 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. Welcome back to IBM Vision, everybody. I'm Dave Vellante with Jeff Frick. This is theCUBE. Check out ibmvisiongo.com. It's our digital interactive experience. We've been here day and a half, wall-to-wall coverage. Marcus Hearn is here. He's the worldwide marketing leader for Watson Analytics. Marcus, thanks for coming on theCUBE. Thanks for having me. Yes, so Watson Analytics, all the buzz of this show, it seems to have supercharged this business. You know, the sort of stodgy governance, compliance, you know, reporting. All of a sudden, it becomes predictive. You must be excited. Very much so, absolutely. This is, I've been in the analytics business in software for about 22 years now. And this is that point at which you see that inflection, much like computers had, I think, in the 90s with the graphical user interface that Microsoft developed for Windows and then Macintosh, of course, by Apple, really kicked it into the sort of general populations use, right, beyond just computer kids and stuff like that to everyone. And that's, you know, I think what Watson Analytics is in terms of analytics for people, really takes it from analysts to everyone. Well, and you know, the promise of this business for decades has been to give us this 360 degree view of our business, to be predictive. And it ran into technical hurdles in the sort of late 80s, early 90s, even early 2000s. And all of a sudden, you know, the Enron disaster was a real boon to this business, got a little lift. And now we've kind of come full circle. Why is it now that we are able to fulfill that original vision? Well, you know, I'd love to give you a sound bite. There's never one answer, right? But I think you can get a few things down as being the contributors to this, the key drivers. And one of them's like the ability to collect all this data. All right, you remember back in the Y2K problem, that existed because people were trying to save memory because memory cost money. And so instead of putting in, you know, the full date, they would truncate it. And hence that issue. So back in the last century, there was a real drive to sort of compact, to save on space, to just get the minimum of data you need to sort of run your business operationally. And that has since gone away. So the amount of data we can collect and hold on to and store and archive has just exploded, right? The advent of big data is not so much that, you know, suddenly more data is being generated. It's that we are capturing it. And then when you've got all of this data, the true value in that data on its own is not so good except to fulfill your orders and run your business. The true value is when you apply analytics to it. So as people see that happening, as competitors do it, as innovations built from the information that's derived out of that data, it just, it bends snowballs in effect, really. So you have this sort of confluence of the ability to collect it, some people who do it and are successful and then just drive that innovation and change. And as we all know, anyone who doesn't keep up tends to go away. Well, you've been following this business for a while in more than following, you've been, you know, in this business for a while. And so you remember in the mid 2000s, even around the time that IBM had acquired Cognos, there was a line of thinking, you'd pick up the Harvard Business Review and you'd see executives talking about how gut feel, trumps the data analysis. Nobody's talking about that anymore. Sampling is kind of gone. There's been a real mind shift, hasn't there? Yes. And sampling, that's a great example because when I started my career, that was what you did. You couldn't really analyze the entire population of information you had at your hands. If you had a big population, you could usually get hold of a sample and then computationally, that's all you could deal with. And from that sample, you would try to understand what the population looked like. So we still do it today with polling, right? With elections, because you can't ask everyone before an election who they're going to vote for, you take a sample. The big difference is now with, you know, the data and with the computational power and what we can do is like take the whole population, understand exactly what it looks like and be able to predict what new entrants are going to look like, right? So it's a whole paradigm shift, going from like sort of forecasting what it looks like to predicting what new things look like. And to sort of bring this full circle, this kind of ties into what executives we're doing and still do today. They, through the course of their careers, understand and learn a lot and that is a sample, their experience and which they build out an opinion that's usually a very good one. They get paid well for it and so forth and do well with it. But now there's this ability to bring in all this additional information. So the sample is still relevant. Their opinions are still relevant. It's not, you know, that it's no longer there and it's all about the data. It should be, I think it was done well today in the keynote, combined, you know, you can't just rely on math alone, you have to have human interaction and that's actually one of the points behind Watson, is this ability to learn and to be cognitive, not just straight math. Straight math is going to save you a lot of money, it's not going to innovate. Well, I think Jordan Ellenberg summed it up today. I mean, essentially you can build a model and there's so many variables in that model you can turn knobs and make it say anything and that happened a lot and probably still does in organization of lies, damn lies and statistics but people would take models and they would tune that model and calibrate it to tell their story and defend their story and say, oh, I've got a big model behind this. Now whether or not they were making the right decisions, who knows, you know, only time would tell but it seems like this notion that you guys are putting forth of the sort of citizen data scientist, this sort of self-service model, you know, the Wikinomic starts to explode the data and they put it into the hands of more people and you start to get sort of counter-intuitive, you know, thoughts and potentially actions being taken. What are you seeing as sort of the state of the citizen data scientist? So it's a great term. I love the term because it sort of encompasses the thought process we have here and the state of it is that it's you, me and everyone having the ability to use analytics to form an opinion rather than just sort of hear things, discuss things, look at aggregates of data to actually use something like Watson Analytics, put the data in there and explore it and find out the predictors of outcomes, find out true causes, really drill into it in a way that makes sense to us to have a more educated opinion to then sort of like contribute to the conversation in a business. There was, you know, often I'm asked about, you know, how do you, what do you think about the shift in decision-making? It's not so much that sort of decision-making is shifting anywhere from, you know, data scientists to line of business. It's expanding, kind of like trying to think of many concepts here at once, but it's raising the ability of the entire organization to understand things analytically, not just a small part of it to contribute that into the conversation. Everyone now has an analytic basis or the potential to something like Watson Analytics for their opinions to then drive the conversation and drive to decision-making all the way up to strategy, right? Even at the level where strategy is sort of developed and it'll always be developed at that top level and then driven back down and having analytic decision-making happen operationally. So with the thousands of decisions that are made a day by systems or by people and call centers to have those driven analytically and the strategy and everything in between, that's the shift. Yeah, and really using cloud to get it to that person, a cognitive interface so they don't have to be a data scientist and give them that ability to have just more, more data to make their decisions in a broader set of employees across the organization, we find over and over that's a huge part of the old innovation equation is enabling more people to do more within your broader population. So I think the other neat thing is this, the concept that Watson Analytics is this overlay God knows what applications actually delivering the insight that I'm looking for in the way that I phrase the question or the data set that I threw in there. I don't really care which of the many, what the wealth of apps that you guys have, I just want the answer. So let Watson go in and figure out which of the pieces of the puzzle can it leverage within your suite to come back with something that I can work with. It's interesting you say that. We found very early in, back in the day when we were building out what's analytics and how alpha and beta was one thing people would commonly ask is what's the one thing, right? I can see this predictive model I can see all these contributors to an outcome. What's, you know, if I could only change one thing what is that one thing? And it's interesting because you give that to people like okay here is the most boiling the ocean down to the one key factor that you should try and control or change to optimize your outcome. What happens then immediately is people want the detail. They wanna go in, they wanna understand it and they wanna say okay so if I can change this what else can I change? And can I do this in combination? It's interesting how, you know, people strive for simplicity because then that drives their questioning. If you give them the complexity first there's this understanding and this innate thing of like well this is not my complexity this is not the picture I wanted built out let me start at the start and build it out and that's what what's analytics allows as well. So you and I and I do it often can start with the one thing and finish there and many people like to then go further and explore it a bit more and this is usually where the stunning insights come from. This is where the aha moments happen when you start with that one thing and sometimes it's the one thing typically it's a pretty obvious thing, right? If it rains you get wet but as you sort of drill through and find out then you find out all these other contributors and all these other things you can actually start doing something innovative, right? And that's the really cool thing about it. And I think that most organizations sort of give up have historically given up because it's just it's too hard to go to that deeper level they're beat up or it's too expensive or by the time they get to the answer the market has changed so all of those sort of symptoms have created sort of a backlash in this business and I see Watson Analytics as sort of changing decision-making from what I was called agenda-driven decision-making to your term analytic-driven decision-making. So how did Watson Analytics come about? Our understanding is that essentially the classic Watson crowd deliveries an arms dealer delivers services up to anybody who wants them and so how did you compose those services? Can you take us through sort of the introduction of Watson Analytics? Yeah, so, I mean, this is a long conversation how they're stuck. There's a lot of things, you know, there's decades in the making practically. But really what it boils down to is that there were teams at IBM, three of them in very particular teams that knew that we had the ability to build into the interface, into the offering. A lot of the intelligence that a data scientist brings to bear and this is understanding what algorithms are most appropriate, what data preparation steps have to be taken given how the data look, you know, this is not magic, this is a science, this is a discipline. If you have data that looks like this, you must do this to it to prepare for analysis. If you have results like this, they mean this in the context of your data and typically that's what we hire data scientists to do and we'll still do so because it can get quite complex. But we knew we could build a lot of this in. A lot of the discipline, the understanding, the interpretation of results and these three teams were doing it from different angles, right? You had the business intelligence, the predictive analytics and the cognitive analytics were all trying to build in this layer of intelligence. It's like, we're all doing the same thing. Let's bring it all together, have one layer of intelligence across these three key factors that we then combine under the covers and that's what's an analytics. What's an analytics is really, you know, a lot of marketers try to think up taglines and probably the tagline we think up is data scientists in a box. It's not actually the tagline. That's the concept of what we're all trying to do is bring that experience, that intelligence, that understanding and discipline into the actual product. The same way that, to use an analogy, we've seen this happen with technology constantly, particularly computer technology. When computers first started, everything you wanted to do had to be programmed, everything. And then they came up with a scripting language. So you didn't have to program straight, you could use scripts and then they came up with a graphical user interface. Now you can talk to the thing and look at phones. I mean, my kids know how to use a phone. So this progression of building in the layers of intelligence to broaden the scope of usability is exactly what we were up to with Watson Analytics and we came together and delivered. So what about, to stand that theme of usability and user experience, it's key in terms of how you interact with the system because behind the system is complicated stuff or average business users. So what have you done to extend that user experience to the citizen data scientists and how far can you take that? Well that's where Cognitive comes in in a big way. So you'll see when you, I mean, Watson Analytics, anyone can go on and use it now, right? So you'll see if you upload a spreadsheet, it will immediately use Cognitive Analytics to go through your data and find things that you should be interested in. It'll start straight away and say, you don't even have to know the question you're asking. You can just put it up there and it immediately comes out and says, you should look at this, this and this and this. Or give me a question, I'll answer it. Or just dive straight in and start slicing and dicing and analyzing yourself. So that concept, that ability to look at data and to know what's interesting in there is critical. That's key because I think pretty much every other offering out there relies on you rolling up with a question or a hypothesis or something in the data you're looking for. Watson, to my knowledge, is the only thing that really has the ability to do that for you. So in terms of a user experience, that's a big step. Well I think, because it gets you over that first time, I think that's an interesting thing because you talked about a path of discovery going down a question tree, if you will, based on prior results. And as you said, the hardest part often is just getting started, the data prep, all these nasty ugly things that you have to get done before you really get into the work of applying your knowledge, your insight into that data. Exactly, and the other thing that people do unknowingly, although not always incorrectly, is they roll up with assumptions. Everyone rolls up with assumptions. I roll up to sales data, I'm probably gonna start looking at deal sizes, revenue, average lead time that led to it. That might not be the critical thing in there that I should be looking at. Maybe it's geography, maybe it's the sales team on board, maybe it's the competitors, right? So that's one of the good things about Watson Analytics is it can find you things you might not have thought of. I remember the first data scientist I think I ever interviewed was Hillary Mason at Hadoop World, and she said, the hardest part of my job is knowing what questions to ask, and that's the art of data science, is what questions should I be asking in order to achieve some kind of business outcome? If I'm hearing it correctly at this event and others, you're essentially dramatically compressing that elapsed time to the right question by assisting individuals, citizens, data scientists, and getting to that, right? Because I don't know if Hillary knows the right questions, she's a smart lady, right? But I don't know if she's asking the right questions, she doesn't know if she's asking the right questions, it takes a lot of time and energy. Is that the right way to look at it? I think so, and it's not necessarily they might not be asking the right questions, it could also be, are they asking all the questions? So Hillary could very well be rolling up with good questions, but she's only got two of the four. Right, right. And then when she sees the other two, wow. Yes. Look what I can do with that now. So is this happening in the real world? You have any examples? Oh, absolutely. So this happens all the time. Like we have the SPSS predictive portfolio. See this every single day. The difference being that you have to be a data scientist to utilize the tools and get it done. So it's absolutely happening. Now it's happening with people who are not data scientists using WhatsApp analytics. And we saw Mark from Mueller up on stage today, I thought, great example of how he just sort of flew in there and did it immediately and his CEO kind of had that aha moment this changes everything. That's happening all the time. And you know, I mean, WhatsApp analytics, there is a, you know, it's free, right? And all the analysis, all the visualizations, everything you need to do that Mark talked about is there straight away. So that's picking up steam quite quickly. I was gonna say, Maurice, it's not like people have to go through a complicated business case to do this. Not really. It's sort of infinite ROI. I mean, it's blatantly obvious. Let's try it. No, I mean, the paid for versions, we don't block any of the analysis or any of the visual, nothing like that is in the paid for. The paid for goes out and extends in different ways, like massive storage or you need a whole team sort of on there with Twitter data access or something like that. But if you're just looking to get started and do data science on your data, I mean. So functionally it's the same. You might resell the fire hose or give you. Analytically and visually in terms of functionality, it is the same. There are some functions that are in the paid version. You don't get in the free version. And yeah, but not in terms of analysis or visualization. Yeah, you're talking about access to data, Twitter data, okay, the storage, you got like, the storage is collaboration features. This is a number of others. Talk about the business decision to go to market that way. To put a wrapper on a bunch of products that you've had in the portfolio for a long time, offered as a cloud-based free service to get people to try out. There had to be some interesting conversations about that because that's, or maybe I don't know the whole story is that that doesn't sound like kind of a typical way to launch a product from IBM. Well, not from IBM, not at all. This is actually very innovative, very new. You know, the team's quite excited. We challenge this every day and IBM, you know, as Ginny says, we're going through a transformation. This is one of the key elements of that. And really it boils down to, in order to, you know, really capitalize on analytics and to move IBM forward with its customers and make us essential and get them to understand this is about creating an ecosystem. This is not really about creating revenue, although that will come in other ways. This is about creating an ecosystem in which we raise everyone's analytical ability, understanding and appreciation therein of what it can do. And that will just ignite a fire in organizations. Mule is not going to stop where they are. There's no question they're going to go on and they're going to start doing things like operationalizing predictive inside. They're going to look at cognitive, rolling that in. And that's what this is intended to do and that's what it's doing. We have a huge use of population going. It's growing every day by leaps and bounds. And that was the point. The point was to build out a new sense of what analytics is to give everyone interface in there and then to show them the paths to how to really, really, you know, fire up their organizations and make them effective and optimized. We just have about a minute left, Marcus. I wonder if he could just give us a glimpse of the roadmap, maybe show a little leg. I think I'll be there. So we announced the professional version, of course. We had a personal version already and we have the free version. What's coming is an enterprise version. I can't give too much detail on what will be in that, but that will absolutely be aimed at, you know, the top level of deployment across an organization. We've got Forecast incoming, which is a big in demand by a lot of our users. We have integration with a lot of other IBM products coming as well. So for existing IBM customers, they'll be able to see direct connections into their offerings to be able to pull data and do analysis straight away. You know, we really intend this to be, in the end, the analytics interface for practically everything IBM provides in addition to probably things other people provide. And that's in a nutshell. There's going to be more than that. We have continuous delivery. They give me something every week. I can't market it fast. It's not only a shiny new toy. It's a secret weapon, secret ingredients. The analytics interface for all. I like it. And the menu that IBM is delivering to its customers. Marcus, thanks very much for coming on theCUBE. It's great to have you. All right, keep it right there, everybody. Jeff Frick and I will be back with our next guest. This is theCUBE. We're live at IBM Vision in Orlando. We'll be right back.