 Live from the Fairmont Hotel in San Jose, California, it's theCUBE at Big Data SV 2015. Okay, welcome back everyone. You are watching theCUBE. We are here live in Silicon Valley for Big Data SV in conjunction with Hadoop World and Stratoconference. This is theCUBE, our flagship program. We go out to the events and extract the seeds from the noise. Our next guests from RapidMind are really successful fast growing startup, fresh funding, Michelle Chambers been on many times. Welcome to theCUBE, president and CEO at RapidMind and Ingo Mirsva, CEO, doctor, I should say. Welcome to theCUBE. So congratulations, we're going to get the signature here. The famous author, Michelle's many books but it's going to get this quick signature at the end here. You're writing books, you're starting companies, you guys are growing. You guys have the hot hand right now, the Scuttlebutt here at the show in Silicon Valley is this Boston company is kicking butt and they're not from Silicon Valley. So that's awesome. We're on fire. You're on fire, you're rapidly ascending up and but Big Data has been in your blood. You guys have been scratching this itch for a while. So give us the update on RapidMinder, the company, the story, the fresh funding and what's going on with the product and why are you guys so excited and why the successes is coming very rapidly to you guys. Well, should I start? Sure. It's really almost a crazy week. Everything's happening this particularly. The product release on Tuesday, we had the funding release yesterday. We are on strata those days. You could say really be on the roll. So what's happening right now, so in the Boston area really it's not just snowing there. There's a lot of cool stuff happening. So it's really the time that's actually Big Data and Hadoop is becoming a reality. If you compare this to like a year ago, only like 10% of all the people have been asking for pushing down machine learning into Hadoop classes but now it's everybody. Everybody wants to do this and wants to bring it actually into the end of their business users. And that's the biggest change in the last couple of months. If you just, well, you're just in the middle of this storm, that's what's happening. So it is a storm. Certainly in Boston got tons of snow and you can't stop joking about that. The joke we heard earlier in the week was if the patients just admitted to deflating the balls, God would stop snowing. And so of course we're huge patriots even in California, but in all seriousness, Boston has always been a hotbed. Dave Vellante and I run SiliconANGLE Media. We have our office in Marlborough. So we always talk about this. Boston has a great DNA of systems guys going way back during the 80s generation. You look at what DEC did. You look at all the stuff on 128, now 495. It got kind of caught in the dot com bubble, but now you're seeing a resurgence this new generation, data analytics and dupe. It's all systems guys and like math geeks. This is like a good time. So yeah, are you guys seeing the same thing? I must say a lot of action going on in Cambridge and all around. So is it, am I misreading this? Is it like getting it right? So we're in Cambridge and I have to say it's the absolute hotbed for big data and for analytics companies in the area. And you know, when you look at the differences, I think between East Coast versus West Coast companies and a lot of the heritage does go back to the DEC days and kind of the whole generation that got created out of that around enterprise business. The way I always characterize it is I find that East Coast companies are very grounded in solving business problems. You know, something that is very commonplace in a financial services, a telco, it doesn't matter what industry, but it's very grounded kinds of stuff. And what you have happen out here on West Coast is a lot of innovation, but you don't know what problem you're actually solving typically, right? And so what's happening now in Boston is really convergence of both of those, right? There's innovation that's for the sake of innovation, but it's also solving then being very quickly applied to real world kind of business problems. So one of the other things that Ingo didn't mention is that this week we also are in the Gartner MQ in the leaders quadrant again, for the second time in a row and the Forester wave is going to be coming out pretty soon and we'll just, you know, see where we land there as well. So we've got a lot of good stuff going on. So a startup to be a leader, that's a real accomplishment, that means a lot. I know Gartner, a lot of people always look at Gartner as kind of the guidepost to there. So congratulations, that's an awesome thing. Thank you. And first of all, we love startups because it's really action. So I got to ask you on the funding. So you just talk about the funding, how much cash did you get? What are you guys going to do with it and the plans? So it's a $15 million round. So it's led by two Boston based firms, Ascent and Longworth. And also our insight investors, EarlyBird, which are really one of the top firms in Germany where we originally are from. So, and OpenOcean, those are the guys behind MySQL. So we bring together really the top European firm, the understanding of open source business models and then really now new American investors to the table as well. Okay, so talk about the product now. So you get the product release. Why are you guys so successful? Obviously in successful startups, you have the team, you guys have a great team. Product market fits important. What is it about your product and the market where the fit is? What's the key success there and what's that enabler for you guys? So let me talk a little bit about how we view the market. In the advanced analytics space, there are very traditional players that have been out there and that market has been dominated predominantly by SAS and SPSS and even R, which is a 20 year old technology. And when we look at that market, that market was really dominated by single compute engines, by mostly structured data and doing ad hoc or offline kind of analytics. Very little of it went out into production and the types of people that we're actually using that were mostly mathematicians. So quants, actuarials, statisticians and what you see unfolding in the marketplace, when you have a market that is that large, you typically would see continuing innovation but you can see in those markets that the innovation is actually slowing down but at the same time you have a whole host of new startups in analytics space and a lot of that got fueled by big data. It's also being driven by the IoT space and around cloud and what you see happening there is that market's a lot, lot larger and it's because what everybody's trying to do is get to the business analyst. And everybody talks about these magical data scientists as though there's one ilk and really there's two flavors in my mind of data scientists. You have developer data scientists which is what you have a lot of buzz around out here and those are the people that are doing R and Python and Scala and then there's really the applied data scientists and the applied data scientists are people who are gonna use existing algorithms to solve real-world business problems and drive real business value and Hadoop is maturing to the point where these data lakes are really interesting but now you have CEOs pounding their fists on the desk saying I absolutely have to have the value out of it. And I've spent all this time and money building infrastructure which show me the money, right? And so what you want- You mean from the data lakes? The data lakes are sitting there. Okay, we did this, we captured everything. Now we need to get some action out of it. And I can't just go to 10 data scientists in my organization and get the value out of it. I want to empower hundreds of people in my workforce in the line of business to create that value. So data lake is a service in a way, right? Enable the data lakes to be accessed by normal people or business people. Access is the first thing but I like the fact that we are really talking about getting actions out of this. So it's not really like, okay, I get another insight can base my decision on top of that. We're talking about predictive analytics here so we are predicting the future. It's a very complex topic. That's why we have those data scientists in place. But if they become a bottleneck, nothing is happening. Well, there's two things that we see in the queue coming out over and over again. One is there's a lot of confusion around what a data scientist is. And two, there's a spectrum of different skill sets. And then also there's an operational issue of how do you operationalize a data scientist role? And then the last thing is there's not many data scientists out there in the pure sense of the word. We're talking about quant job, math, unique creativity. I mean, when spreadsheets came out, it was like, okay, I just want you to spreadsheet. So to us we see data scientists, you can't scale that and have the operational efficiencies. So is that true? And then what do you guys do to solve that problem? Now we're coming to the product. That's exactly the problem we are seeing in the market. This bottleneck, it doesn't scale. It's not getting operationalized. And it's even a really bad decision to invest into data scientists and then using them for what we call table stake analytics. So there's no point in actually taking those highly paid resources and letting them do the boring stuff. Or the non-high yield least expensive steps. So you're saying use the most expensive resource, the pure core data scientists to create an enablin model with this leverage for average people that the tasks are important, but you don't know the ROI yet. So let that develop. And rapid miners really, there's interface making this happening. So we're taking those, well, traditional data science tasks, make them available to business analysts, data-saving business users. So everybody's like becoming a tableau for advanced analytics really. That's what rapid miners is. I mean you probably have a lot, I'm sure you have a lot of happy customers coming to the table saying, hey, thank you. So what's the status of the customers? How many customers do you have? How do you guys price it? So we're an open source product and then we have commercial packages that start out with freemium models. We have over 600 customers at this point and our open source total community is a quarter of a million users. And when we talk about our community, we're talking about people that are using our software for an average of five to six hours a day. So these are hardcore people that are- So open source code, is it open source as in the source code or are you just offering it for free? So we have open source where the entire source code is available and then we also have freemium packages. So we use a different model. It's a new open source type of model. So there's the traditional open source that's the like Hortonworks, right? Where it's support and services. Then you have open core models which are value add IP components just like Cloudera has. And we've used both of those models in the past and currently we're on what's called a business source model, which is really very new open source strategy. And what that means is that your current release is a, your commercial version and it starts out with a freemium packaging and then your trailing release, your last release is then open source. And so that's a really key part. So it helps us to continue to drive development. Okay, so let me get this right. So business source is you offer a free version of commercial use, no source code, just package, right? For use, right? And based on business, natural business value drivers. Yeah, and so they use it, okay, I'm playing with it. And then the next release of that package is open source. Or is that, did I get that right? You got it backwards, okay. So the current release, which happens to be 6.3 is our commercial packaging. And there's a starter one. And that is limited in terms of the number of data sources you can get to and the amount of memory because that's where you do your analytic crunching. And then the packages up from there are paid packages that get rid of those governors. And then the last release, okay, is a free release, complete lease open source. Okay, so it's development of more in it? Yes. Integration applications. And that's the algorithms and whatnot. People continue to extend our package, not just through algorithms, but also through connectors and through pre-built models as well. So we call those accelerators, and we offer, we have pre-built accelerators in our platform, but then we also, our community contributes back in terms of accelerators as well. So we have 100 outside of the company committers to our open source project. Yeah, I mean, why not make the goodness available in all formats, right? Absolutely. And then get paid for when it's working, right? And so your triggers of memory and data source, that's your little product marketing feature that ratchets your revenue up. Exactly. Ratch is not a good term, but it increases your value. That's more value. Yeah, so one thing which is really interesting about this whole model and I call this really, it's a community-based model is what are the great things you can do with this community? So if you have 250,000 users and they're using RapidMiner on a day-to-day basis that many hours every single day, can you maybe learn from them? Can they give something back than just money? So, and this is exactly what we are doing. So we have a concept inside of the product we call Wisdom of Crowds, because I said before that we are giving RapidMiner to the hands of business analysts and data-savvy business users, but that doesn't automatically make them into data science heroes. It doesn't transform them right away. So can we do something to help with this transformation? And this concept of Wisdom of Crowds we invented is doing exactly that. We're looking into how the 250,000 people are using RapidMiner, how are those analytical processes looking, and can we learn something from them and give recommendations to all the new users we are getting, and that works very, very well. It's like on Amazon. And do the users opt-in for that? Is that just part of the terms of service? They are opting in. Do they know they're being used? Yeah, well used, but so they are part of a huge family. I won't say exploit it, I'll say leverage for the benefit of society, right? Well, it is not that. It's like Wikipedia, it's exactly the same thing. You're contributing and you're getting something back and that's exactly how our users are feeling. So it's kind of like the open source ethos, the more you put in, because they can take the value out themselves, right? So you're taking the gesture data of all the interactions, kind of building the best practices for folks, algorithmic. In fact, it's the biggest knowledge base for best practice in analytics in the world. So you're taking advantage of the crowd, the crowd interactions and the usage, turning that into gold for everybody else. Exactly, so it reduces this data science skill gap, it bridges it, so it turns people into those data science superhears. So we're doing this by recommending what are the best data pre-processing steps, we're recommending what are the best models to use, so this whole model selection problem is going away. It's very disruptive and, Michelle, you know, we've talked about this in the queue before, and certainly Dave and I have talked about it. You know, back in the old days, you'd have to get a reference implementation, hire some firm to come in, scope it out, what do people need, a long timeframe, you pay a boatload of money, and it's long, and you get a couple of use cases. This is our best practice. Now, you guys have this tapped into the crowd. That flywheel's rolling. And it's free, but everyone benefits in it, and you get a better product features. It's like how it's awesome. Right, so we use our own machine learning techniques to create these recommendations, and what happens is everybody that we go in and talk with about it is like, oh my God, this is like gold for our business analyst, right? Because now, the most novice business analyst, somebody like me who just graduated out of an MBA program can come in, and they can immediately add value to an organization, to a line of business. But also, even the most seasoned analyst or data scientist can actually learn. Because now what happens is, you know, we're creatures of habit as humans, right? We kind of do the same old things, right? We get into our grooves, right? And it's just, that's the way we are. The way we're wired as human beings, right? So now what happens is, this is not a black box, it's very transparent. So the recommendation says, ah, the next step that most of the people in the crowd would do is whatever, let's say it's a logistic regression, that's the step they would take, and you're like, oh, well that's not something I would have thought of, okay? Or maybe let's try that, okay? And then oh, by the way, it comes up with recommendations on how to actually fine tune that model. So the reality is, once people build models, they usually have to fine tune them, and oftentimes they just don't even bother, so that doesn't increase the accuracy and therefore the value to the organization. But now what they do is get recommendations in line while they're building it on how to actually have the optimized model. So instantly the company gets value out of it, and gets premium. So collective intelligence of all the other things in context to their tasks. Absolutely. So its recommendation engine meets kind of collective intelligence, if you will. It's exactly this idea, yeah. Perfect description. Well, it's awesome, so it is. Can I get the part, no, is that okay? We would love for you to have the product, John. By the way, this is all I think about 24-7, by the way, with our CrowdChat product as well. No, the crowdsourcing thing is a big thing. That is a huge deal, in my opinion, because that's going to give you more source data for product management, product innovation, and that's fantastic. So you're on to something huge there. Yeah, so we introduced this feature capability late last year, mid last year, and we've continued to add to it. And so this year you'll see additional recommendations that will come out with along the same lines of leveraging the wisdom from the crowds. Okay, so that's awesome. I think you guys have a great, great success. No wonder why you're in the leader quadrant. Great stuff there on the positioning and the market. Awesome. Technology innovation, the enabler that you guys have. What is your core disruptive enabler from a tech standpoint? Is it the in-memory, is it the spark, is it the product? What specific, under-the-hood things that you guys have that are making all this stuff work? So we run in multiple compute contexts, which is amazing to me, right? Because when I look at, I've been in the analytics space a while, right? And everybody has sort of a slice of the answer to really what the market wants. And what the market wants is the reality is people have mixed environments in their data centers, and they're going to continue having that for quite some time. And so what we do is we handle all types of data, which is a precursor to big data, but it also includes binary text, images, you name it, it really is all data. And then what we do is we run in-memory, we run in Hadoop, so we do the push-down processing to run in Hadoop. Now that's pretty magical, right? Because today, if you want to use analytics with Hadoop, you've got a lot of coding people that are doing some pretty hardcore coding. And now all of a sudden to empower a whole new constituency around being able to extract the value out of their data lake is huge. We also run push-down in-streams processing. So we run in Apache Storm, and we've got in-the-labs Apache streaming, or excuse me, Spark streaming as well. But it's not quite stable enough to release the GA products, but we are toying with that as well. Streaming's fantastic. Streaming, and the real-time space is just super hot, and it fits the adaptive machine learning techniques that we have extremely well. The other thing that we do is we do elastic compute up in the AWS cloud for cloud execution, and then we also run in database. So whatever your environment is, we kind of work inside of that kind of context. Yeah, that's fantastic. So you have the kind of nice bit, so it's really the idea and the execution has been a key part of it too. You guys kind of mapped out some of the use cases, like just the hassle of Hadoop in general is a lot of maintenance involved, a lot of sysadmin, a lot of standing up some gear. So making it go faster in that process of the lake is set up. So you guys help people who are setting up these data lakes, right? So that would be safe to say. Absolutely, and as you alluded to earlier, there is no magical data scientist that's out there. You have to, in an organization, it's in context of an entire team. And so one of the things that we do is we facilitate collaboration so that you can bring together line of business people, data scientists, IT folks, all together business analysts to collaborate because it's a very visual environment, right? And that makes the analytics that accelerates the time to value for organizations. One of the things that we do is in addition to providing these guided self-service analytics, we actually provide something we call accelerators. And accelerators are predefined, and you and I have talked about this before, John. So I know you're going to be all over this, all right? But these predefined use cases. So we have built into the product things like customer churn and direct marketing and sentiment analysis and predictive asset maintenance. And what's the beauty of this compared to a packaged application is that with advanced analytics, it's not very generalizable. It is difficult to completely extract the value because it's so closely tied to your data. And so what these accelerators do is give you a 60 to 80% fit, and then you can tailor them to your specific data so that you can maximize the value. But at the same time, you got a quick hit in terms of being able to start somewhere. You can look at it. You get a quick taste. You get a quick taste. You get a directionally correct vector, and you can understand it and then decide what to do. It could go into production. It could go into production from there. So it's that, it's that. Versus growing the Hail Mary and will the answer come back? Am I, oh, I screwed up the whole thing and you waste a lot of time. That's the alternative, right? That is the alternative. And you actually can always lift the curtain. You see really the underlying process and you can use this as a starting point. So for integration, because honestly predictive analytics, when you think about this, it's not about feeding more information into the one big decision maker and then you make this one big decision. That's it. The biggest value of predictive analytics is if you have millions of what we call micro predictions. And you actually automate and operationalize those processes. But in order to do this, you need to deploy the analytical process. So the accelerator is a starting point. You like the result, you like the outcome. And if that's the case, you use the starting point, just press a button and you get into deployment mode. And that's another thing which I was missing for 15 years now that actually deployment was as simple as a single mouse click. Because it's ridiculous. After you built all those processes and did all the work, if you need another year to get to deployment, the value is almost nothing. The lag is, we'll kill you in the lag. The lag will kill you. Well, they bring up a good point. In the old days, you were restricted, but now with the new technology, I'll see with unstructured data and now all these platforms, it's rapidly, so you can do new things. Compute's not an issue, right? Get in memory, that would spark it. It's pretty bad ass. And so you guys have nailed that. So I got to ask them, well, that's first of all, I think first of all, the signaling concept of bringing in, most analytics folks are like, old school, okay, I'm gonna attack this corpus and pound it all day long and send some queries in. But now you guys are doing something interesting that I see some successful firms doing in other areas. I've been to go in the social sales space is they're signal driven, right? So they look at the signals and they use them in context to doing things in the main area. So you guys are analytics. So this wisdom of crowds, looking at getting a percentage of accuracy to get a look at it, iterate through it. So the iteration, the signaling is the new kind of thing that's happening right now. Yeah, so what's happening, it's a great point because the traditional analytics space was very slow to change your models, all right? And it was the frequency in which you did that was just unbearably slow. Now what's happening is everybody wants to essentially have adaptive models. They want to be able to very quickly look at model shift, model drift, because the reality is, as data is coming out of this fire hose, your model gets outdated relatively quickly and it depends on what your space is. And what you want to be able to do is very quickly detect that, update your model and then push it back out into your production environment. Well, you can have that chasm that we just talked about between development and production. It has to be eradicated in order for us to move to that type of environment. So what's happening is the traditional SEMA and CRISP models where people were developing models in this very water flow approach is really going away. And what's happening is people are using much more agile approaches, which really means that you got to be able to experiment and get fast to fail models, right? And most environments don't allow you to do that because they're not really designed to do quick prototyping. Yeah, and you're predicated from either outdated provisioning hardware, infrastructure models and or old software development paradigm. So you're screwed on both fronts. Exactly. Well, that's why the cloud's so awesome. I talked to Andy Jassy at Amazon. We talk about experimentation all the time. Experimenting is critical and I think you guys have a great model so I'm stoked about it. And I believe that's the right direction. I think the new way to do things has to be fast. And that's why the streaming thing's interesting to me because now the data flow of access of data quickly is part of the evolving data fabric, if you will. The data's got to come in and blend in with other data and that's going to be the key to good analytics and accuracy basically. So we have a couple minutes I want to get some general questions. So congratulations on all that stuff. Great conversation, we're going to do another time but I'll ask you about some of the things going on in the industry. So EMC just announced a data lake foundation. Are you guys involved in that or are familiar with it? What's that about? You guys didn't get serious? Is it good? Are you involved at all? What's going on with that? So the open data foundation that they just announced is one where it's interesting as a consortium. We're not involved at it at this point. It is something that I think that the industry has mixed results with because here you have an open source community and then you have this consortium that looks like vendors trying to drive their agenda where on the other end you have open source which in theory is not being driven by any agendas. Now the reality is open source is- Is on agenda. It has its own agenda, right? So yeah, so there are agendas going on but I think that it's going to be interesting to see whether that actually succeeds because I think it has a perspective, there's a large perspective around it's a heavy handed approach to what has been really a very strong open source driven community. So you think there's a bolting on a foundation just to kind of get in the game and see what happens? Okay, that's good feedback. Well I got to ask, so on the customer side, how should CIOs be thinking about this next gen environment? If you go in and talk to CIOs, obviously they have a lot of legacy where you know that. So okay, I got legacy, not a full rip and replace but this transformation going on it's a journey as they say but I'm a CIO, look I got to drive new apps faster, I'm going to convert your infrastructure as fast as I can. There's all this data fabric stuff going on with these open data platform stuff. I don't want to deal with none of this nonsense or this nonsense or pure do this and that. I just want to get the apps out the door and scale them up. Wei, how do you guys fit into that conversation? Well, it's actually a perfect fit. So let's start with what is the big idea of big data is actually not just having more data than we can handle because that was really always the same problem. It's that we actually do this on commodity hardware, that we actually do this at a reasonable price. And that's automatically something that CIOs are interested in. So yes, it's the speed of delivery but it's also at the same time at a reduced cost because otherwise the whole scaling wouldn't work at all. And so RapidMiner fits into this naturally because, well, we are making use of those bigger data structures, data lakes, whatever they are. And at the same time also, we are speeding up this whole development cycle down to deployment. So a CIO would really be foolish if they're not realizing, okay, this has now totally changed. I have the opportunity for every new project, for everything around unstructured data, for everything where Hadoop is playing a key role here to actually invest into a new platform here. And also this is a starting point for a migration path. But a migration is not happening overnight. It's always a journey, as you just said. So Ingo and I talk about this as the big data candy store. So when CIOs come in, they often say, I want to displace a particular vendor in the analytics space that's pretty large, okay, because of their cost structure, right? And they come in and say, are you coming to be my knight in white armor and save the day for me, right? And they're like, and of course, typically our salespeople will have said yes, right? And so now we get set up for, all right, we gotta come in and be the shining knight, right? And so the first thing I do is I say, that probably isn't happening. And they're like, really? And they're like, aren't you here to sell me? And I'm like, yes, I'd love to sell you, okay? But the reality is that you have organizational change in an organization, and people are going to be resistant to change, wholesale change like that. So what you're much better off doing is having what I call the big data candy store over here. So you have your legacy environment over here, let people continue doing what they've always been doing. And over here in their big data candy store, okay? You give them new tools, all right? And they come up with learning. And then the candy store, which kid, where do you want to be if you're a kid? Exactly, exactly. That's a good metaphor, you know? Subliminal, you know, messaging, yeah. So you come over here in the big data candy store, and now people get comfortable, they start seeing how easy it is to do things. And what's now going to happen, instead of having a push strategy in the organization, you wind up with a pull strategy, okay? Where people are going to say, can I use this for that type of problem as well? So land and expand is your growth. So just like Tableau grew and Splunk and other winners, land and expand. No big monolithic sale platform. No one wants another platform like they want a hole in the head these days, right? I mean, you walk in selling a platform, it's just like, you know. Well, and it's the reality of, I mean, if you're moving from any technology to a new technology, there's risk associated with that, right? And so as a CIO, you've got to pick your battles, right? And decide where you're going to fight those battles and what your risk is appropriate for your organization. So final question. What are you going to do with the money? You've got 15 million, you've got to spend it. Obviously consortiums can be a distraction, certainly. If there's a big plan, you know, come by our grid for customers, you'll jump into anything that works, but you had limited resource, 15 million is not a lot of cash. Certainly with growth, you'll be, you'll be, you know, doing some stuff, but even that's not big dollars. Uber just raised a billion dollars. They're still private, but I mean, 15 million. What are you guys going to do with it? What's your priorities? Obviously product and some customer growth. Is it the standard blocking attack when we start up? What's the plan? Yeah, we're continuing investing into the product, of course. And this is really something that is important for us because we need to actually stay ahead. We are ahead right now, so we want to keep this position. So that is one important thing. Another very interesting aspect for us is building out our ecosystem. Predictive analytics, as I said before, is as strong as if it's embedded, if it's operationalized. So, but in order to do that, you need a strong ecosystem around RapidMiner. And, well, we started out of Central Europe. Been very strong there for many, many years, moved here into the United States only last year. So we are still in the beginning. So you got to build out some marketing. You got to build a community here in the U.S. You got to get that core base up on the commuter side. So we have a very aggressive community plan this year and that's what part of the funding will go towards is that RapidMiner is pretty unusual in that we've had a marketplace for quite some time and we are going to be giving back even further. So one of the initiatives we launched this year was RapidMiner Academia and that really goes back to our roots and foundation out of universities. And so now RapidMiner is available to professors, students, researchers at no cost. So there's a lot that's going on in the company, not only around product innovation and of course building out the sales and marketing engine here in the U.S., but then also going back in terms of give back to the community as well. And you guys are going to stay in Cambridge. That's headquarters for the... Even with all the snow, we're staying in Cambridge. No, it's global headquarters. Global headquarters in Cambridge. Not a bad place. Cambridge is a hotbed. Yeah, MIT. Cambridge is great. And then great universities in Boston. And we've got great talents that comes out of our networks there as well. So. All right, go Boston. I'm a big, big fan of Boston Tech, so I always root for them. Great entrepreneurs there. I think the VCs could up their game a little bit, but don't want to digress this interview. That's another topic for another time. Thanks for coming on theCUBE. I really appreciate it. Guys, congratulations. Michelle, thank you very much for coming on. RapidMiner, great success. Rapidly rising. Data mining, analytics. It's hot. And again, it's only going to get better. It's early days. This is theCUBE live. Extracting that data and sharing that with you. Thanks for watching. We'll be right back after this short break.