 At Big Data SV 2014 is brought to you by headline sponsors WAN Disco. We make Hadoop invincible and Actium accelerating Big Data 2.0. Okay, welcome back. We're here live inside the queue. We are in Silicon Valley for Big Data SV. I'm John Furrier, the founder of SiliconANG and Joe and Jeff Kelly from wikibond.org. Our next guest is Amila Frem from Neo Technologies. Welcome to the queue. Thank you. Thanks for having me. We met when you guys were a couple people in the valley. John Callahan and True Ventures introduced us and boy, what a whirlwind for you guys. Back then we were kind of talking, high-fiving each other, saying, yeah, this graph stuff is going to be brutal. This Facebook platform could take off. It could take off. It could take off. I think there's a pony in there somewhere. Billion users, Actium users later. Just one example of many, and we just had Joe Halestin on from Berkeley, which by fact, they're talking about graph databases. You get nodes and edges and math kicks in. Graphs are a wonderful thing in computer science, as we know. So congratulations, very relevant, and as an entrepreneur, great success. Give us a quick update on the company, kind of where you guys are now. Funding, staff, what your plans are. Yeah, so we're active in the graph database space. We're the most popular graph database out there today. If you look at any kind of objective measures, at least, in terms of number of deployments, tweets mentioning us, which I know that you guys track, things like that. It's pretty evident that Neo4j is the leading graph database today. In graph databases is this alternative way of looking at data, where you embrace the relationships. That's the key difference between a graph database versus any other model out there. So you have nodes, and you have relationships between nodes. And then you have key value pairs that you attach to both the nodes and to the relationships. Turns out to be a very effective way of managing complex data. And any of the big web properties out there, the Twitters, and the Facebooks, et cetera, the world, have all implemented their own proprietary stacks. They had to in order to get to the scale where they are today. Now Neo4j exists as an off-the-shelf solution. And it's used by, this is new for us, it's used by 100% of the Fortune 1 company. That's right. That's true. That's right. You can't argue with that. So Walmart is using us, for example, and 35% of the global 2000 is using us in production. In production, not POC. No, I mean, it was funny, because we're also open source. We have a community edition, which is available for free. And we have an enterprise edition, which is on a subscription. And a lot of open source companies throw up a lot of beautiful logos on their slides. And it turns out that they're actually just users, not customers. But now, so we have hundreds of customers, 35 plus, that are in the global 2000, and are using us for real mission critical stuff. And tell us about the business model. So when you say customers, you've got the open source edition, essentially. And then you've got, I guess, what you can call an enterprise edition, and package some services around that. Sure. Yeah, so sort of the philosophy in my book is that when you run an open source business, you want to segment out the people who are more time than money, for the people who are more money than time. So if there's a student or a hobbyist out there, you can try to sell them on it, but it's never going to work. Like, if you have a lot of time, you can work around any. It doesn't matter if you're open source or if you're proprietary, you can work around it. As long as it's software, you can work around any mechanism in there to protect the software. So the people are more time than money. You want to give them your stuff for free. But then the people who have more money than time, i.e. big companies, you want to give them a valuable offering and try to get them on a subscription hook. And that's what we have with our Enterprise Edition. So really, the key differentiation between the Enterprise Edition and the Community Edition is clustering. So with the database, it turns out to be really important if you have production deployments to have a cluster system. So if one machine goes down, you don't want the entire website to go down, for example. So that's available in the Commercial Edition. And as well as we have a high performance cash, and we certify it so that it's good for certain operating systems, and of course, also support. It's really hard enough for the Enterprise and make it a viable option. So talk a little bit about the evolution of the company. I mean, you guys have come a long way. John mentioned you guys met a few years back. Where are you in terms of size of the company? You mentioned customer attraction, which is great. But give us kind of an update on where you guys are. So we're still small, although when John and I met, we were tiny. So in comparison, we're bigger, right? So we're 60 people today, spread out That's a decent size. When you go, you see some of the companies that these at Strat, at smaller events, there's some very small companies. It's actually not bad. So we're 60 people, and we're spread out across 11 countries, just because as a CEO, I enjoy the pain. They have a big fund advantage. You need a graph database to manage all that. We do, we do need a graph database. Luckily, we have one. And headquarters here in the valley in San Mateo, where we're 25, 30 people, something like that. And so yeah, so that's sort of the state of the unit. So you get some good traction here at Strat. I'll read some tweets from our crowd chat monitoring that says from Russell Whitaker, it looks like I may have an actual immediate need application for Neo4j at work after seeing Emil's presentation at Data Science Strat a conference. So hey, congratulations. Thank you. You get some instant traction right out to, you know, spreading your wisdom out there. Next one, they had a guy, Dan Woods, ask the question, what is the downside if any of graph analytics engines loading data from anywhere for analytics? Mm, the downside if any, loading it from anywhere. Well, I think it's one of those usual problems where like crap in, crap out. Right, if you take all of your data and put it in there, you know, if you don't have valuable scrub good data in there, you're not gonna get good answers. And I think that's sort of the key perspective on that. I think that, I agree that the traction strata this year is really good. The way that I actually introduced my talk at Strata was that it turns out that graph databases, when you look at sites like DBEngines.com, which track traction of all the database projects, graph databases grew the fastest of any category in databases, bar none in 2013. Grew the fastest of any category in databases, bar none, which is pretty impressive, right? At the same time, it's the fastest growing one but at the same time, it's also probably the most misunderstood, mm. Because people equate graph databases with social. They say that graph databases are only good for social. And the dangerous part of that is that it's partially true because graph databases are really good for social but not exclusively so. So we have a lot of traction in financial systems, we have a lot of traction in financial services and in telecom, in retail, we're used for recommendations, fraud analytics, network management, geologistics, shipping, there's like a wide range of use cases. Why didn't you? That quote, by the way, got a lot of retweets. You must have said that on your presentation because Paige Roberts from Acti and said, the biggest misconception about graph databases dashed are only for social dash Neo4j. Yeah, that was the entire title of my talk actually. Okay, but let's talk about why not social? Give some specifics, social is graph-based relationships, et cetera, and unstructured and what loosely structured. What are the besides social? Well, so to be clear, social is a great use case for graph databases. It's not just the only use case. Yes, exactly. Phenomenal for social, check the box. Check, yeah. And that's what I mean. I think like Zuck has popularized the notion of the social graph, right? So people know graph equals social to them. I think that makes perfect sense. I think that's why people think of it in that way. But then, I mean, really anything where the relationships are important, it turns out that actually my perspective on that is that the difference between information and knowledge is relationships. Like what your viewers are doing right now when they're trying to figure me out is that they're looking at data about me and they're trying to gather data about me. And if they're like name, first name Emel, last name Ephraim, age 35, that tells them something about me. I call that data, discrete data, data on me or flat data. But in order to really understand me, if they understand that I grew up in Sweden, which is why I have this Swedish accent, right? I live in California. You good at Winter Olympics? I'm good at Winter Olympics. I have a wife called Madeleine, a daughter called Nomi. I work at Neo Technology. Like all of those things is what gives me color. It is what makes you understand me. And that's all about the connections between the unknown concept and the known concept. So if that is true, if the difference between information and knowledge is in the connections and the relationships, it's mind blowing to me that there's not a database system out there that fundamentally embraces relationships, right? Except for graph databases. The graph database do do that, which I think is why they have this sort of wide appeal. What's your biggest challenge right now with your business? Obviously, you're in a good spot, relevant to the market for social and data in general, because data is great for, you know, we talked about the shape of the graphs. There's also computational upside of having the ability to have arcs and edges and nodes. And so graphs are a beautiful thing from a math perspective. So it's obviously very relevant. So what's your challenge, what's your opportunity that you're going after? Yeah, good question. So it used to be that the big challenge was the market perception of graphs. And that is still a challenge. But for example, today, Forrester released their take radar report on the operational database management, enterprise DBMS. And in there, they predict that graph databases will be used by 25% of enterprises by 2017, 25%. I mean, that's pretty significant from two guys at the Starbucks Cafe in 2007, right? We shouldn't invest in it. I mean, we shouldn't invest in it. Which is pretty cool, right? So that used to be the biggest challenge. It still is a challenge. I'm not delusional, or I probably am. But in this particular case, I understand that most people don't yet know what graph databases are. Having said that, I do think that we made significant strides in that one. The biggest challenge for me right now is very simple, hiring. Finding awesome employees. We're looking for data scientists here in the valley. We're looking for engineers worldwide. We're looking across the board. And that is, I mean, I'm sure, mirrored by a lot of companies here in the valley. And it's a real challenge. Good luck with that. And it's really hard. It's very competitive. Obviously, Facebook and everyone else wants these guys too. So it's a startup. So let's talk about that. So you're in San Mateo. You're going to move all your operations. So I'm going to be in Sweden. You have other development centers around the world. Talk about the global aspect of it. Yeah. You guys are funded from Europe. Yes. To European BC. We have a European heritage. You know, I grew up in Sweden. The company grew up in Sweden. We still have Malmö Sweden. It's our engineering HQ. And we have all of our engineering in Europe, except we also have the one guy in Kuala Lumpur and one guy in Auckland, New Zealand. And it's a little bit spread out. But engineering, sort of the gravity of engineering is in Europe. And then the gravity on the commercial side is here in the valley. But as I mentioned up front, like we're in 11 countries just because I enjoy the pain of managing a global organization. If we could just go back to some of the use cases and some of the application and the different industries. So the forest report, 25% of enterprises are going to be using graph in some form. So that's going to span verticals just by definition. So what are some of the things that you're seeing in, you mentioned, financial services. What are some of the nodes and edges that people are trying to understand in that market, for instance? Maybe we could talk about retail or others. So let's start with financial, then go to retail. So financial services, there's a huge leading investment bank in finance. Typically I can't talk about the customers, unfortunately. But a huge leading investment bank that all of us would know the name of. And their problem used to be that when they onboarded a new trader, it took them two weeks until that person was up and running and had access to all the assets, the media assets, the documents, and the collateral inside of the company. It's a heavily regulated industry when it comes to access to those kind of things. And it turns out that when they did root cost analysis on that, it was actually because the software platforms they were using, it was so difficult to compute whether new trader got access to this particular resource. Because if you look at that as a node and edge kind of thing, node and relationships, the new trader and the new employee or every employee is a node. And every resource is a node and resource being a document or something like that. But then it's not just a direct relationship between them. It's the trader belongs to a group. The group belongs to one or more groups. The content is aggregated into folders and collections. And you have all these kinds of access control list relationships between them. It ended up being a many, many multi-way join in relational database, which is a minutes response time or an hour response time. With a graph database, it's millisecond response time. So in that example, it was like the files and folders and documents were nodes and the individuals were nodes and then whether they had access and if they belonged to different groups and how they're organized were the relationships. Interesting. All right, so let's take, we hear a lot about Ritza. How about something like pharmaceuticals? Pharmaceuticals, we don't have a lot of traction yet in pharmaceutical in sort of the traditional one. We see a lot in bioinformatics. And bioinformatics turns out that when you dive into that, there's a lot of very graph data. Actually, in our body, there's a lot of graphs already. Someone, be it evolution or a divine being or I'm not gonna go into that. Someone or something has already figured out that a graph model is a very efficient way of representing information. In fact, that's how our brains are structured. Neurons, synapses to other neurons that are connected in a big network or a graph. Turns out that inside ourselves, protein networks are graphs as well. And when you start working with that kind of data, it's all very graphy. Interesting. Now a little controversial use case, but obviously the NSA story is out there and the use of graph type of technology to understand that terrorist networks and bringing in all that metadata, I mean, it's all related to the NSA scandal and the things that they're doing. Does that have the potential to give a kind of a bad name to graph databases and some of the use cases? Yeah, I mean, I think it does. And I think graph databases are a hammer. It's an ethically neutral technology. It's a tool, right? And with a hammer, you can bang someone in the head or you can build a house for them, right? And if the first, you know, umpteen guys that you see with the hammers start banging you in the head, you're not gonna like the hammer after a while, right? So I think that for us, I'm personally, let's say not a huge fan of, you know, the NSA type applications. Having said that, I think for us, we need to lead with and make sure that all the amazing things that graph databases enable are the stories that people hear first. Knowing that it is horizontal technology, it can be applied in a good way and in a bad way. Yeah, I mean, I think, you know, even beyond just kind of the graph database market, I mean, do you feel like it could have the potential that whole scale to, at some point, turn to the, I guess, the commercial sector, the industry that we're in? Because, you know, right now, a lot of the IRS is directed at the government. But, you know, as the public starts to understand a little bit more about all the data that's being collected on them, you know, from my perspective, it's an issue as we as an industry need to get out in front of. I mean, what's your take on the potential for that to impact kind of your ability to do your job and? Yeah, I mean, I couldn't agree with you. Appreciate the question. Yeah, I couldn't agree with you any more. I do think that it can completely reflect on us. And I think, you know, with good reason, because, you know, it is a neutral thing, like amassing all this data and then analyzing it and being the tool builder for all that is a neutral thing. It can go in both ways. And I think the answer is to have an intellectually honest debate about it and talk about what do we as a society feel are the right things to use these things for and the wrong things. And then just myself or my company, I try to steer us in the direction so that we don't engage with those kind of things. And that's sometimes doable, sometimes not, but that's sort of my own personal moral compass on that issue. Yeah, I mean, it's one of those issues. Tech is an enabler and you can enable positive things, negative things, it's how you use it. And that gets to the heart of all those kind of decisions you make inside a company, for instance, about how you're going to use data. Just because you can do something doesn't mean necessarily. Should, yeah. So talk about the plan. I mean, I see a lot of things to navigate, hiring, you mentioned, getting some quality people, obviously, Key. What about funding? What's your plan? Looking for more funding? What's the next step? Yeah, so we raised $25 million to date. We have lots of cash in the bank. We're in the fortunate situation where we get some funding from our customers, which is the kind of funding that I like. It's called in, value, creation, gets paid. Yeah, so that's good. We have significant runway. If we stop hiring right now, we can cash flow positive if you want to. Having said that. You want to grow. We want to grow and the market's exploding and we know that the big guys are entering the space. We know that, I'm not going to mention names, but all of the big guys who are currently invested in the database space are coming out with graph databases. We know that, well, it's public now that Teradata launched what they call a graph database. One could argue, technically, whether that's true, but what they call a graph database as part of their Aster story in End of Q3 last year. So that is happening. So now is not the time to take it slow. So we are investing and we want to do that. We may even race another round in 2014 and just accelerate because the market is exploding. Well, you guys are a hot startup, a great area. No one in the history, no one where you guys have came from. You know, the entrepreneurs that make it always have an itch to scratch, always have the persistence, I got to say, in following your work and the team. It's been fun to watch. You guys have kicked ass and nailed it on the graph side. Congratulations. Thank you. And you're going for it. And you know, same with us. You know, we had the same thing. We felt strong about our media that we built this media business and it's been great. You know, the fifth season with theCUBE, we already had 100,000 views here live on the live stream, so it's been great. Fantastic. Thanks for coming on. Appreciate it. Great to see you. Neo Technologies, Neo4j, great buzz. I mean, just so many tweets. You're winning the crowd, so to speak. Great community. Thanks for coming on. I mean, thanks for coming on. This is theCUBE. We'll be right back with our next guest after the short break, live from Silicon Valley, all the action, all the startups, all the major big data innovation here at Big Data SV.