 We have Ron Bodkin, who's the founder of Think Big Analytics. Ron is an engineering dude, he's a geek. He built a solution with Hadoop and NoSQL. Previously, he was the VP of Engineering at Quantcast, which crunched zillions of events in their architecture. So we'd love to have scientists and engineers on. Ron, come on in and take the middle seat. Nice to meet you. Hey, thanks. This is an example of social media. We did a Twitter handshake. I said, hey, any startups want to come on theCUBE? It's walk-in time. Open mic night here at the opening night before the cocktail hour. Ron, thank you for coming on theCUBE. Thanks, John. Alex Williams, he's our senior writer in the services angle, DevOps angle and covering a lot of our cloud and coverage as well with Clint Finley. So tell us about yourself, what you're doing right now, your startup, what you've done in the past and then we'll get into some fun conversation. Sounds great. So I'm founder and CEO of Think Big Analytics. We started the company back in August of 2010 and I came to that because I had previously been VP of Engineering at Quantcast, which is a real pioneer in the big data space. Quantcast does direct audience measurement and real-time look-alikes for ad targeting. So Quantcast may have been the first company to put Hadoop in production. I'm sure stating that there'll be somebody who will challenge me on that, Hadoop in production in late 2006, so very early in the cycle of technology. And while in my time there, I led teams that did large-scale data processing, ultimately multiple petabytes of data process per day, did predictive models, both inference to figure out from anonymous activity on the internet, the characteristics of users, their interests and demographics, as well as predictive models for what are called look-alikes, for people that are likely to respond to ad campaigns at large scale. And the experience we had there of working with Hadoop using the Green Plum database for fast analytics on that data and doing predictive modeling led me to feel like there was a huge opportunity in the enterprise to apply the same techniques, which is why we started Think Big Analytics to be a partner to customers using these technologies. So I have to ask, because you worked at Quantcast, I am a web geek from 1995, the original web all the way through to now, and now ex-computer science product guy now running a media company, how come Quantcast and compete numbers are so different? Well, there's a big difference that... Tell us why. I want to hear this. So, you know, one thing Conrad Feldman, the CEO of Quantcast says is that our panel is as bad as everyone else's, right? Which is panel-based methodology really doesn't work well for measuring... You're better than they, they suck. Quantcast is hard, I mean, Compete is horrible compared to you guys. But Quantcast, the difference is Quantcast does a good job of directly measuring, putting a pixel up and measuring every activity on a website. Sure it does. You know, they... You're supposed to say it's really light, I was looking at the angle. If you tag correctly. Okay, Derek, he's our key guy, he knows all the tagging tricks. Okay, we'll get that done. Okay, so, you know, it'll measure, it should measure your audience accurately. But Compete also does extrapolation, right? And Alexa is the old school... They use panels, right? So they use small fractions of the audience. I mean, this is the revolution that big data brings is you can measure everything. You can count all the users. You can directly identify behavior, count it up, instead of taking a 1% sample and saying, well, we'll find a way of correcting all the biases in people that install a toolbar. So who's doing a good job in developing the visualizations of like web data that anyone can use? So there's a lot of innovation in graphical environments. You know, we see more of the innovation as it starts to flow into integration with data environments. So you've got on the one side pure plays. You've got companies like a karmisphere and a data mirror and some stealth mode startups that are providing visualization analytic tools for business analysts and data analysts in a big data environment. You also have a range of BI companies and the tableaus and click views and micro strategies that are integrating their tools in with Hadoop as well as like a Pentaho and a Jaspersoft. But then also you see a lot of people that are doing custom applications where they'll take the best and breed toolkits for visualization and put a services API in front of a big data environment to let them visualize. It's a space that needless to say there's constant evolution going on. So let's talk about your company. I'm just doing a quick hashtag here. Pump you guys on Twitter. Get the traffic up. Get you a little more traffic. Watch this guy run. I really like it. We'll analyze it later. The engineering. Well, that's good. I mean, here's what's exciting, right? I mean, we're all geeks, but a lot of the best new startups are by pure geeks where the innovations on the tech side. So let's talk about Think Big. You guys implemented a Hadoop back end with no sequel. Take us through that and what was the objective. What do you guys do? What's the value proposition that you're offering and talk about some of the tech behind it? Sure. So we think Big Analytics are professional services firms. So we help customers use big data technology, but we think that there's an amazing amount of innovation that's come out of the open source community. You know, hundreds of millions of dollars of investment around Hadoop, around some of the no sequel technologies like HBase, around analysis tools like R. So we are helping customers educate them on what they can do. What are the business possibilities? How do they use it technically? Helping them envision what's possible, put together a strategy. So you're a services company. Correct. You're not a product company. That's right. But because you build four customers Hadoop systems. So we build solutions. We help them set up environments. We build custom analytic applications that integrate data and ultimately do production processing of data. And we help them with data science, helping them get signal and models. Well, I don't know if you know this, but Alex is the editor of Services Angle, our new vertical pub. Yeah, and I'm curious how long these engagements are usually going. I'm hearing, you know, Hadoop engagements can range from eight weeks to 16 and longer. Well, you know, we see that there's a journey. So at Quantcast, we invested in this technology and spent years getting more value out of the technology. And one of the things that's exciting to me is that there's so much you can do as you learn and get more maturity around doing advanced modeling, getting predictive models, creating new products, more advanced ways of processing data. So we see a lot of opportunity for customers to keep improving what they're doing, expanding their use of the technology and getting value. So what limits any kind of services engagement is customers stop hiring services when they're not receiving the value. When they reach a plateau of usability and say, look, we're going to take this in-house and reduce our cost. So we are always striving to create value for customers because there's so much you can do here. We often will start with short engagements. We'll do service strategy engagements to really focus on road mapping, what we call a brainstorm. What can we do with this technology? How do I apply big data? What are the use cases that are going to really benefit in my business and what are the first things I need to do to get started? We'll then help customers sometimes with the proof of concept, what we call a jump start. And we'll... Not related to the jump start here at Strata, which is a... Unrelated. I'm not going to say who came up with the name first. How many... Yeah, most of the trademark dispute for jump start. The first use wins. No, seriously. How many employees you guys have? So we're a little over 25 employees at this point. So I asked Bill Schmarz at EMC and they have a huge consulting practice in there in the throes of all the BI nonsense around. Oh my God, I spent millions of dollars in trying to get to the unstructured side of the house and get predictive in real time. I'll ask them the same question. I'll ask you is, when you go into a client, okay, we say the burning house example. Every room's on fire, but one is really on fire. What's the biggest thing that you go after first? You got to put out that fire. Is it a fire? Is it an opportunity? What thing do you usually get sucked into first just saying, I got to address that? Well, we see a lot of people looking ahead and saying there's so much opportunity around big data that we can use this to really get a jump start, pardon the word, if you get a real lead on our competition that there's so many strategic things we can do by looking at data, by creating new products or being more effective in the way we operate our business. Now that being said, we also see a lot of organizations that take the opposite approach and say, well, we know we need to get there and there's a lot of opportunity, but a first step for this would be to help us get more of a handle on our data warehouse to let us manage the data more effectively. It's bursting at the seams and we have such a backlog of requests for things we want to do in the warehouse. So can we bring in Hadoop and start to supplement our warehouse and create a pool of data that we can then do some analysis on and start to add value to the business, ultimately building models and getting to the higher value use case. So hence why the service company, you got to come in and actually do some hardcore engineering. That's right. Pretty much. I mean, you got to come in, assess. Do you develop custom tools? We do. So we work with, we partner with the best and breed product and technology companies, but we fill in any gaps that are out there. So we do things like have environments for installing integrated cluster tools beyond just Hadoop. You know how do you install ganglia, Nagios, how do you integrate with systems monitoring? So deployment. Java. Yeah. So getting environments up and running reliably. We have, we open source a number of utilities. We have frameworks for data processing that we've used again to complement what's already out there and working well in open source. So we want to assemble solutions for customers, but we create technology where there's gaps in the market to make it easier for our customers. So are there instances where you're saying, ooh, I wish I had this tool and because it would do, it would be perfect for this situation. And then do you build your own or what do you do? Well, as much as possible, we look to work with partners that are investing in a space. So like if there's a gap in Hadoop or H-Base, we want to work with a partner to have them or partners to encourage them to contribute those capabilities into that roadmap, that technology. So we, likewise, we cultivate relationships with data science tools, with development tools. We want to see the space continue to rise and get more productivity. So we're not trying to become a product company. Instead, what we're trying to do is deliver solutions to our customers. And we think that there's a huge need out there for it. If I could just ask one more question then. What gaps do you see in the ecosystem out there? Well, we talked more about the BI tools and analysis and it's still an emerging space. There's so many use cases to be supported. Standardized reporting, making it easy for less technical business analysts to dive in and work with data. The tools that are out there today are more focused on power users. So making it easier for them to have access is one we see. We also think that there's this sticky nature that for unstructured data, which has really been a sweet spot for big data architectures today, you can't have simple point and click ETL. But as people start to do more and more in this environment, they want to have the right tool for the right job. They want simpler tools for working with the data where it's appropriate. And there's some investments, but it's still early days in that space. Okay.