 Ryan Off, who's a data scientist with Bitly, and for the folks who know SiliconANGLE, know that we've had a Bitly person on theCUBE at Strat in February, Hilary Mason, great guest dynamic, a total data geek. She was fantastic. Brian, I'm assuming you're the same. Welcome to theCUBE. Thanks for having me. Good to have you on, man. So Bitly, obviously, is swimming in data. You guys are a company that powers the shortened URLs for anyone who knows what that is. You guys were default on Twitter for many, many months or I think about a couple years. Now Twitter has their own service, recognizing that that's some serious valuable data. So you got booted off Twitter, right? So you're, but you're still out there. You have distribution on other clients, right? Yeah, we're on, we're still on Tweetdeck. We're still on Twitter. Twitter's currently using t.co to wrap all the links, but Bitly still gets all the information. We're heavily used on Facebook, a couple other social sites. Yeah, and you guys, obviously, not only just do the shortened URLs, which is a great way just to kind of create kind of a new resolver, if you will, for the navigation. You guys also track the analytics. So when you shorten a URL and you share something, you guys are tracking origination and destination, kind of source, destination, traffic, knowledge. So share with us. What's the data science like at Bitly? Give us a peek inside the Bitly machine. The data science at Bitly is an interesting job. We have access to data flowing through all types of networks. So we have data coming from Twitter, from Facebook, direct traffic, email. And through that data, we can get location information, country of, like country with the refers, how information is jumping from one network to the other. All that analytics that I've described is free. If you put the plus at the end of a Bitly URL, you get to see all the free analytics. And we all have enterprise analytics for our enterprise customers who have the custom URL shortener. So they can get a little bit more information about sentiment, reputation monitoring, a little bit more geographical information and stuff like that. You know, I remember just, it seemed like a decade ago, but it was only a couple of years ago, the big rage in real time search. You had Collectika One Riot, which when Collectika went out of business, One Riot was sold to Walmart.com and with Cosmix. You guys were around, you guys were being incubated, I believe at the time. People were building these real time search engines like Google, like thinking that people are going to really actually sit there and watch a screen go really fast. But it's different though. I mean, it's not necessarily the same search experience, but it's about discovery, right? So can you share with us the data science involved around surfacing discovery, whether that's analytics, connecting people, some of the magic that has to happen behind the scenes? Well, in the old traditional way of search engines, you basically wrote mini crawlers, you put them out on the internet and you tried to discover new leaks with the social media, the real time web. The content comes to you. It's basically instead of spiders and crawlers, you're more filtering pipes and it's a fire hose and how well you can handle it. And that's the nice thing about Bitly, especially as we're starting to do a little bit more research. It's not older content, we have the freshest content because it's the stuff people uploaded just minutes ago. So the data science component is how do you take interesting things from a torrent and make sense of it, make sense of it in real time and provide information to people in a timely manner that they can make some decisions based on it. I mean, we're totally geeks on this whole movement of what this all means in terms of the new social web. I mean, Silicon Angle, our motto is where social science meets computer science. So there's a lot of math involved, there's a lot of computer science, there's a lot of kind of sociology in this real time web. The question I want to ask you is, because a lot of folks are new to this, like real time, when does it all mean? Talk about this notion you mentioned, content finds you, right? That's kind of most intelligent content based upon you now connected to the network. You guys have source destination event when people share stuff on Facebook or share something on Twitter, which is now a norm in the internet society. People are sharing with mobile phones, et cetera. You have source destination events. So you guys are tracking these new channels and marketing people and people in general say, oh, there's new channels are out there, social channels. I mean, social channels basically means basically distribution of content in social networks or the web, mobile phones, social networks. Explain to activity streams and this new paradigm. What is it, is it real? How real? And what are some of the factors involved in that? I believe very much it's real. It's real in the fact that just the number of devices people are using to connect to the network is changing and adapting in ways, 10 years ago, we're just not everyday occurrences. Just the amount of people who have internet-based mobile phones and how they use it and how they use social networks and how they gather information. I mean, back 10 years ago, people had the palm that could connect to the network and that was a boutique item that was very expensive. Phone.com, remember those days? Exactly, and now, you know. Crappy browser, typing in URLs. And now it's ubiquitous, nearly. And so that changes how people access information and how they produce it too. So updates are not necessarily coming in huge, long blog posts. They're coming in quick tweets about a plane going into the Hudson and that real-time nature of it and how that creates a critical mass. I mean, we're seeing examples. Obviously, in Egypt, we saw the government turmoil of those all social media-based Occupy Wall Street was really generated around essentially connected people using what's called crowdsourcing techniques. It's basically more that people are just in communicating faster, right? Well, I think at Clay Shurkey or William Gibson, I don't know, somebody much smarter than I, just said technology becomes interesting when it becomes cheap. And so all of a sudden, as access to these networks became ubiquitous and everybody had it, people started using it differently in the same way that when desktop publishing became much less expensive, it became a much more interesting industry. People were able to publish their own small magazines and stuff and it's the same with social networks. You can find, if you have a good voice, you can find an audience and you don't need a huge amount of infrastructure behind you. Well, technology becomes really interesting when it's free. Yeah, yeah. And you guys play a role in that. So when you sit down as a team to figure out what data products, if I can even say that, that you want to build, what's the objective? Is it utility for the user? Is it, does it have to be some kind of monetization pot of gold at the end of the rainbow? Talk about that a little bit. Not necessarily. Right now, just as a scientist and a position that I am in the science team, it's a dream come true. It's a joy because we basically get to take the data that we have and think what interesting things could we sift from it? What could we build? What could we, what models we could build? What observations we could make? And we, and Hillary's an amazing leader and she basically, you know, shepherds us and leads us in that way. And as we build stuff and it becomes interesting, the business guys, the infrastructure guys, they see it and they build on top of it and it just works its way up into a product. What surprised you at Bitly? I'm asking you, you're getting access to all this great data. You're applying techniques and methodologies, you're experimenting, I'm assuming, I'm just speculating but probably accurate using open source tools, community detection, distribution points, trying to analyze these routes and how people share and all this stuff. So it's a slew of variables, right? And has anything come out of the Woodward State? Wow, that is cool. It's... Can you share anecdotally anything there? Anecdotally. So when I was, I first got there, I started to look at just how people use the iPad. Like what time of day do people use different devices? So if, you know, it's earlier in the day, are you more likely to use your cell phone or desktop computer or so forth? And I actually, so I was able to plot all the usage of different devices through Bitly. And I was able to see, you know, what time of day is someone more likely to use their BlackBerry. And we saw this really interesting traffic pattern where people use their iPad much more in the evenings. And, you know, they use the computer during the work day at a time, but you can see how people use different devices differently. And I didn't think I would have that much granularity with the data. That I could see how people use devices differently are what countries the iPad has actually showed up in and so forth. So that was really surprising that I could get that much information about usage patterns. Talk about some of the things you're working on with at Bitly that are priorities for your business in terms of the data science and obviously it's an important area. We heard of, you know, Jeff Hammerbacker's putting a team together. Everyone's trying to hire this new person that's now, has never existed before really in a computer industry kind of sense. So talk about the kind of projects you're working on and then some of the geeky details, you know, Hadoop, obviously the big part of that and some of the elements you're using. Right now I'm working on a couple of products. One is just kind of differentiating the quality of the content we're receiving what refers give us really good quality content. The other is to try to, we're creating a couple of models that we're trying to use to predict how many clicks a link is going to get, how fast it will decay, how much traffic is left to go to it based on patterns of usage. And to that degree I'm also starting to look into different people who have a tendency to access information earlier than other people who are really on the cutting edge. Mike Juer, who was my colleague who also put together the presentation with me, he's been working heavily on trending. Trending topics can bitly kind of come up with a trend and how can we put our own spin on trending that's different from what other people are doing. So that's what we're doing. We just moved from, we just moved to Verisign. We just signed a joint agreement with Verisign so we have a data sharing agreement with Verisign so we're in the process of trying to go through all that dot com data and figure out how it matches up with bitly and maybe using it for some exciting projects. Great, you can see things early too. I mean like from a virality standpoint you can probably see things early. I heard someone chatting about when Alzeer website gets started, gets velocity, you can see some things early. What's his name, that was killed. You guys, I think, didn't you guys pick that up early? Was I reading something about that or? Which, are you talking about the Bin Laden killing? No, the, what's his face, Libya guy. Oh, Gaddafi. Gaddafi, yeah. Yeah, that. Didn't you guys see that early? I think it was, I didn't read that somewhere on a blog post. We did start, that link got shared pretty quickly. That was, you know, my, I started to interrupt. My, you know, I was always talking about Charlie Senate, Global Post. Global Post basically broke that story and I'm sure it was a shortened link and they had, they actually bought the video from a local guy there, it was amazing. So it's real time predicted, well, it's happened but you can see that early, right? Yeah. And that's the news breaking, that kind of thing. Well, that's the thing is, even with trending, you know, there's words that are going to be heavy. There's ubiquitous terms, you know, unfortunately Justin Bieber is always on the social. Lady Gaga. Lady Gaga, unfortunately. The Kardashian plan, but, so the trending is trying to look at when they go outside their norm. Like if you have, you know, 50,000 clicks an hour, if you got 50,000 clicks the next hour, that's not interesting. If you get 1,000 clicks an hour and you suddenly get 50,000 clicks, that's interesting. Something's happened there, something changed. So it's trying to find, trying to strike the balance of telling people what's the latest news. But then also adjusting that to what the norm is. Changes to the margin. You know, Jeff Hemibarca yesterday said that he looked at the skill sets of his team and he pointed to, he had said I had data analysts and research scientists and it sort of put them together. That's how I came up with data scientists. Does that apply in your world? I mean, that's our team. We have, you know, Hillary and myself have CS backgrounds. Mike comes from a heavy mathematics background. Dennis is just an amazing programmer and we recently got a new person, Anna, who is a physicist. So it's bringing a lot of people. You need the philosophy major and you got to get the liberal arts in there. I mean, come on guys, you get the diversity. You got to get the diversity in there, but you know. We take all who can hack. Well, we're really big fans of you guys. We've been following you guys for years. I personally have been following you guys since you got your funding and it's just been great to see you guys emerge bitly as a company. We know you got a ton of great data. We'd love to collaborate with you on stuff. If you ever want to work with Silicon Angle and publish trend data, we'd love to do that with you. We have some things that we've been doing on our own as well as instrumentation around audience is something that's important to us. So again, we think what you're doing is really important work and congratulations for just an amazing run and continue the growth. I appreciate it. Thank you very much. Yeah, thanks for coming on theCUBE. Thanks for having me. I appreciate Brian off at Bitly. Watch these guys, that real value add. I think that's going to be a real future around analytics. This real time piece is critical. The speed of business is accelerated and you guys are part of it. Thanks.