 Okay, we're back here live in New York City, this is ground zero for Hadoop World Stratoconference and of course the big data NYC event to put on by SiliconANGLE, it's our first CUBE event, we actually had our own event. I'm John Furrier, the founder of SiliconANGLE, my co-host Dave Vellante, our next guest, David Smith, VP of Marketing and Revolution Analytics. Welcome to theCUBE. Thanks for having me, John. You guys have some announcements today, but obviously the big discussion is analytics. It's two Hadoop worlds ago, we've been here now, our fourth Hadoop world, Mike Goulson said, the big tsunami of apps are coming, they kind of never came, except for analytics. Everyone wants analytics. Tell us a little bit about what the announcement is and how that ties into the froth and the demand for just multifunctional analytics and visualization, et cetera. Yeah, that's right. Today we've announced the availability of Revolution R Enterprise 7, which is a big data, big analytics platform based on the open source R language. And kind of over the last few years, there's been all these stories about companies collecting these massive volumes of data, but now the question has really come up, what are we actually gonna do with that data? I mean, it's one thing to be able to query it, but it's another thing to be able to actually ask that data meaningful questions about, which are my most important customers or where should I build my next store location or what should I be doing with our marketing programs? These are questions that data science can answer and data scientists using the R language can do that with Revolution R Enterprise. So I'm reading your press release and it says moves the computation to the data for high performance analytics. I mean, that seems to me the way it should work in Hadoop. And I remember when I first started to learn about the whole notion of Hadoop, the idea of shipping function, five megabytes of code, petabytes of data. So this is very much dovetails into that. I wonder if you could talk about that from an architectural standpoint and is that different, how different is that from other approaches? Yeah, I mean, for one thing, these algorithms are very mathematically complex. You know, it's one thing we've been talking about doing just simple queries of data on Hadoop. But when you then talk about doing things like regressions or predictive models or tree models, computationally, these are just orders of magnitude more complicated. So it's really important, especially when we're talking about big data. First of all, not to have to move the data anywhere else. And secondly, to be able to use the computational capacity of the data platform itself to build these advanced predictive models. Now we think about Hadoop as being a data storage mechanism, but also recognize that these Hadoop clusters have dozens, sometimes hundreds of CPUs, computational processes in them. And if you can apply them to these predictive models, you've got this computational powerhouse, a massively parallel scheme that you can use to actually build these models. Dave, I've got to ask you because this brings a question out. This analytics conversation is really part of a bigger global conversation around enterprise readiness. And that's been a big theme around the big data world. The transformation is shifting back to the clouds. You've seen a lot of activity on pure build-out. Data centers, modular data centers, Google's going to have a data center in the San Francisco Bay area, maybe they talk about that. So you're seeing really cloud and the hybrid cloud really emerge into a viable construction. That is only going to fuel more enterprise platforms. So I want to get Dave's perspective, then I want to ask you, David, Dave, Smith to comment on this, because there's just two things. There's an analyst perspective, then there's your perspective of, okay, customers are confused about what platforms to choose from because with the transformation of the underlying infrastructure under the hood, whatever, how do they maintain that investment protection? Dave, I want to get Dave Vellante first to your perspective of the choices in the mind's eye of the CIO. Okay, because analytics is the killer app. It's like email was once the killer app, analytics is now the killer app. So how does the enterprises pave the road or are they putting gravel down? Or what are they doing? I think the smart CIOs, or maybe it's not even CIOs, the smart practitioners know that there's always going to be some new technology coming on. So they have to work into their processes, the ability to absorb these new technologies in as seamless a way possible. So that's why you always hear so much focus on people and process. And the smart organizations that we work with are constantly looking at the skill sets that they have. It's almost like the Jack Welch approach. Remember Jack Welch would say, okay, 10% of the staff gets fired every year. Always wanted to, well the smart CIOs are doing the same thing. Look, we have to update our skill sets constantly because it's people process and technology and technology is frankly the easiest part of that. So I think that's the framework that the most successful companies use in order to be able to adopt new techniques and technologies. Yeah, I'm certainly hearing the same thing from CIOs. There's so much disruption going on just in the data platform arena. When we talk about the distinctions between Hadoop and data warehouses and data appliances and cloud and cloud appliances, CIOs just aren't sure what's gonna be the long-term contender in terms of the data platform itself. And it may well be a mix of those because each of those have their own characteristics. But when CIOs think about the analytics that they're gonna build on top of these data platforms, it's risky to think of committing to analytics that only works in any one of those given platforms because you don't know what's gonna be around in a few years' time. That's why with Revolution or Enterprise 7, we developed this write once deploy anywhere feature. Data scientists can write analytical methods in the R language and have those same methods run in Hadoop or in an appliance or on a server regardless of what's going on. Yeah, that's where I was getting at. So that comes back down to they wanna move now but they wanna have investment protection. We used to talk about it in the market headroom, right? I wanna have some headroom to grow into. Can you talk a little bit more and drill down on that because that's interesting. They wanna make some informed decisions now. What do you say to the enterprise customers that's saying, hey, we can get you up and running today no matter what your choices are strategically on platforms or tech, we're gonna have some headroom for you and some extensibility. Yeah, I mean organizations for one thing should be thinking about investing in data science because it's data sciences that actually gets the information out of the data. We hear a lot about the skills gap for data scientists. Some companies are saying it's hard to hire data scientists but I think that those companies that are built on a legacy platform for analytics it's hard to actually recruit this new generation of data scientists that do know Hadoop, that do know open source but the companies that I've seen be successful have set up new data scientist groups based around open source platforms around the R language. Data scientists are graduating from universities every year trained up in the R skills and are ready to actually dive into that data and find out what new insights are available there. So I wanna ask you about the right once deploy anywhere approach because from an economic standpoint it's very alluring cause it's a one to many model. What's the enabler for right once deploy anywhere? Yeah, so fundamentally what we've done at Revolution Analytics is to rewrite some of the existing algorithms that come with open source R but as what we call parallel external memory algorithms. So unlike open source R which runs single threaded we design these algorithms to run with multiple streams of data at once and to be distributed across parallel data architectures and what that means for us as developers is we can take those same algorithms and port them into a Hadoop platform or into a database platform taking into account the unique performance capabilities of that platform but hiding those details from the data scientist. From that point of view they're just programming in the R language and those computations happen to take place into whatever data platform they're working with. So where's, talk about adoption a little bit cause you got the college crowd, they're very familiar, much more familiar with R and you got the SPSS crowd that sort of wants to learn more about it. So I wonder if you could talk about that dynamic a little bit and talk about what's driving the adoption. Yeah, so I think the distinction between R and SPSS they're a little bit sort of chalk and cheese to use an English expression. Bless your cotton socks. Yeah, R is more of the data scientist programmer type of environment whereas SPSS is any more of the business analyst. I think the real distinction is probably between R and SAS. SAS is kind of the legacy player in advanced analytics. Been around since the main frame era but SAS today I think is kind of where COBOL was in the late 90s. Okay, so that's good. Nowhere basically. Positioning right there. That's the quote of the night so far. Okay, but you mentioned that SPSS is sort of more, still, you're saying, maintaining the business. So everybody talks about analytics for everyone. Is it the intent of revolution analytics and are not to necessarily go after that space or is it to evolve there over time? Yeah, I mean there's really three pieces of this puzzle to make these killer data apps. There's the data. There's the data science and advanced analytics and then there's the user interface with the business application. Our focus is on the second of those. So enabling the data scientists to work with data to build these predictive models and then integrate those models into business user applications. Our strategy as a company is to enable those business user applications through partnerships. So for example another thing that we announced just recently is our partnership with Alteryx. It's a very elegant drag and drop environment for the business user to be able to connect their data to business to predictive models and to outputs but they don't need to program and they can also take advantage of the skills of data scientists who can deploy these advanced models into that nice drag and drop interface for them to use. What kind of metrics can you give us guidelines for people who are sort of following revolution analytics and are what kind of numbers or any kind of progress can you give us on proof points? Yeah, I mean probably the easiest thing to look at is that the R language itself is quite easily the most widely used statistical software today. Several recent polls, including the recent RECSA analytics poll have ranked R as being used by more than 70% of data scientists and it's the number one choice of more data scientists than any other software. It's just growing at an amazing rate. Now how about the company, can you share with us where you would headcount these days? Yeah, we're about 100 people so we've doubled our headcount over the previous year, doubled our regular news over the previous year and we intend to stay on that track. Awesome. Okay, we are here live inside the Cube, we're live in New York City for day one. This is our intro preview on the opening night of the Big Data NYC event here at the Warwick and across the street is Stratoconference and Hadoop World, the big main show as well as Big Data Week here in New York City. We're here for two more live days, wall to wall coverage, we're gonna be packed and we would not be possible without the support and generous underwriting from Hortonworks, Wendisco and other supports in the community. Thank you for your support. We'll be right back with our next guest after this short break. Thank you.