 Live from the San Jose Convention Center, extracting the signal from the noise, it's theCUBE, covering Hadoop Summit 2015, brought to you by headline sponsor Hortonworks, and by EMC, Pivotal, IBM, Pentaho, Teradata, Syncsort, and by Atunity. And now your hosts, John Furrier and George Gilbert. Okay, welcome back, everyone. We are live in Silicon Valley in San Jose Convention Center for Hadoop Summit 2015. I'm John Furrier, and that goes George Gilbert with Big Data Analyst at Wikibon.com. And this is theCUBE, our flagship program. We go out to the events and extract the signal from the noise. We have two special guests here, the founder and CEO, Praveen Kankaria, and Anand Venigubal. You got that? You got it. All right, good. So you guys, head of street analytics, the CEO here. First, before we get into what you guys do, what's your vibe of the show here? You guys are seeing everything out on the floor. You're talking to folks. We're doing our best to share the data here on theCUBE all day, yesterday, today. What's the vibe? What's going on? I mean, consolidation? Not happening. Maybe it's consolidation, but this packed house here, bigger names, enterprise focus. What's your vibe? What's your take on the show? Share to the audience. What's going on here on the ground? I'm not so worried about the vendor community. What kind of consolidation's happening or not? I'm more fixated on the customers and the enterprise customers who are walking around here while we are talking here. I'm seeing a very different vibe this year. People are on to real use cases. People are not demystifying what big data is. It's not a small POC. Somebody should de-risk the company and figure this big data thing out and go in one corner, take half a million dollars and figure this thing out. Do a no use, throw away a discardable POC and come back and say that we can do it when we want to. We're way beyond that. We have real customers, real large enterprises walking around here who have, who are solving real use cases and creating transformational benefits. Hanna, what are you seeing? Same thing? Yeah, reality. Yeah. Real money. People see the value. The ROI thing is like, hey, you know what? Our ROI is we don't start doing something. There's no ROI at all. Exactly. So, you know, I just want to harp on one thing that Praveen said, it's real now. It's real now. The biggest companies in the world are spending real money seeing real value and hence buying. Are you seeing commonality between or even like categories of applications that are taking off? There's a bit of a bias in what I'm saying but I'm hearing streaming a lot, right? Yeah. Well, real time drives that. Yeah, real time. That's not, I mean, bias in the sense that streaming helps you guys. It is a fact, right? People are really enthusiastic about adding a rich customer experience, creating a rich customer experience by adding the most recent information and blending all of that with a powerful- It's interesting you said that because you know, even though you claim to be biased because that's an interest you guys have, you kind of look what you want to see but one of the things that I do in the queue at every event is I kind of have a mental model of if I was going to do a tag cloud in my head of what I hear the most and or top three conversations I'm in, you know, real time is probably what I hear the most. Yeah. Okay, besides Hadoop, I mean Hadoop, obviously would be the big fat font but real time is huge and that affects a lot of different things. The database, the in memory, streaming, geo, this is fast data, it's certainly data flying around and harnessing it is challenging because it's a whole new paradigm shift. So I would agree with you on that. Let's take a step back. Tell us about impetus of the company, speeds and feeds, employees, where you guys located, what's the company about? Then we can dig in some questions. So we are a product and services company largely focused on big data and as part of that we do, we help large enterprises deliver the promise of big data by way of services and along the way we have introduced products where we've seen massive gaps and the combination is actually turning into a very explosive benefit for our customers. So we go in there and we don't get out till the problem is solved. Now whether our product, Anand Venugopal here represents stream analytics where the stream analytics will create a head start and in some cases not. It's just a data at rest problem but needs to be solved with some smart algorithms crafted by our data scientists or just smart data engineering but we don't leave till the problem is not solved. And customers want that right now. It's services heavy right now because it is across the chasm. Some say the early adopters in the industry's across but that bridge to the future still is going to bring on the fast followers in the rest of the industry. So even Gartner says some good numbers, 50% of the people will look at Hadoop. That's still a massive number. That's everyone going to cross over the chasm. So it's not necessarily the enterprise themselves hasn't literally crossed over the chasm. So it's services heavy. So I would agree but I got to ask you what specifically in the services do you see the most of that will get the product market booming? Because at the end of the day the customers are building their products. They have a chasm to cross themselves. So the first thing is Anand mentioned from his bias perspective, real time streaming analytics is a big area of operation for us. That's where customers have burning use cases and we're able to go in with our tool set and smart people and we're able to solve, create proof of value in a very short amount of time. One voice over IP company called us and I'll give you half an hour. This thing is really good. Can you really solve this problem without writing a line of code? And we went, we were able to do this. Ranging to pretty large use cases where people are looking at offloading the traditional data warehouse because that's choking. The price points are not feasible, not viable for them to scale further. I got to ask a question that's going to lead into how you guys relate to the big data space in Hadoop. Kind of as a setup to that. And it's kind of a trick question. So take it for what it's worth. Analytics, a process or a product? It's from a customer standpoint. Process. It's a means to an end. The end is not analytic. So next step is, well to me you can argue both sides but process improvement right now seems to be another, not on the tag cloud in terms of most buzzword but what I'm reading out of the trends is everyone's transforming their businesses end to end, right? So like big data impacts every part of the business because you're measuring everything. So if you can instrument something, you can improve it, that's the thesis. If you believe that, then the question is, okay, people going to scratch your head and say, oh boy, I can really make a difference. So the process changeover is across the board. So I see analytics first addressing process and then product develops out of that, outcomes and whatnot. So what are you guys seeing in that transformation and how are you guys working with customers because they need to trust their partner? Right. They got to have a tech person, they got to have a team and it's a full on engagement. Yeah. So I disagree with calling it just for sake of process improvement. It's about improving anything and everything because if you can gather data on anything and everything and analyze the data, you can improve. So we have customers, hardware companies who are improving their products because earlier they were not able to analyze all the data they were gathering. The data was sitting in some cold storage. Now they're able to analyze and get the analysis to their engineers who can improve the products in their very next release cycle. So in other words, you could improve a product or a process that you can instrument it. So what I'm trying to say is I agree with you and yet I disagree that it's much larger than just process improvement or product improvement. It's about product. I guess I really disagree with myself but no, that's the point. This is the beauty of people. This is the beauty of my big data. Beauty is in the eye of the beholder. If I have a problem, I call my problem, I call what it is. I don't say, oh I need MapReduce or I need streaming. I just got to go in and figure out how to architect that so I got to, you know, either I have staff on, people on staff that can do that or I got to find a partner. That's why I'm focusing on the services piece because at the end of the day, who are you calling, if I'm a customer, who do you call to do that, right? It's like, okay, so that's kind of my next question. So take me through an example. I'm a customer, say, hey, I just had a big board meeting blow out with my team. We had a debate about processing products. We've got 10 different opinions on what to do. We all know we got to go to the next generation. Come in and help me. What do you say to that? Yeah, sure, I'll be right over or whiteboarding. Take us through the engagement process with you guys and the customer and take us through that historical journey there. Sure, the beauty of impetus offerings is that no matter where the customer is in their evolution cycle, we have an offering to start them where they are and take them all the way through the journey of realizing business value. So with a major airline, we're doing use case discovery. Right? You're doing use case discovery. Use case discovery, right? Where are the opportunities for them to improve customer experience or revenue or shape cost off, right? And prioritizing them. That's a pretty unusual way to get into a customer where they're like, help us figure out what problem we should solve. That's like trusted advisor territory. Exactly right. That is the territory we've been playing in since about nearly two decades now. We have been trusted partner to our customers and it is the journey is continuing in the realm of big data. And you might be surprised, but that is true. They actually want help in when you bring wisdom across industry verticals, people want to hear about not just their industry vertical, but what's, you know, a telco guy wants to know what the healthcare industry is doing and how can they take their lessons from there over here. That's what we do. We help people cross learn from across the ecosystem. And some of these are horizontal best practices that can be applied to. Well, as best practices as customers want to partner, it's going to help them figure things out. Exactly. Because a lot of things are happening in real time. Look at Spark, for instance, next week, we see a lot of tsunami of activity on Spark. And then the people will be speculating, customers will be getting seeing all this greatness and like, hey, I want that, that too. And it's not like, Spark's not a shiny new toy, it's relevant, right? So we can see that. You can obviously go, okay, I can get that. But now operationalizing Spark. It's a whole different bottom. But John, that comes much later. The first problem is what is the use case I'm going to solve? But then I'm going to solve through storm or Spark or some proprietary platform. That's very secondary. I think if you go talk to enterprise customers, and particularly the executives who have to go into that board meeting and justify where they are and how are we differentiating compared to our competitors, they have to, they're not dealing at a Spark or storm. Yeah, they don't talk that language. Whatever. They're dealing at, you know, what problem can I solve? Yes. How fast can I solve? And how much of a benefit can I produce? That's why I agree. I think that's why it's challenging when customers come to these early adopter industry movements like Padoop, for instance, which has been great. The linguistics are not connecting. It's like the language of the customer is, hey, I might get fired, or this is the psychology of the customer. I might get fired, or if I'm successful, it might fail by growing too fast and now I have more costs that could bloat it out now for, okay, now I'll take that success because the board says, double down. Yeah. So these are the challenges, an architectural challenge, if you will. And we have seen behavior changing in customers based on business priority, right? They become less attached. Like just like Praveen said, they become less attached to the technology. You know, we have seen across the distros, there's obviously polarity, right? And even in the customers, right? Some one distro supports Spark, one other distro supports more of storm. And there are customers who know, we are putting a call center analytics case in front of the customer and saying, hey, we solved this problem. Here's a sample of the screenshot. And the customer, no matter what distro they have, like, I want that. I want because I have that problem. Who are you coming in? When you come in with these, you know, business solution, like case studies, who are you calling on? And then, you know, then who do they introduce you to to say, go now, work with them? Very good question. So we've seen multiple models. We've seen some very large, and I'm talking about very large enterprises, Fortune 50, as an example. You go in and in some companies, the business leaders are driving this whole process. We had one studio called us, this executive called us that and said, we're releasing a new movie and we don't have our analytics in place. So somebody told me, you guys work on Hado. On what? Hado. Hado. Hado. Hado. But the guy was looking for his problem to be solved. Now, he doesn't care whether it's Hadoop, no sequel date, whatever you bring to bear. Hadoop connected it for him. He sees the buzz for Hadoop. Somebody told him that. These guys are pros. Call them. They know that Hadoop stuff. Can you get us some Hadoop? But then, what was the step? You said to him, okay, we can, you know, help you figure out like the demographics and who's receptive to the movie. Then what is the step from there? Who does he connect you with? To actually go do it. So they have people who are one or more levels below who are more hands-on at a business level. But is it traditional IT or is it like a rove? And then IT comes in. IT definitely comes in. So where I was going was that business leaders take, start driving the process and then bring in IT. But we're also seeing another very interesting model which actually is very hot and I think it's spelled very good for such companies. IT leaders are calling us, hey guys, we need you to come in as a partner and help us ferret out use cases and sell these, the solution for these use cases to business so that they can give us more fun. And so that they don't lose relevance. They don't lose relevance. They're outselling to business. So which is very hot in the monstros. Yeah, they just want to sell that through because they're champions. I mean, this is great. And I love this about early adopter markets where it's growth is fast. The stakeholders who are your champions know what to do. Then they got to kind of heard the cats internally to kind of get it in soul. And they're going to have questions. Wait a minute, we have an Oracle system that does that. Why would we want to move off a relational database to handle this unstructured data? There's no value that's been proven out of that yet. Prove the value, then we'll, well, so there's always that circular hole you dig deeper and deeper of the no ops who say no, no, no to everything. So that's the first step you're saying. So get the use case identified, sell it through, then what? Make the first use case a success. And in a major credit card company, that's what happened in early 2012. In three and a half months, we took their first use case, which was a real-time offers and recommendations engine into production, and I have the date, it's June 12th, 2012, and then we went live. And it was a three and a half month assignment and they were dramatically surprised. It was a 51 order root cluster. And after that success story, we've now gone on to implement what 20 or 30 different use cases successfully in production. Over three dozen use cases for that same customer. And the cluster is now over $2,000. So I got to ask the CEO question for being to you and to come in and share with us the top three conversations that you have that you're having with customers and what specific large enterprise, not small, large enterprises that, and why are they going, and what is their top concern? So top three conversations you're constantly having, if there's a pattern, you can identify. And then like the top enterprises, what's their mindset, what are they working on? So one is a lot of customers are dealing with pure data engineering problems. It's not even analytics. They're drowning in the amount of data. So they just need a better database, a data infrastructure to deal with all the new data that's coming in. And everybody knows. Is it the amount or the fact that they can't sort of make sense of it, the cataloging almost? So dealing with data, getting the data, ingesting the data, cataloging. Yeah, data management, that's not a dupe issue. It's just like. They find it when they need to x years from now. Yeah. So this is one area. The second area where people did not just come up looking for a solution, but as soon as you show light on this, is this area of real time analytics. Yeah. And we just. And that's something that they haven't seen before, right? Every other day. It's a very big jobs like kind of mindset. They don't know what they want you because they haven't seen it. Absolutely, yeah. And we're actually, we're surprised by what we're learning on a daily basis. Every new prospect conversation is teaching us a new application, a new use case. Oh wow, we had not thought of this. And this morning, Anand actually gave a talk on a topic which was given to us by a customer that they used our streaming analytics platform as a big data enterprise services bus. They actually coined it for us. That's nice. And he went and spoke and he had 150 people in the room. It's great when customers do your marketing for you. Absolutely. It's fantastic because you're so successful. All right, so let's talk about the streaming services and the data warehouse modernization. Because that is a trend. Streaming's been out for a while. We've seen it in real time. But now we're also hearing here, you know, offload, moving off, migrating out of data warehouses. Because there's now coexistence between large data warehouses and then the new needs for, I'm seeing new use cases that I've never seen before, whether it's internet of things and or other different, what was considered weird use cases now going mainstream. So talk about the streaming and the data warehouse modernization. Talk about both? Yeah, talk about both. We've got a couple of years left. There's obviously a relationship between the two and I'll draw that line, right? So it all started with the adoption of Hadoop. It's great. People have become open to adopting open source and now they're looking for a streaming technology that's compatible with Hadoop, right? And the Kafka storm or Spark stack, either of those are designed to sit conveniently on Hadoop any distribution. And we're currently based on Kafka storm moving to Spark streaming, adding Spark streaming support as well. So it's a powerful stack and we are able to actually overcome conversations with where the competing technology is proprietary in nature. People have become so friendly to open source and they've been waiting to get liberated by from proprietary, the hold of proprietary vendors that they will, they're positively biased towards open source oriented technologies, right? So we are a streaming analytics platform based on open source, but we've solved the problem of deep skill sets and complexity that people need. We make it drop dead easy visual to put together applications to the point where in a month we were able to produce 12 different screens and use case scenarios in a month for a major telco. And the beauty of what we produce and the functionality what we produce rocketed it up all the way to the corporate CIO in terms of the presentation, right? So that was, that's streaming analytics, dramatic productivity to achieve real customer value. Can you give us a quick sort of recap of the use case that required both the big data, you know, the data warehouse and the streaming, so where, you know, where they came together? Sure, in this case we're talking about use cases that essentially needed streaming sources like a set of box data coming through and you're taking the set of box data, you're matching the customer ID on the box to who the customer is, checking whether that customer actually, for example, opened up a maintenance ticket or not. So when you call in, John calls in, the call center agent already knows that, oh, this is John, he's probably a pretty frustrated guy right now because his set of box has been misbehaving. And the first thing the agent tells you is that, don't worry John, if you're calling about the set of box issue, we've already figured that out. We have a replacement job on the way Wednesday morning. So the streaming, the data quality is real time, so it's not stale data. Oh, he called two months ago to complain or about his bill. It's actually, no, no, I get other data sets unified in to that moment of the touch interaction. Right. Engagement or whatever you like. So you're able to now blend real time data with the historical data and provide context sensitive service to the customer. That supports your systems of intelligence thesis that we're working on. Okay, we're getting the book here, but I want to give Praveen the last word. So what's your next move? What are you guys up to? Website, how do we folks contact you guys? What's next in the company mission? What's going on with the next steps? So for us, I think we see delivering the promise of big data as an opportunity of a lifetime for us as a company. We're investing in products and services and the proof is our customers, large Fortune 50 companies, willing to work with us. We are not a Fortune 500 or 1000 or 2000. Not a reseller of our, yeah. And yet these large companies are very comfortable in working with us. And it's solely because of the value we can bring. One other trend which has taken hold in the last five, seven years is information technology has gone on the revenue side of Fortune 500. Prior to that, it was largely seen as a cost center. Yeah. 70% of your cost is to operate your business. IT was a cost center. And now IT is on the revenue side as well. And that's where, so IT has become an area to differentiate the company. Is it any particular department? Is it like marketing first or? We've seen this in marketing first, but I think we're seeing it all across. Anand mentioned this credit card company which actually, I'll share a story. Another financial services company, they had a massive rule-based system for fraud detection. And we actually helped them build a new system which is based on supervised machine learning. So they could discard all the hundreds and hundreds of rules that built, which were not relevant for anybody and everybody, to a fully supervised machine learning-based system where the system is constantly learning from their own data. And they have cut down false positives tremendously. The system paid for itself in four months. Okay, Praveen, we got to get in the wrap here. Thanks so much for sharing your insights. Congratulations on your success. Thanks for coming on theCUBE. We appreciate it. We are live at Hadoop Summit here for three days of wall-to-wall coverage at theCUBE. I'm John Furrier with George Gill, but we'll be right back with more coverage right after this short break.