 Live from San Jose, California, in the heart of Silicon Valley, it's theCUBE. Covering Hadoop Summit 2016, brought to you by Hortonworks. Here's your host, George Gilbert. Welcome back to theCUBE. My name's George Gilbert. We're here at Hadoop Summit in the San Jose Convention Center. And we're here with Sri Raghavan, distinguished guest from TerraData. And actually fresh off a TerraData influencer some of it a couple of weeks ago. That's correct. So Sri, couple of things to talk about. TerraData we've known for having the best decision support database, big iron, you know, flex the performance muscles. They're changing their stripes. Tell us how. First of all, George, thanks for having me here and TerraData here. We're very happy to be here. We have a lot of respect for it. We keep on in theCUBE. So it's great to be back here with other people having participated in various parts of it. So TerraData has, of course, as you mentioned, we've been in business for 30 years and we started off doing a lot of work around the enterprise that our work has. We've actually had a very substantial, big data solutions as well as practice for the past five plus years. So we've been very active. We've been putting out solutions. We've got a lot of customers who have purchased our big data solutions. In particular, the one that I'm representing today in the company is this one called the Aster Analytics Solutions. It's what we call the multi-genre advanced analytics solution. And it is something we're very proud of and we've had a lot of customers. And so, you know, that has been central to the big data story that we've been telling, we've been trying to tell and we've been very successful with it. So that was, Aster became part of TerraData's pivot to big data when you did the acquisition, which was what year? The acquisition was in 2011. So it's roughly going on about five years now. Okay. Yeah. So everyone's talking about now analytics, you know, predictive analytics, machine learning. What's TerraData saying to its customers and its prospects about how it can help them on that journey? Yeah. So typically, you know, whenever we start having these conversations with prospects and customers, the question has never been about technology. We never approach it from the standpoint of, hey, look, we have this great technology called Aster and we've got this integration with Hadoop and we've got machine learning graph and text and sentiment and what have you. We typically try to approach it from the standpoint of, hey, let's first try to figure out what business problems a particular customer or an organization has. So it's the conversations of vernacular has always been in that vein. And our goal is to provide, to be the best advisors that we possibly can be to the customers towards resolving these kinds of questions. So let me give you an example, right? A couple of years ago, one of our big customers today, they had the first question they had was around, hey, a lot of our customers are churning. This was a big bank. And so, but they couldn't quite figure out what their rationale was behind all this churn that was occurring. And so the kinds of questions that we asked them was more around, okay, what are the different kinds of, you know, challenges that you have through which your customers interact with you? And they came back and said, well, you know, we have people who come directly to our branches. We have people who actually come and engage with us in through the web, through our substantive and substantial digital presence. They also talked about how people call it their call centers and talk to them and all that. So we did, we looked at, you know, these are completely disparate sources of data that come through these channels. And the idea is, you know, we started looking at the various paths that people take when they interact with the bank through these channels. And we did a path analysis, you know, we have our own patented path pathing algorithm called NPATH, which we applied to this issue to be able to determine, hey, what are some of the antecedent conditions to churn? What are the four or five things that people do before they say they throw in their hat and say, look, I don't want you to be a part of your bank anymore. So for us to be able to understand that there was a churn problem was very important, which then subsequently gave births to a number of business questions. For churn to occur, people must be interacting with you. Okay, if they interact with you, how do they interact with you? What are the different frequencies with which they interact? So then we started asking questions, where's the data located? What does the data look like? Are they all the same data? How do we bring them together? Then we start, that's when we started bringing the technology piece into it. So to answer your question in a fairly elliptical manner, it's about the business problem. Then we focus on the process and then it's only then do we care about the technology and bring that in. So how long did this engagement take and sort of what technologies did you have to bring to bear and who ultimately was the customer? You're right. So the question in terms of how quickly we were able to do it and of course one of the first unsatisfactory answers is of course it depends on the use case, but typically we've been able to resolve these things sometimes from a matter of hours to typically one or two days, where we're able to show some material difference between the pre-solution days where we went in and explained our solution to where we are today. So the amount of time spent is very short. Of course, there are many other use cases that get built on top of that and that extends engagement quite a bit in all that. So typically on average it's a matter of days before we show very, very good progress. Now the question as to who typically consumes it and again this is where it goes back to the whole point of leading with a business question because typically a lot of these questions are asked by the business. I mean, most of the time in the past when the conversations happen around big data technologies it's always how do we talk to the IT about delivering the solution? But as I'm sure I'm preaching to the choir and you've seen it yourself, a lot of these conversations are generated, they've begun with the business side of the organization. Guys are responsible for customer satisfaction or people who are responsible for getting customer retention or to increase new members. When I used to work as part of a JPMorgan Chase back in the day, one of my areas of responsibility was new customer acquisition. As a person who was responsible for delivering insights into customer acquisitions, it was my job to be able to recommend to the business, hey, these are the four or five strategies that we are able to recommend based on the insights we got from the analytics. In the case of JPMorgan you might not have had that technology bring to business. You would have had a path, a favorite or an optimized path, I assume. Chet, take us back to this bank. Where did you find the breakdowns where they were experiencing events that led to churn? Yeah, and the answer to this question is never obvious upfront, right? Because all you do know is as a point of visibility the only thing that is at least intuitively available to you to look at is the fact that somebody walks into your bank and says, I'd like to close my accounts with you today. So typically that is your first view into the extent to which the customer has been affected by your services or the lack thereof. Now, of course, there are many other waystations that the customer has passed with you in their interactions with you. So it could be that potentially they've come back at some point in time in the prior to that moment and said, look, I really am very upset with the fact that you charge me 30 bucks for this check which got bounced and I have $10,000 in my savings account. You can't really do this. It's terrible and all that, right? So having said that, then what they realize is, look, I know today has been the culmination of a lot of events. So that's when you start looking at what are those four or five events that have occurred? And that's when you start looking at there are, we have, let's say, five million customers. Now, obviously every customer has a certain path. You know, everybody's idiosyncratic in terms of how they interact with any given institution. But what we look at is among the five million customers, is there any chance for you to be able to crystallize five paths that, you know, it's a standard 80, 20 Pareto rule kind of thing where, look, five paths account for 80% of the churn customers, meaning that there are, you know, six or seven things that happen in these five paths. Every time we need to consider when a new customer starts going in the path, and we start flagging them and going. Here's the other thing that's important that the bank did, and I'll let you ask the question, but this is an important point. Not every customer needs to be given the same level of service, right? And the bank, sometimes, you know, the bank comes and tells us, look, we realize you're upset as a customer. We get it, right? But then you also have to be realistic. You are very upset with us, but you have $5 in your checking account. You overdraw your account maybe 15 times a year, and you don't really keep any substantive amount of balance. So really, the fact that you're upset with us really bothers us, but frankly, it is not worth our while to make you happy because where is the bank for the buck going to be in a sense that you're not going to be profitable over time? So these are some of the questions. That's when the insights are available to you, and you operationalize them by taking certain other things like, okay, what's a lifetime value of this customer? Right. So these are some of the things you do to be able to deliver proper value to the company, and we do a lot of that. Okay, and now you have a broader, a much broader range of tools. It's not just raw database. Correct. It's the consulting and analytics to bring these increasingly repeatable solutions. Absolutely. So you ask a very important question, George. One of my colleagues, really smart guy by the name of John Thuma, one of the things he tells me all the time, look, we're in the business of delivering what he calls human analytics. And the term is a great term. I've always thought of that, but I've never used that term, and he's absolutely right. Here's the deal. Our job is to be able to provide analytics, not just for those four and a half data scientists that you have in your payroll, right? So to that extent, we want to be able to have the entire organization be part of this data driven journey. What does it mean? Which means that, okay, I need the line of business manager to be able to develop certain decisions based from the data that she is able to see. I need the head of insights to also do the same thing. I need the CEO of the corporation. Now, how do you do that? Well, one of the ways by which you do that is to provide a solution where you're not restricted by your knowledge or no knowledge of being able to code in, let's say, Java, Perl, C++, Python, and so on, right? So the deal here is for you to be able to say, okay, as a company, are we providing certain apps that people with good visual interfaces that you can click and paste or rather, which you can drag and drop to be able to get to, being able to execute certain analytics which run in the backend. And so that's exactly the trajectory that we've taken as a company, to provide solutions where you have the ability to be able to implement them without knowing what the inner plumbing looks like. Okay. That's a holy grail for us. So let's leave it at that point as part one, and we will pick that thread up in a part two interview at our next show. But that is, it's a much more solutions-oriented message, obviously, with TerraData no longer being a database supplier, but a trusted advisor, and sounds really intriguing to get one level more. All right, this is George Gilbert. We're at the Hadoop Summit 2016 San Jose Convention Center, and we'll be right back after this. Thanks.