watson Analytics. Incredibly easy and yet there it is, it's doing predictive analytics. And then they kind of move up the chain to more sophisticated tools. Once you start, it's kind of hard to stop, right? And as these users are finding meaning, they're finding things that are informing the business, you get better and better and better until you're actually able to change how you do things with data. Does that make sense? Yeah, it's like buying a boat. You thought I was small, keep getting a bigger one. If you're, yeah, exactly, right. People process in technologies. We talk about it all the time on theCUBE. And they always say technology is the easy part. Of course, talk to the guys in the lab and ask them if what they're doing is easy. But nonetheless, you're bringing the technology to the table that we haven't seen before and it's game-changing. But I'd like you to discuss the people part of it and the process part. Let's start with the people. What kind of skill sets do I need to create a predictive organization? Because I've got this entrenched disinertia of, well, let's turn the dials and my agenda-driven decision-making versus the citizen data scientists that you guys are promoting. So let's start with the people. What kind of people skills or mindset changes do you prescribe to organizations? I think at the heart, it's not, a lot of people will start off and they'll start talking about the skills, right? You need to understand statistics. You need to be a data, and I don't think that's true. What you need are people who are passionate about data, who will let the data tell its story, right? There's people who are afraid of that and there's people who aren't. And the mindset that you want to have at your organization, if you really want to be data-driven are people who are not afraid of what the data will tell you. Because sometimes data will tell you things you don't want to know and you don't want to hear, right? So you have to, once you take that off, I think the possibilities are endless. Does it help to have statisticians? Of course it does. Does it help to have data scientists? Of course it does. Does it help to have DBA administrators who really know data? How do you structure it? How does it live? What does it live with? Text analytics, all of these different kind of technical analytical capabilities are important, but without this mindset of I'm gonna let the data tell me its story and I'm going to feel good about it, no matter what that story is. All of these other skills are not as impactful, right? And you can learn bits and pieces. I said it's a team sport. You don't need that one smart guy who knows everything, right? People come to the table, they come from their backgrounds with lots of different analytical puzzles and it goes back to technology. I'm a marketing manager. I love predictive analytics. I've been in technology. I think all my productive life, right? All of my career has been spent in technology. So technology can help all of these different folks work together, some technology, not all. And the technology that enables everybody to work together, right? If it's backed by that mindset of I'm gonna let the data talk to me, it's incredibly powerful. That's what we see transforming businesses and professions and industries. So I want to explore that a little further if I can. Sure. You've got a lot of experience in this space. So you meet a CEO at a cocktail party and he or she says, hey, Jane, help me. I'm frustrated with my organization. They've made decisions based on their own agenda and I want more data. What should I do? Okay, you got this Watson analytic stuff. Let's assume I bring it in, but how should I organize my teams? Should I get a bunch of DBAs? Should I get the data scientists? You were saying that's nice to have. Is that where I should start? Where should I start? I would tell them to start with folks who understand that data's not scary, right? I'm going back to that. Folks who are comfortable working with data. Folks who are used, I wouldn't say used to because nobody's used to it. Folks who are comfortable with this notion of, okay, I have this data lake, right? We talked about data lakes. What's in it? We're curious. We talked about data oceans, by the way. We like data ocean. I think eventually it's going to be, I don't even know if it's eventually, I think now it's really more of a data deluge. I can't say it, right? It's not a swamp. Some places it's a swamp. People will start off like, oh my God, I don't have enough data. I guarantee organizations, they have plenty of data. It's not a matter of finding the data. It's a matter of being comfortable enough to say, okay, here's my big data store. Let's see what's in it. I'm going to see it. I'm just going to point something at it and see what's in it, and I'm not scared. Those are the kind of people you need to bring to the table, I think. It's that mindset, and I'm sure you've seen it. Yeah, absolutely. So I'm going to communicate, no question. So I'm going to now communicate to my organization, hey guys, we're going to be data-driven. You hear about this data-driven, so we're going to make decisions as a team. We're going to put the data in the hands of all of you folks, and let's talk about it. Let's analyze it. Let's not have somebody hiding behind a model with a set of assumptions that we don't really understand, that's going to take us hours and hours and hours to unpack because it's so complicated. Here's the data. This is how we're going to go forward. So that's a starting point. I think it is a starting point, and it's being for the CEO and for IT and for the technical leadership, I think it's important to allow them to let users access data, right? Because some of it is, in the past, it's been a matter of we need to control it, we're scared what if it gets out, single version of truth, whatever. Let folks play with it. If you're really going to be a data-driven organization, you have to sometimes take some of those breaks off, let people lose, let them play with the data, let them kind of understand it at their own pace in their own way, and then you can all sit down and gather those perspectives, and that makes for a very, very strong process. I was going to say, but there's still this cognitive leap that they have to make. Even if I have all the data and I do reports, traditionally it's been backward-looking. I want a data-driven decision. I'm looking at data as to what's happened, modeling it, doing it ever. It seems like there's another leap then to then let the machine come back with the predictors with some confidence level that I feel confident enough, ah, maybe I screwed up the moment. Maybe I put the wrong data in. How do people kind of get over that hump between looking back and being kind of data-driven in terms of what happened versus really letting the data set the course for them? It seems like a bit of a leap. It is a leap. It's hard. It's not a trivial, we talk about that, it's like a switch. It's like self-driving cars, right? Which is just- It is. You know, when do I take my hands off the wheel and have the confidence that we've set it up the right way and this is good stuff that I should get behind? Yes, I agree. So you do need, as an analyst, right, somebody who may be used to, I hate saying the old way, because it's not the old way, it's just a way. You're used to a certain way of looking at the world. You're used to, okay, here's the stuff in front of me, this is what's happening today. Now I'm going to trust this algorithm, this funky thing, to tell me what's happening and not only am I gonna trust that it's doing it right, quote unquote, whatever right means, but I'm going to actually change my behavior based upon whatever this thing is telling me. Is that something that is very easy for a mind to do? When we talk about it, of course not, but when you're actually sitting there and you're free to do this and you're able to do this yourself and you can see what's happening, right? You see it on the screen. So if a tool, again, technology comes in here, right? If the tool is accessible enough so that you're comfortable using it, I think it can help you overcome some of those objections, right? If it's something very difficult, then you're going to be more hesitant. You're going to think it's more of that black magic, right? It's a little bit black box, but if it is accessible, if it's something where you can kind of control the environment, you can look under the covers, you can validate it, there's guidance, right? There's communities, there's other folks talking about it. I think it can help you bridge those gaps as a user. Well, I would imagine most importantly is that you can have little victories, you can start to get a positive feedback loop that, okay, I did it, it worked. Okay, now maybe I'll trust it a little bit more, I'll go out a little bit further, I'll incorporate a bigger decisions based on the data that's coming back. That's one way of doing it, or you start small, you do something little, you gain that victory and you say, okay, I started this on a little Excel spreadsheet, right? I'd make Excel data, look at this, I build a model, I'm so proud of myself. Let's add in more data. Let's try a different technique. Let's see what my data science buddies are doing over there and let's incorporate their smarts. You see what I mean? So you can grow in so many different ways, but it always starts, I think, with feeling that you can actually do it, period. Do the analysis. Yeah, so I'm excited, I want to jump right in. So I got this data lake or data ocean, I've got some financial data and I got sales data in there, I got my social data in there and I want answers, I want answers to my questions, I want to know what questions I should be asking. Okay, so I have all this data. How can you help me? We can help you do pretty much anything that you want, of course, but the first thing that I would start with is okay, you have all this data, that's great. What we would start with is what are you trying to achieve? Right, you have to start backwards before you can go forward. So I want to understand my business, I want to understand where my opportunities are, how I should be optimizing, let's say my sales teams. Good. What products should I be investing in? Those are the kinds of questions that I have and then what other things, what are my blind spots that I should be asking? Yep, see, and right there, what you just said, that's what will guide what you do next, right? Because you're clearly starting off from that sales perspective. Other folks can start off from, I am really worried about fraud. I really need to understand the risks to my business as opposed to what are the opportunities. Somebody from like police departments, we work with police departments quite a bit, they come to us and say, you know what, I need to understand how to deploy my troops. And all of these questions will assume certain, how do we say it, the application of a certain technique potentially, right? So one may be an associative, what are the things that people do together? That's what I need to understand. What are the drivers for X, Y, Z? That's what I need to understand. And once you set up the business problem, you understand the kind of data you have, right? Where it's coming from. That's where the technology A will help you put the data together in a meaningful way to achieve that business problem, right? Help guide you through what are the various, there's so many different analysis techniques that you can apply. I mean, it's an endless thing. But what you're trying to achieve and the data you have will figure out the stuff in the middle, right? So then you do the stuff in the middle and that's the fun part. That's the part I really love. And you try a whole bunch of different things. You figure out which one gives you the best answer. And the best answer could be the most accurate, given the technique, right? The kind of mathematical accuracy metric, or it could be what makes the most sense to your business. Okay. And then you take the next step. In practically speaking, I put this data into your predictive analytics cloud. I structure it in some way, or how do I get started? So one of the advantages is that you actually don't have to have a really predefined structure in the first place. If you want to put it in the cloud, that's great. Because I don't really know. And that's fine. And that's most people are like that and we wouldn't want to say, okay, your data must be perfect. If your data is not perfect. Can't help. Yeah, yeah. I'm sorry. That's not a good message, right? Right, of course not. Data's never perfect. Nobody has perfect data. So we have the ability to work with your data wherever it's sitting. It's in the cloud. Fine, it can stay in the cloud. It's on-prem. Great, let's keep it on-prem. You have it mix and match. That's fine. Let's mix and match it. You have that kind of, let's say, freedom to do whatever it is that makes the most sense, right? Sometimes you may say, you know what? I have this data sitting here and I really think that I want to move it to the cloud. Awesome. Or you say, I don't want to move anything anywhere. That's awesome too. But wherever that data is at, so to speak, and that's my Chicago coming out, wherever that data's at, you can work with it. And one of the things that we do from kind of this performance perspective because of this data deluge is we push the analytics closer to wherever that data is sitting in the first place. So instead of doing all this moving things around, let's take the analytics to the data. Let's make it more efficient. Let's make it sing and dance and come back to your business problem. Okay. And so to engage with you, I need some software, right? So I acquire some software licenses from you. I got to run it on some hardware somewhere whether it's in the cloud or on-prem. Or both. Or both, absolutely. And then that software that I acquire from you is what? It's Watson Analytics. It's some other predictive software. It's a combination of things. What is it? It is whatever makes the most sense for your users what you're trying to achieve. So Watson Analytics? Absolutely. We also have a full predictive analytics portfolio from the SPSS acquisition and that has Modler which is a awesome workbench and it can be deployed in many different ways. There's statistics which is used by millions of people worldwide. So it's really the tools that make the most sense. Okay, like you said, if I have R, I can use R. If you have R, you can use R right alongside of the capabilities we have natively and that's an advantage, right? Not everybody wants to program. Some people wanna program everything. Some people wanna program something. Some people wanna have somebody else program stuff for them. That would be me. And the advantage is we allow you to do any of that. If you wanna program it 100%, awesome. If you hate this notion of code that you just wanna do GUI, great. Or the people in between where you have the code behind a GUI that somebody like me can use and I don't have to touch the code but I can do all the fun and cool stuff that the code allows me to do. Does that make sense? Yeah, it does. So okay, great. So we're at the cusp of this sort of new, I feel like we're at the cusp anyway of this new dawn of this new era of analytics. Where do you see it all going? Let's be predictive. Where do you see this in five to 10 years, Jane? Let's be predictive. Where's my crystal ball? I don't see it. I think in five to 10 years you're only going to see things. More people are going to be predictive analytics users. I think you're going to see this. We won't have this kind of fear of data, right? You're gonna see this across the board. Everybody will be a predictive analyst, I think in some way. I also think what's going to happen is you're going to see predictive analytics getting infused and embedded into applications. You see this happening now but I think you'll see this more and more. So pretty much anywhere you have data it can benefit from that forward-looking view, right? Right now we're kind of scratching the surface but I think in five, 10 years as the tools get smarter, as we have access to more types of data, things like geospatial, right? We only started playing with that. So the analysis gets more sophisticated. The user base, the people who are doing it, I think will explode. I think everybody wants to do this now. Think of Watson analytics and everybody can do this, right? And it will become a lot more pervasive. You'll see this in more and more and more applications. Imagine a world where analytics is as pervasive as search in applications and everybody can use it. Jane Hendricks, thanks very much for coming to theCUBE. It was great to have you. Thank you so much, thank you. All right, keep it right there, buddy. We'll be back with our next guest. This is Dave Vellante with Jeff Frick. We're live from IBM Vision. This is theCUBE. We'll be right back.