 Okay, welcome back to theCUBE. This is live coverage of Hadoop Summit in Silicon Valley. In the heart of Silicon Valley, we are in the San Jose Convention Center for day two of exclusive wall-to-wall coverage of Hortonworks and Yahoo. And the ecosystem, the community of Hadoop and Hadoop Summit. The hashtag is Hadoop Summit. If you're watching, you can always hit the hashtag Hadoop Summit. We're tracking it, we're watching it. We're getting great commentary and great production from the folks out there. And feel free to ask us questions that we don't mind stirring the pot up and also answering your questions and asking the guests additional questions. Again, this is theCUBE, our flagship program where we go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE. I'm George, my co-host. Hi everybody, I'm Dave Vellante of Wikibon.org. Abhi Mehta is here. And Abhi is a good friend of theCUBE, a long-time contributor. Abhi is the CEO and founder of Trasada. And somebody who has just been an awesome visionary in this business. Abhi, welcome back, great to see you. Well, thank you so much. It's good to be back with the original crew from almost four years ago. You are the original tech athlete. I think when we were talking about theCUBE, we were always talking about being the ESPN of tech. When Dave and I and the team, Mark Hopkins, put this cube together, we're always talking about this. It's like talking sports to us. We're big sports fans or whatever you're interested in, having in-depth conversations. When we were at Hadoop World, the original Hadoop World, you were one of our most prolific guests talking about the Industrial Revolution and data factories. And since then, it's been such a great ride. We really appreciate your collaboration. You are truly a tech athlete, Abhi. And you're certainly a great guy, knowledgeable, and an entrepreneur. So you're out there getting dirty, building a business. Way too dirty. Yeah, you're only just leaves up being an entrepreneur. You're the janitor and chief executive at the same time. So a lot's changed. So I got to ask you. I'll see you're a tech athlete in our world and we've been following you and you've been contributing on our news desk. What's changed in the past year, in the two years? Absolutely, but first of all, congratulations to the two of you. I remember CUBE being a small desk at Hadoop World. And I had said that one day, the next time, in like five years, I'm on the CUBE, we'll have makeup people, you know? Getting us ready for the CUBE. This is a beautiful setup, so. Well, Horton works in our sponsors too. We don't have the makeup yet, so. We don't think our sponsors are underwriting. It's been great. We're working on the makeup and hair. This is, I'm very proud of what you guys have done at the CUBE and you truly are. You know, you've made watching tech really cool and a shout out to the crew at the back and Mark and Jeff and you guys are awesome. So what's changed? I think a lot has changed. I was talking to a customer last weekend and we sat back from a very interesting conversation on a unique application we are bringing to market. And he asks me, but you've been doing this, this is our third year. Truseta has achieved what I wanted to achieve in our three years. You know, we've done, Truseta, what I call the fun way. Some people call it the hard way. Like what you've done with the CUBE and your ventures. We haven't raised VC money. You know, we say the best capital is revenues. We have a lot of revenues. Our customers are our best supporters. We have an office in New York now. We've expanded to a new vertical. And we're finally seeing this shift. In 2011, there was still more skeptics than believers on big data. We've seen that change. In 2013, that number has flipped. There are a lot more believers in big data than skeptics. So that's the first thing that's changed. I think for all the big tech companies that we have witnessed, kind of validating the platform, there's a big issue with the validation we'll talk about in a second. But truly, everybody believes big data is here to stay and is a transformative force. I've said multiple times with you gentlemen, this is not a tech revolution. If we make Hadoop purely a storage platform, we as an ecosystem would have failed. And I'm not ready for that failure to come to our shoulders. So this is a business revolution, everybody gets it. The second thing that's changed is we are past the credit crisis. So there's a business change happening and we are hearing the business community now finally getting it. The reality has been, it's taken two years to push back the pain of the credit crisis and people are finally gasping for air and saying, all right, the bad news is behind us. I know I have to transform my business. Seems like this term big data could be that force, but I don't know where to go. So there's a realization in the market that there is a new normal. In the new normal, people like you, me and John as customers Dave have different aspirations, different behaviors. In order for businesses to monetize that, the old tools just don't work. The days of selling a database are over. Doesn't matter what you say, doesn't matter if it's proprietary, doesn't matter if it's six years old or a dupe, you can sell a database to any company because you can't monetize a database. What you can monetize is inside. So we are seeing the businesses saying there's a new normal, I need to understand what my customers want and I don't know how to do it. The third thing that's changed is the fact that people are finally realizing that the tectonic shift in technology, big data just being one of them, mobile, social, the commoditization of the analytics stack has firmly taken root and we now see it in the quarterly results of some large tech players. That's not sales guys not being able to sell. Sales guys aren't being able to sell because there's nothing to sell. There is no one answering the knock on the door. So those three things combined, social, mobile, commoditization of the analytics stack is putting a tremendous amount of pressure in three big areas. Understanding what to do, there's a new term being coined, it's called omni-channel. I don't know if the two of you have heard it, I would love to get your perspective on it. And an omni-channel is rapidly becoming way bigger than .com ever was. So just as we saw a rush in everybody doing a website in 2000. Yeah it's called hashtags. Hashtags, yeah exactly. People are now saying I want to be omni-present with my customer on a phone, on an iPad, in social, in online, and in brick and mortar. And people are saying that's part one, how do I reach my customer? Omni-channel. Number two, what should I collect from the channel? There's lots of interesting, the channel is not just a means to reach the customer, it's also a means to collect information. So I should be able to look at John and collect his mobile information, his social information, his hashtags, his tweets. So how do I collect it all? And number three, once I've got the platforms and collected data, what do I do with it? I know there's gold in Facebook likes. I don't know how to monetize it. So that's a change. Let me ask you a question. This is awesome. So first of all, we know it on the channel as we live it every day. This is what theCUBE's all about. We throw the confetti of data out there and it travels down the multiple omni-directional, bidirectional channels, omni-channels. But the Obama campaign really revolutionized micro-targeting, so macro-funneling, micro-targeting. This is the new marketing. So I got to ask you, because you talk to customers, you're only an entrepreneur and knowledgeable and prolific, but you talk to customers and that's really impressed with how you do that. But I got to ask you, when you meet people that don't get the omni-channel kind of concept there, they haven't really crossed over mindset-wise. Because it is a mindset. Absolutely. What do you hear from them and what do you say to them to kind of convince them that this is a really explosive, disruptive, innovative opportunity? And you hear it all the time. CIOs are on Twitter. How many times have we heard that? Or I don't get why, what you're coming from. What's that? Hashtag, is that a marketing thing? How's your value in hashtag? It's a communication revolution. What do you say to those folks to get them over the hump? There is, we have this really fascinating thing. We have built that to say that we call it a data pyramid. This is how I answer the question, by the way. And I think I'm seeing the skepticism change, John, this in the pyramid, John, what we did was something very fascinating. We built the pyramid and we overlaid the pyramid very purposefully in a way where the bottom part of the pyramid was existing data assets. What we call transactional data, identification data. It's extremely thick, rapidly available and every industry vertical, whether it's retail or banking or pharma, gets it. They go, oh yeah, we do that today. We know who our customers are and we collect transactional information about them. I said, absolutely, great. So they're kind of, I got a hook, right? Then we layer on top of it what we call interaction data. This is data that is, so the first two layers of the pyramid tell you who you are and what you do. There's a fundamental flaw that every customer agrees with me on, on transactional data analytics. Analytics has a fundamental flaw today. That flaw is when you use transactional data to predict the future, the issue becomes transactional data is telling you what you have done in the past. What you've done in the past is typically not an accurate predictor what you do in the future. A great example, Dave uses his credit card to buy beer and wine and shops late night because he's a partier. Now Dave is married and when you get married, Dave, you don't call your bank and say, yeah, you don't go out, first of all. But also you're not calling your bank and saying, hey, just let you know I'm married, you know? Don't send me the old offers. You don't call Walmart a target and say, well, I'm not going to be buying beer at midnight. Well, maybe you are buying more alcohol now, but that's your problem, Dave, you know? But the fact is, if you just look at transactional data as a retailer, the coupons they would send you, the marketing messages, we're based on a behavior that has fundamentally changed. You can observe that behavior through interaction data. So the bottom two layers tell you who you are and what you do. Interactions tell you who you do it with. At the top is what I call emotions. This is the golden nugget in analytics. Everybody's trying to predict emotions. And I can tell you, you cannot predict emotions. So golden rule of analytics, you cannot, statistics, you can't predict the future. So what can you do? You can observe emotions. How do you observe emotions? It's through hashtags, it's through tweets, it's through likes. What is the best predictor that you are gonna buy a Toyota or a Honda? Well, it's the fact that you tweeted, I'm looking for a Toyota. No analytical model will tell you you're looking for a Toyota. So when I draw that pyramid, they go, oh, I get it. You're telling me what I do today isn't bad. I'm like, yeah, you're right. You're a good smart person. What you're doing today is not bad. It's just not sufficient for the future. And how's that gonna change your business? So here's why it changes the business. Analytical models today are written within those layers of the pyramid. That's a problem. The next generation of article models have to be across the pyramid. So you have to be able to vertical. You have to be able to look at emotions and interactions, attach it to transactions to see what you actually do. Are you capable of buying a new car before I called you on the tweet? Yeah, okay, do you have the way with all of it? Two of them together. I need a new Ferrari. You got it. Not capable. Or when John is tweeting and saying exactly, when John tweets and says I want a Toyota, well actually John can afford a Ferrari, you should call him and say, John, no. You should buy the Tesla S, you know? John's jealous that Ray Wang just got a Tesla. So the three layers. That's the, that's the. So let's go through the pyramid. So layer one is what? Who? Layer two is what? What's layer three? Emotions, why you do what you do. So who what why? Why you buy? Why you buy? If you write analytical models that combine it, no CIO, no chief marketing officer, no head of retail will argue and say, I don't need that. They do need it. The problem is, you can't go and tell them to stop doing what they're doing. Number one, you have to then tell them, anybody can write models within the layers which they do. You got to write models across the layer. So this is. That's the transformation. So you mentioned a new vertical. Retail is a new vertical for you guys. It is, absolutely. Financial service is your traditional vertical. Correct. Now is this what you're doing in that new vertical? Yeah, so you know we've always been very deliberate in entering vertical because we believe that the future of big data isn't in selling commoditized tools and platforms because with search and fast SQL, those layers will be rapidly commoditized. So we are making a very, we've been, we had a very, the reason why we're in retail, it's a great story. We had a phenomenally powerful social marketing app and we couldn't sell it. No financial services customer would buy our social marketing app. I guess the reality was, well they were dealing with some other issues. So I got to ask you hypothetically, let's just say that I was going in to meet a venture capitalist and I wanted to explain to them what this is. How would you describe it to an investor? There's a lot of startups out there are pioneering omni-channels, multi-channels, using the new paradigm of communications and creating new value proposition. So what do you say to a venture capitalist who's evaluating, oh well we already have a horse in the track there. What is this, what do you would call this? I mean, because it's disruptive. It's not yet in a category. So what do you call this thing? Hey, I have a blank. We continue to call it an analytics application. The issue in the market is there's too much confusion among the world's platform and applications. They're too intermingled. No one knows what the one means. Ron Conway nailed it when he said that three years ago, if you had mentioned the word platform, VCs shut their doors. Today, platform is the hottest buzzword in the market. You don't have a platform, I'm sure, yeah. The issue is to answer your question in two parts. But the VCs aren't always a predictor of the revolution. And I was going to say that. Very few VCs truly understand what the future of this data analytics stack looks like. It may not even be a stack. And the very few VCs who understand that are willing to admit they don't understand what the future looks like. And I think those VCs, Palantir, could never raise VC money. There is money from Peter Thier. Why? And everybody's- Because he had vision. Because he got vision, exactly. So when we talk to people, we say there is a new suite of analytical applications coming about that will fundamentally rebuild the stack in a way that no one can define. The best way I define the stack, we draw a big box and we draw a drag line in the middle. And we say they'll be creators and they'll be commoditizers. And the VCs are going to pass. Exactly. The VCs will do- They'll pass. They'll probably do what- Not a big enough opportunity. I saw a lovely video on Nikola Tesla and the VC on YouTube. I don't know if you guys saw it or not. And he talks about his idea of wireless charging and he says, so is it a wireless play, an energy play, or a mobile play? He goes, that's way too much losing focus. The reality is, in big data analytics, it is a omnichannel play, whether you like it or not. It is a data play and it is a technology play. And how it is going to be fused together is up for grabs. Here's how I answer the question. The way I answer the question is that three questions are omnichannel that need to be answered, John, and all those big new companies to be built. Number one, which channels should you play in and should you not play in? There's tremendous confusion in the market amongst retail, especially amongst retailers who will lead the omnichannel charge. Should they have a Facebook app? Should they have a Twitter platform? Should they have a Pinterest app? Should they have a Foursquare app? What should the mobile app do? Tremendous confusion. That's number one. What platforms should you play in? Number two, what data should you collect from those platforms? No one knows. No one has a clue. Should you connect every single like from Facebook? Should you collect the friends? Should you collect the fact that I also like Ritz Cracker Falls? No one has a clue. And we're doing some fascinating work in that area, by the way. Where we have actually proven to the world what is a like worth. I can tell you what a like is worth. No one has a scam. The last issue in omnichannel is if you can now figure out the platform you're in, collect the right data, and figure out what is the right best product to offer you. What channel should you offer it on? If I know that Dave should buy a Tesla, should that offer be delivered over email as a direct mail in a beautiful black card with silver lining? Bannerad and Twitter. As a bannerad on Twitter, as a tweet, through your mobile as a pop-up, no one knows. Those three are massive issues in omnichannel that need to be solved. So the first one, which channel to play in this confusion? Second one, there's no clue about which data to collect. And then selection of products, the delivery and what's the problem there? Confusion. The problem is they don't know. No, the problem, there are two problems. Finding the right product offer to make to increase wallet share, make it the most predictive. So we are seeing hit rates. This is a statistic for you. We are seeing hit rates on using omnichannel well, where we have taken our customers on the right channels, collected the right data, figured out that person X should be made, at a segment of one by the way, X person should be made Y offer, booking rates average of 20%. When direct mail is 30 basis points, email is 3%. So the problems mount is confusion on the channel, those channel to play in. There's no clue about the data collect. And the third thing is they have no ability to deliver. So it was confusion, clue, delivery, availability. No, no, no, no option. Marketers will tell you that their email, their direct mail, they hit a wall. They can't squeeze any more out of that level. Every marketer that you talk to will tell you that. So they're looking for, and they know there's gold out there in these Amazon. And what they're also seeing to add to that point, Dave, is not only do they know there's gold there, they've seen Amazon go mine the gold. They're like, so you go talk to a retailer, and you ask them, what's their biggest competition? Guess what the answer is? Amazon, no one even mentioned Walmart anymore. Because Amazon, when you think of it, so if you were to ask me, what will be the most predictive, there's information arms race going on, which you've heard me say, no one knows the right platform capability you build. That's where Pivotal will struggle. Pivotal is going to struggle because the platform, the way they define it, isn't what a platform is. They're talking messaging, bus, and metadata. That's not the right platform to build apps on. There's different, and there's a different conversation. There's different platform capability and big data to enable apps. And Salesforce gets it, not Pivotal. It's a very interesting thing that we just talked about. But it took, you talk to them, and they see Amazon sitting there having the richest profile of what you buy. So Facebook knows what you like. Google knows what you search for. Amazon actually knows what you buy. Guess who's the most valuable data source? And not only do they know what you buy on books, but they know what you buy on groceries, and on shoes, and on, and so in the retail world, if you were to combine that, that's a mixture of Neiman Marcus, and Nordstrom, and Walmart, and Target, and grocery stores under one roof. And they have reviews, so they know what you like, and they know what you're searching for, and they know what you buy. And they have a platform where the marginal cost of doing it is zero. So then what people are scared about is Amazon has this thing called third-party resellers that they can sell on the platform. Guess how much money they make a year through that platform? Six billion dollars, which is straight profit. Right, the marginal cost, like you said, is zero. The third-party retailer is nothing. Because Amazon built their platform for the peaks, and they're saying, well, I'm not going to sit on it. It's the cost of the signup. Exactly. So that's where Amazon, Google, and Facebook, it's their battle to win or lose, and it's an information arms race, and other retailers are saying, we want that capability. The good news is they realize that it comes from Hadoop, they realize that companies like ours that can live with the king. So we have a retailer who is redoing their entire future infrastructure to deliver products and services, build products and services, and make the right customer offers Omnichannel on Tresena's technology, which you guys, hopefully, Jeff will write about in two months. That is exciting, because they say that capability will make me compete with Amazon. And that's the revolution that's coming. We are seeing more customers do that. So real quick, what did you mean by Salesforce gets it with Heroku and building apps? Yeah, so I think Salesforce realizes that in order to monetize any data, the platform capabilities starts with three core components. It has to be data on it. Salesforce has data on it. There needs to be certain infrastructure components that need to be free, that gets what comes with Heroku. And then there need to be certain enabling platform capabilities, I call them engines, that allow for an app developer. So if we all agree that the future of big data is segment of one analytics, how do you get to segment of one? You're not getting to segment of one by a metadata platform and by fast SQL search and linguistic search. You get that by data. You get that by data, but then an engine that can infuse the data at the individual level. We have that, we call it tree. Then graph is becoming the killer app. The killer engine on Hadoop is graph. Facebook has it called graph search. LinkedIn has it called PYMK, Twitter has it. So those two are becoming core engines. So Salesforce has that capability. Doesn't Pivotal have the data through EMC kind of? I don't know, I don't think so. I don't think EMC has data. EMC is in customers who have data. Yeah, exactly. So if you can pull those three things off on a platform, you'll get people to build apps. Guess who has it? It's Salesforce. It's Splunk. So those companies are more interesting than Pivotal. They're the new application development platform. Absolutely. I think Pivotal and Clutter will get there, by the way, but it doesn't exist today. Yeah. Okay, Abhi, great to have you on theCUBE. We really appreciate it. Prolific, it's exciting. You've got a lot of passion. We love that. And as you guys said, we're a tech athlete. And also an entrepreneur to say that congratulations on doing some cutting edge work. Thank you. Great to have you on, CUBE alumni. This is theCUBE. This is SiliconANGLE, Wikibon. I'm John Furrier with Dave Vellante. We'll be right back with our next guest here at Hadoop Summit Live for day two, exclusive wall-to-wall coverage. We'll be right back.