 The Cube at Hadoop Summit 2014 is brought to you by Anchor Sponsor, Hortonworks. We do Hadoop. And headline sponsor, WAN Disco. We make Hadoop invincible. Welcome back to The Cube. We're live here at Hadoop Summit in San Jose. I'm Jeff Kelly with Wikibon. My next guests are Ashish Gupta, who's the CMO and SVP of Business Development at Actian. And Michael Hiskey, VP of Product Marketing at MicroStrategy. Guys, welcome to The Cube. Thank you very much. Ashish, your first time here. Thanks for coming on. We appreciate it. Michael, you've been on before, and the Cube alum, you know how it rolls. So let me start with you, Michael. What are your thoughts on the show? 3,200 people representing 1,000 companies. What's your take? I love coming to Hadoop Summit. First of all, this is the best community event. You know, there's such a different feel West Coast, East Coast, Hadoop Summit, Hadoop World. I always find that interesting, but at both events, every year, they have the same talk track, which is kind of interesting. This is our year. You know, this is the year Hadoop goes bold. This is the year Hadoop kind of gets bigger and goes enterprise. And that's exciting for us, because we come from a typical enterprise background, and it's the integration that we're talking about here that makes it exciting. And Ashish, what are your take? You know, I think this is my first Hadoop Summit event, and it's really interesting, because we keep hearing about this whole idea of Hadoop and big data, and the fact that it is almost here, as Michael said. But it's almost like a person who's sitting with a party dress on, but not knowing which party to go to. And what we announced today is actually changing that. And what I hear from all the other vendors that I hear today is that this is changing. It's going from what used to be these unending POCs to now going into production. It's going from what we said with the fact that you don't need to use age-old tools like MapReduce to get to that data to SQL in Hadoop. And that's very exciting, because it's going to open this world up to a much larger set of users. And I think this is the year, most probably, to make that all come together, hopefully. Yeah, well, certainly I think, you know, we've, having covered this market now for a few years, one of the big knocks on Hadoop was it's, you know, you needed to have a PhD, you know how to do anything in it. You know how to write, and you have to know how to write MapReduce jobs. And it was really, as you say, unending POCs, where people were doing some interesting things, but it was often data scientists, kind of off in a corner with their little cluster, because those were the only guys that knew how to access that data. And now with Yarn, there's the opportunity to really open up the platform to other end users. And so talk a little bit about what you guys are doing and your approach, because it's an interesting one. You know, we hear a lot about SQL on Hadoop or SQL in Hadoop. Talk a little bit about your approach to that problem. Yeah, absolutely. You know, when you think about SQL and Hadoop, it almost comes down to the whole idea of the big data innovators problem, right? If you see what Christensen writes about the innovators dilemma, I think big data is going through that same innovators dilemma. It's because a lot of the market is locked out. It's just shut out of being able to access Hadoop data. And it's almost like the skills that knew how to access data. The SQL programmers didn't have the data anymore because they don't get access into Hadoop. And while the folks that have the data don't have the skills anymore, because it's very hard to find people who are skilled to get MapReduce code written to get into that data. So we've changed that entirely with Actions latest announcement with Hadoop SQL edition. The idea behind this entire announcement is that how do you bring industrial grade SQL in Hadoop allows you to get security, compliance, allows you to have data governance, allows you to bring in data from all different sorts of applications and then analyze it at a speed that is much faster than anything that's out there. But that end-to-end solution so that you can get the right insights is what we've announced here and allowing SQL to start accessing that. And that's why this partnership with MicroStrategy is really important. Yeah, well, absolutely. Because it opens up Hadoop to this world of SQL and all those kind of BI tools that people know and love and have been using for years and now can be applied to Hadoop and all these clusters that people have deployed but we're having trouble actually really leveraging the value. So Michael, talk a little bit about that and MicroStrategy's approach to Hadoop relative to yarn, things like Action and SQL on Hadoop. So really, there's two really interesting points here. Now SQL is the lingua franca of what business users know and love. And the business users, the interesting thing that we hear every day is that with our business clients they're getting more and more sophisticated with how they use data and how deeply they get into data analysis even at the business user level. But in the end of the day, I love coming to Hadoop Summit but in some ways this is the conference for unicorns. The unicorns are data scientists because here everyone's a data scientist and we see lots of data scientists on the job board they're posting for many of them but when we go out into the world we don't see so many data scientists. They're very rare, they're in rarefied air. So unless we can get Hadoop data into the hands of business users it's never going to cross the chasm where the rubber meets the road. So the way we see it is that data scientists can be able to purify their models and curate their models and have a way to publish those and get them out to business users because if they can't get them into the hands of the tens of thousands of enterprise business users those that are using MicroStrategy today in many cases they're never going to be able to get big data into the hands and make actionable insights. So at MicroStrategy we're not doing our own SQL in Hadoop capability at this time so we really have sought to partner with innovative companies like Actian that are also in the enterprise space are also pushing the envelope for big data analytics and have developed a great way to access SQL in Hadoop so that our users can get interactive access to their business data. So talk a little bit about the impact of bringing Hadoop into your ecosystem might have on how end users are using BI. So because if you're, is there an impact? If you're using BI tools to look at structured data maybe you've got a dashboard and you're looking at a cut of the data that you're familiar with with Hadoop there's all sorts of different data sources in there. How does that apply, how do you apply that or how do you leverage Hadoop with a BI tool? Obviously you've got to be able to connect it through things like SQL but are there other things you need to do to make it more appropriate for this new environment with all different types of data sources? It's a great, I said a really excellent question. The important thing is that for business users they don't really want to know where the data sits and they shouldn't have to know that. I talk with clients every day and clients tell me quite often that the to be data system of record will be Hadoop. Maybe that's not tomorrow, maybe it's not the next day but eventually all the data will live in Hadoop. In that environment they need to be able to have interactive access on it. Now today they want to be able to access Hadoop data alongside conventional data warehouse database data so the data can reside wherever it lives today. In a big data world the last thing you want to do is start moving data around between platforms but by separating out or abstracting that layer so that the business users using MicroStrategy can interact with the data in real time and the IT department can decide where and how the data is persisted. Really separates the two in such a way that it benefits both entities so that the IT department can dynamically orchestrate and provision data how it makes sense and the most economic platform possible and business users can interact with data using SQL in the tools that they know and love because changing a business user's application is harder than changing his relationship. It certainly is. So let's take into some of the use cases. So you've got customers across industries what are you seeing out there when it really gets down to well I was asked earlier in a panel I was sitting on like how do you recommend going about if you're advising an end user organization to start evaluating things and I say well the best thing to do from my perspective is talk about ask the vendor for some customers, some references that meet some of your requirements and if the vendor can't do that I'd be concerned. Now I know you guys have lots of customers and give us some color. What are they doing with the new capabilities you're building? It's really fun to talk about customer use cases. This particular one just steps out right on top of everything else. One of our customers, a telco customer was using CRM information to predict where that churn would happen. It was a mobile operator. And they had about a 1.1% churn rate not too high but when you look at it over 30 million users it's $160 million a year to reacquire those customers that churn. So they looked towards us and they said look CRM is a backward looking database. How about we've used some of our customer log data some of our network log data but all that sits in Hadoop. How do we get access to it? How does the marketing guy who's used to using CRM get access to that Hadoop data? And we solved that problem for them. Not only did we bring in customer log data we also brought in network log data. So when that call dropped with the geospatial location of that person added to it the social media intakes and outputs that customer was doing. And we allowed them to go from a 5% recognition to almost an 80% recognition of who's going to churn. They actually found 80 out of 100 needles in the haystack because they got more data sources. They were able to iterate that much faster and they were able to get the accurate timely results or the insights as you said so that they could save that customer. That is a beautiful use case. I mean they went from 5% accuracy in terms of predicting who's going to churn to 80%. I mean that's if you think about and if you apply that to the numbers you mentioned you have $160 million to reacquire those customers when you apply that and you can do the math. It's not a insignificant amount of money. That's exactly right. What about some other industries? Do you have retail, finance, what other things are you seeing out there? You know the Action Analytics platform is available to just as a Lego plate to everybody and specifically in various verticals like healthcare, finance and retail. What we found is that if you can get to the timely insight in the right time the next best offer. We work with Home Depot and Home Depot actually uses what used to be a three months worth of data and sampling based on these old systems that just couldn't handle that data to now with us using three years worth of that data so discovery analytics across a huge amount of information inserting in weather information to say this person is going to need a sump pump because there's a storm coming and give that next best offer. The Action Analytics platform allows them to do that. In addition to that in the financial services industry how do you reduce the number of false positives on a fraudulent credit card announcement? And that's an important balance because you don't want to upset your customers but you obviously want to prevent fraud so how are you helping them do that? So what they do is they utilize information that they collect in terms of where the person is where that cell phone is for that person using the cell phone to call your home suggests that you're very close geospatially to where your credit card is so you can reduce that false positive. The challenge though is how do you analyze it and how do you look at all that data in a timely manner and our platform enables them to do that. It takes all these data inputs, analyzes it super fast just like we're doing with SQL in a Duke the announcement we're just making we're about 30 times faster than Cloudera's impala. Now is it important to be fast? It's important because you want to be asking the questions getting the answers and asking more questions based on that so that you don't have false positives so that you don't give an ad that's not specific to the segment of one, that one user that's coming in and you don't have that customer churn and that's our focus at Acton as a company is how do we get those insights, those signals from all this noise of data? Well that's a challenge is the speed from the perspective of time to insight and the time you make the decision to invest in the technology and the approach to the time you actually can build and deploy an application that's actually going to answer a question. If you're trying to build something from scratch and you're trying to pull together different pieces of the patching the Duke ecosystem can be a challenge. I mean you need some smart people and it's going to take some budget and time and it's interesting here as this ecosystem kind of matures and new players that have been doing this for a while from a more traditional approach such as micro-strategy and acting to degree and then marrying that with some of the capabilities that the Duke brings to bear. It's really opening up some new possibilities. Absolutely. Break that innovators dilemma in the big data world bring new modern software you're running on modern hardware to give the right insights and our partnership embodies that but it opens this whole world of SQL users millions and millions of users right into the world of Hadoop which is where all the data is going to be sitting. And it's where they're going to get new insights that can move their business forward. And that's really what it's all about at the end of the day. I mean we talk a lot about the technology and the different components and it's but those are critical but they're critical enablers. It's ultimately about delivering the value. So that's exactly right. Well guys Ashish, Michael thanks so much for coming on. I appreciate it. Great to have you on theCUBE for the first time and we'll have you back many times I'm sure. I look forward to it. Thanks again. Please stay tuned. We'll be right back with our next guest.