June 14, 2012

Reuben Smith on RIM, building web communities, engineering education in Canada

“All the content...is licensed under Creative Commons. So you don't necessarily feel like you're giving something to some behemoth corporation...The company exists kind of for the community". Reuben Smith on RIM, building web communities, engineering education in Canada


Ravi: The date today is June 14th, 2012. I'm here in San Francisco, California, here today with Reuben Smith. Reuben is a senior engineer at wikiHow. How's it going, Reuben?

Reuben: Hi.

Ravi: How's your day been so far?

Reuben: It's been pretty good. How about you?

Ravi: Good, good, good.

I know Reuben from playing soccer, playing pick-up soccer actually. We've been pretty good friends so far. We just hang out after soccer, go for beers, just hang around, watch games, have fun. Reuben here is from the exotic land of Canada.

Reuben: Quite exotic.

Ravi: Quite exotic, yes, actually. I've been there a few times. So tell us a little about yourself like your early life, your childhood, and what got you interested in computer science.

Reuben: OK. I grew up in a small town. I didn't know a lot of people there that were really into computers, but I was.

Ravi: Now where in Canada did you grow up?

Reuben: It's a small town called Perth, Ontario. I basically grew up without a lot of technology actually. The very first house I lived in, we had an outhouse out back, and no running water. My parents were basically hippies living in the outback. [laughs] Not exactly the outback but...

But after a certain age, I think I was probably three or four, we moved into town. Their business got more successful. We got more stuff including technology. I really connected with computers when I was about 12 or 13.

Ravi: Nice. So you must have had an early DOS machine?

Reuben: Exactly, a 386 that got struck by lightning at one point.

Ravi: Oh, wow. OK. So you're playing around with your DOS machine and pretty much later on, you got interested in computer science because of that? You ended up going to Waterloo, right?

Reuben: Yeah.

Ravi: For your undergrad.

Reuben: That's right. I got interested in BBSs after getting a modem and had friends that I played games with and that kind of thing. I got into programming because I always liked that aspect.

David: Reuben, one thing I always ask is, when was this exactly? When was the year? It was probably the mid-90s I imagine or the early 90s?

Reuben: Yeah, the mid-90s.

David: OK, so you're talking about dial-up bulletin boards and things?

Ravi: I was about to say that, the dial-up Internet days.


David: Exactly, yeah. The Internet came along when I was in high school. I was just speaking to one of our interns at work today who said that he didn't know what FTP was, file transfer protocol. I felt old.

Reuben: Wow. Yeah, well, I do too now. I definitely know what that is. Yeah, at some point in high school, I realized I had to focus to get into a decent school. Waterloo is the best school in Canada for computer science. No offense if there are any Canadian listeners out there who didn't go to Waterloo.

David: I've heard that myself as well.

Reuben: It's an excellent school. There's a lot of alumni around here. In general, it focuses on engineering and computer science and mathematics. I originally actually only got in for math. It was a really hard year. I had excellent grades but ended up only getting an acceptance for kind of a general math degree. But in the first year, I got quickly accepted to computer science because those were my highest grades.

Ravi: I'm kind of curious to the difference between college life in Canada, or at least the undergraduate engineering college life in Canada, versus in the United States. I don't know if you happen to know people here who have done undergrad here and compare and contrast.

Reuben: I definitely know people who have done undergrads here, but I don't know much other than the stereotypes of fraternities and stuff like that. I don't know how close that actually fits with real life. Were you part of a fraternity for instance?

Ravi: I was actually part of a fraternity for about half of a semester.

Reuben: OK.

Ravi: And then my grades just plummeted and my dad told me, "If you want me to keep paying for your college, you need to get out of the fraternity."

Reuben: OK.

Ravi: It was a good time, definitely. I wouldn't say it's as much as what the stereotypes portray it to be.

So there's no concept of fraternities or sororities in Canada, or Canadian universities?

Reuben: Not that I know of. Sometimes you do see Greek letters on stuff, but the fraternity life and sorority life is almost nonexistent. Education is also quite a bit cheaper because it's subsidized by the state. I guess there might be more diversity in some ways, as diverse as Canada is. [laughs] But I'd say it's very merit based. In a lot of ways, the people who do best are the ones who are strongest at academics. There's less of a focus on sports and things like that in college too.

David: Just some basic questions. You enrolled for a B math if I'm correct.

Reuben: That's right.

David: Is that a three year program or four? How long does that take normally?

Reuben: It was an honors BMath, so it's four. I did co-op also, so I was actually expected to take five years, but I ended up dropping out of co-op part way through because I had a good job.

David: OK. Co-op is a work study program?

Reuben: Exactly.

David: What is that?

Reuben: At Waterloo, if you do co-op you can do working one term, studying one term and so on alternating throughout your education.

David: For five years?

Reuben: Yeah.

David: Interesting. Then do you get paid for the time that you work?

Reuben: You do. Actually interns in Canada generally are much better paid than here. My boss looked into, for instance, getting an intern from Waterloo and decided it was too expensive.

David: At your current job?

Reuben: Yeah.

David: I see.

Reuben: I'm not going to give specifics, but I was well-paid when I was an intern at Research In Motion, paid better than people without degrees.

Ravi: Certainly. Coming back to Waterloo, what were some of the courses and projects that really inspired you? You were majoring in BMath, majoring in mathematics, but you were still interested in computer science. Were there courses or projects that really got you thinking, "This is my career, this is what I want to do as a career?"

Reuben: I had wanted to go into computer science before. I had decided in my first year that I definitely wanted to do that. And all the courses are the same for both math and computer science students in the first year.

Ravi: In the first year, OK.

Reuben: Yeah, because computer science is part of the math faculty at Waterloo. So I ended up just getting accepted in computer science and had no differences from a computer science student. I did mathematics too. I ended up with a double degree and doing all of the major courses of computer science and combinatorics and optimization. But that was because I had an interest in cryptography during university.

Ravi: Was that your favorite course, for example, cryptography, you'd say?

Reuben: I did a work term before I went to college at a company called Entrust, originally they were part of Nortel. They were a security company that did PKIs and stuff like that. But all the stuff around certificates and cryptography, security in general, I found very interesting and really challenging too. So I wanted to learn that in more depth.

Ravi: Nice. What year did you graduate from Waterloo?

Reuben: I graduated, I believe, it was 2003 or 2004.

Ravi: 2003, so this is obviously after the bubble burst, I suppose.

Reuben: Yeah.

Ravi: How was the job economy back then? Yeah, what were you thinking of doing after? Were you thinking of going for higher studies like a master's?

Reuben: I hadn't actually considered doing a master's at that point. I was thinking just about the job market. I wanted to get out there and work at a really interesting company, preferably a smaller one. My experience with Research In Motion, I had already decided that a really large company with lots of bureaucracy wasn't my thing.

David: Can you elaborate on that a little bit? What was it about that that didn't sit with you?

Reuben: I was at a really small group at Research In Motion. I started there, actually, when there was only about 160 employees.

David: What kind of products were they making at the time?

Reuben: They were making an email pager. It was literally seven lines of email text; you could check your email on this pager thing. They came out with some new products, did really well, got listed on the stock market right before I joined. Ended up doing quite well and by the time I'd left, there were at least a few thousand employees.

Ravi: How big was RIM when you joined, again?

Reuben: About 160 employees. It was a very small company.

Ravi: Wow.

David: What kind of people would use these pagers?

Reuben: Usually corporate exec types.

David: OK, doctors, maybe, in hospitals or no?

Reuben: No, pagers, in that sense, were usually one line and really just a phone number kind of calling you or a very simple message. Whereas this was a way to check your email wirelessly and not on your junky phone, which was all that really existed back then.

Ravi: So this was Internet email like POP and IMAP or no?

Reuben: Kind of. [laughs] The way they originally set it up, it connected to an Exchange server. So it was really intended for corporate use. Later on, there was something, kind of a layer sitting in between POP and IMAP. Their big thing was that when you got an email, it would just get pushed to the device within seconds. You literally would see it before you'd see it on your desktop, you had an email.

David: This was 2003, so about 10 years ago.

Reuben: I originally started working at RIM in '99, actually. So that was, I worked at another mobile company, I think, 2004, 2005, that was doing something similar for a Treo phone.

Ravi: Continuing on, you say you didn't want to do a master's initially, you were actually looking for a full-time job after the course.

Reuben: Yeah. It wasn't as much about the money, but I enjoyed working in a team. I liked academia to a certain extent, but I decided that I wasn't super into the research side of things or the theoretical side of computer science. I still find this stuff fascinating. But, there are people who are much better at it than I am. So, I've always been a strong coder and usually good at getting programs correct. So, I decided that I wanted to work and use my skills.

Ravi: OK. So, coming back to your work at Research In Motion, what was the work like? I presume it must have been like C, and assembly, or...

Reuben: Yeah, originally it was C. I was doing cryptographic algorithms in C. At that point, we were moving to Java on the device, actually...

Ravi: So, like RSA and AES.

Reuben: Yeah. I remember in job interviews, one of my former bosses asked me about RSA. And I just wrote it down for him on the board. I just did it enough times, so it was really easy for me and able to, you know, prove the correctness of it.

David: The paper's short, it's only about six or eight pages, I think.

Reuben: Yeah. Exactly. It's actually extremely simple mathematics. But, it's...

Ravi: But, implementation-wise, there are sure challenges because you're dealing with very large integers.

Reuben: Yes. Yes. Exactly. That was. I found RSA much more challenging to implement than, for instance, like a block cipher, or a hash like SHA-1, or something. Those are actually quite easy to implement...

David: What was the hardware platform in these pagers?

Reuben: In the original RIM pager, I believe it was a 386 running at 12 MHz. So, they later, with their phone, they moved to a chip by that British company.

Ravi: ARM.

David: ARM.

Reuben: Yes, ARM chip.

David: So, it was an Intel system on chip platform in a pager, you're saying.

Reuben: Yep.

David: It's interesting...

Reuben: It ran on a AA battery.

Ravi: Yeah. Was that common at the time, because I didn't think that x86 CPUs were especially optimized for mobility?

Reuben: No. Not at all. They ran at a lower power rate. But, it worked for their purposes. There wasn't a lot of, you know, big chips being put into anything wireless at that point. So...

Ravi: OK. What kind of battery life did it get?

Reuben: It actually, like depending on your usage, it would last between two weeks and a month. And, it was incredibly sturdy, actually. When my boss was demonstrating to me the pager, at one point, he just threw it against the wall, and the battery popped out. But, the thing was fine.

David: Yeah, I'd like to see that with an iPhone these days.

Reuben: I ended up demonstrating that in the same way to some friends, but thankfully, without problems.

Ravi: Well, we all know that Blackberries are known for their sturdiness. Isn't that one of their...

Reuben: Yeah kind of...

Ravi: ...things that they're known for?

Reuben: They're known for a lot of things. But, I think the company has kind of lost its way, more recently. I can't, you know, speculate on what's happened. But, they're a huge company now.

So, at that point, I saw a little bit of bureaucracy creeping in. I felt like, when I was really young, I could kind of do what I needed. I tended to work pretty long hours. If I needed an afternoon off or something, no problems at all. But then, later on, it was kind of like, "Oh, you should check with HR...sorry, you can't," you know. That's a little frustrating when you're like...

I think big companies can go that way and just try to be completely uniform about their polices, but smarter ones and smarter managers kind of fight against that.

Ravi: But, you mentioned RIM losing its way. But, isn't the major reason because of the iPhone, how destructive it was. I mean, not just RIM, other companies have come to mind, Nokia, or Ericsson. These guys used to be big players when it came to cell phone hardware. And, now look at them.

Reuben: I agree. I think that RIM was quite a bit younger than either of the other northern countries, though, in some ways. I think, if there was more of a culture of innovation there, then they might have been able to survive. We were working on very early, but that was one of the projects I worked on was like, a very early web browser on this device. If they had invested properly, or done things kind of like, a little more forward-thinking, like the iPhone did, then I think they wouldn't be where they are.

David: So, what was the state of the data network back then? Can you talk about that a little bit? Like, what kind of bit rates were common and how did people connect to the Internet on these mobile devices?

Reuben: Sure. The RIM pager was actually originally only using the Mobitex network, which is basically a pager network, at the time, mostly used by first responder services in the US and Canada. We basically were charging customers a monthly fee for the device. It was completely independent of how much data they used. So, there was a really strong incensitive to compress everything really well, and that kind of thing.

They did a lot of things, like, if you, for instance, reply to an email, the response was always just kind of a reference to something else, and that kind of thing. It was really, really, focused around being very spare with the bits. In terms of data speeds, I have a very hard time answering that question because I just know that we get packets that were under 500 bytes. There weren't a lot of them coming in. So, I know that the first part of the email would generally come just kind of as one page. And then, if you wanted more, the latency was horrible, so it would be like three or four seconds before you got the rest of the email.

David: Well, certainly, in terms of user experience, the web was more fragmented then. I mean, there is probably a version of the World Wide Web that was in use in computers. And then, a fairly different version that was in use in mobile devices.

Reuben: Definitely. So, if you wanted to do like TCP, you generally connected to a proxy to do it, and it was to make things faster. For instance, you'd have all sorts of knowledge at the higher levels of the stack, like HTTP and that kind of thing because, TCP can have quite a bit of lag in between packets like the SYN/ACK, thing alone. It just would have been horrible over these networks.

Ravi: So yeah, moving on from RIM. So, you graduated with your bachelor's in 2003. So, tell us what you did after that.

Reuben: That was kind of during the economic meltdown of the tech industry.

Ravi: The dot-com bubble bursting.

Reuben: It wasn't a good time at all for getting a job, but I knew a friend in Ottawa, which I'd relocated to, who worked at this processor company. They were working on a Java-based processor.

Ravi: What's the name of the company?

Reuben: It was called Zucotto.

Ravi: Zucotto...

Reuben: They essentially were fabricating silicon, and I was working on the software side, to work on the operating system.

David: Did they do their own fabrication of silicon?

Reuben: They gave it to someone else to do, but they designed the whole chip themselves.

Ravi: Something similar to what AMD does, these days.

David: Yeah.

Ravi: How long were you at Zucotto?

Reuben: Not very long, because they ended up going bankrupt. I think I was there for six or seven months. I think I was the last employee hired. [laughs] It was kind of a hard time for the company, but I was really enjoying the work. I thought it was really fascinating stuff, and because they had laid off a bunch of employees, and that kind of thing, I got to work on these awesome projects that an intern should never be working on. I found bugs in microcode, and that kind of thing.

It was pretty cool stuff. I figured out all sorts of things about operating systems that were really different from the theoretical side I'd learned at school.

Ravi: Like what? Could you elaborate?

Reuben: Dealing with hardware. I got a lot better with twiddling bits in C and that kind of thing, at that point. The scheduling interaction—I'd learned that a little bit in the operating systems course—but realized how important it was, and how difficult it is, when I got there, too.

Ravi: This was 2001, or 2000?

Reuben: 2003, when I got out of school. It was an interesting time. Like I say, I wasn't there a really, really long time, but I got to work on a lot of interesting projects.

Ravi: Because the company went bankrupt.

Reuben: Yeah.

Ravi: You would have just wanted to stay there, if...?

Reuben: Yes, definitely. As it was, they went bankrupt, and a couple weeks later, I got a job down here, through a friend at a company called Seven Networks. They were basically doing email for the Treo and Palm Pilot phones, and some other phones, too.

I ended up working in the Palm Pilot group, and wrote some software. Basically helped them ship a product that summer. But, at that point, about a month or so after I joined, I found out that I got into a master's program that I was applying to at the same time.

Ravi: But you weren't interested in one.

Reuben: I wasn't, originally, when I left school, but with the job market being the way it was...

Ravi: [laughs]

Reuben: I guess I did miss academia a little bit, too. There were more things that I'd wanted to study. Things at Waterloo are pretty fast-paced. Anyway, I did that. Helped them ship a product, but nothing really spectacular there.

David: Did you have any trouble getting into the United States, around visa issues? Any of those kinds of problems?

Reuben: No. I got in with a TN visa.

David: I see. Did you intend to stay, at the time?

Reuben: No. No, it's a temporary visa. It lasts for one year. I knew that if I hadn't gotten into my master's program at that point, then I probably would stay for that whole time, but I knew other friends that had renewed their visa multiple times and stuff.

David: Is TN visa something specific for Canadians?

Reuben: Canadians and Mexicans.

Ravi: I see.

Reuben: Mexicans have to go through a slightly different process than Canadians, because Canadians don't require a visa before they get to the border, but I don't think there's any limits on the number. That's why there are probably so many Canadians in the Bay Area.

David: But it's a non-immigrant track, or non-citizenship track visa.

Reuben: Exactly.

Ravi: Microsoft uses a lot of TN visas to bring in Canadians to the United States as well.

Reuben: Oh, I did not know that.

Ravi: Yeah.

Reuben: But now that they have opened up shop in Vancouver, I presume that they won't have to do so much...

David: It makes it a bit easier, yeah. The trick, though, is that people that come in on TNs, it's not a citizenship track visa, so the expectation is that they either have to renew it or leave the country after a fixed period of time.

Reuben: Yeah, it kind of works that way, but there's been people who have been here for 20 years, because you can perpetually renew them, also.

Ravi: Did you have any family, when you came down?

Reuben: No.

Ravi: It must have been tough. I mean, you must have not known many people.

Reuben: It's true. I had one friend down here, and he had a circle of friends. I ended up hanging out with them and getting into the San Francisco tech culture. Yeah, it's a fun city to live in.

Ravi: How did you find that it was different, versus similar, to Canada and some of the other places that you've worked?

Reuben: I think that, in Canada, the work culture is more conservative in some ways. Not in the sense of people wearing suits to work, but rather people are just a little bit more, maybe, cautious about new ideas, whereas the culture here is like "fail quickly," you know? [laughs]

Ravi: Yeah, one thing I've heard, too, is that equity-based compensation is a lot more common in California. Is that true? Less companies give stock-based compensation in Canada, overall?

Reuben: Yeah, I think that is true. It's probably more common now, but in general, cash salaries are higher, too. They're probably higher in Silicon Valley than anywhere else in the US, though, right?

David: Considering how good Waterloo University is for computer science and math, there must be a burgeoning tech industry up there, in Canada, right? In Toronto, presumably?

Reuben: Yeah. I think there's a lot of strong tech companies. A lot of companies doing consulting, gaming companies—like Rockstar Games is out of there—a number of bigger enterprise software type places, too.

David: I've heard some of the big hubs are in Vancouver, especially for games, I've heard?

Reuben: Definitely. EA has a huge office there.

David: Yeah, Vancouver, and then also Montreal, I think EA has a big office, too, right?

Reuben: Yup, and there's a French gaming company. I can't remember their name now. They made Assassin's Creed, and they're based in Montreal.

David: Is it Ubisoft, maybe?

Reuben: Yeah, Ubisoft.

David: Yeah. My officemate, when I worked at Microsoft, was very into these games and he was from Montreal so he used to talk about this a lot.

Reuben: Oh, nice. I kind of got out of gaming, by the time I got to university, but it's still really relevant to the sector, because most people within technology seem to play some game or something.

David: I think, also, it lends a distinctive type of competence to people. People that are very good at high-performance graphics and C programming, those sorts of things...

Reuben: Very true.

David: Skills can be harder to find, in some places.

Reuben: Definitely.

Ravi: I think what Reuben's referring to is the fact that most of the people who are in tech, they are gamers, rather than game programmers.

David: I see.

Ravi: The tech skillset lends itself to the gaming culture. Like, for example, I play Starcraft, and I know many people in the tech industry play StarCraft. I think the whole concept of being orderly, and having a certain way about things, and multitasking...

Reuben: And figuring out the strategy for something, and doing better at it.

Ravi: Exactly, and the creativity itself, that lends to many things that you find, itself, in programming, I think.

Reuben: I agree. That's going to help a lot.

Ravi: Coming back to Ottawa, you did your master's there.

Reuben: Yup.

Ravi: What was your thesis?

Reuben: I was accepted at a couple different schools, one in Montreal—the school McGill—and it's a great school, but I had a relationship in Ottawa at the time, and I wanted to stay there. I also got paid, like a TA-ship, so that helped make a decision.

I ended up, originally, choosing to study natural language processing, but then fairly quickly, after about four months, I moved into machine learning.

David: This is when, 2006, maybe?

Reuben: I think 2005.

David: OK, so this was seven years after Google had gotten started. They were a big power in machine learning. What else was going on in 2005?

Ravi: Amazon's big in machine learning. They were pretty big back then.

David: They came out with a recommendation engine, and that kind of thing.

Ravi: Netflix, with recommendation engine. But Netflix wasn't big, back in the...

Reuben: Yeah. The original Google algorithm wasn't really based around machine learning. It was more network analysis, almost. They, essentially were optimizing this giant graph of websites, based on the links. Later on, they ended up using a lot of machine learning stuff to figure out keywords of sites linking to other sites, and that kind of thing, and natural language processing, too.

Ravi: Is that what your thesis was about? What was it, specifically?

Reuben: My thesis was about taking the output of network devices—security failures of network devices—and categorizing them in a way that's more usable to network admin.

David: What kinds of network devices? Routers or hosts, or what?

Reuben: Firewalls, intrusion detection systems, that kind of thing; more security-oriented stuff.

David: I see, so you were trying to classify the log files from these devices.

Reuben: Yes. We'd do it in an untrained way, which just meant clustering similar alerts, so the output of that, essentially, you'd have these log devices. You'd see each individual log entry, but you wouldn't get a sense, necessarily, of the attacks that were happening against that network. For instance, say you had a port scan or something like that. A lot of systems that existed before this either used heuristics, like base it on the IP address that it's coming from, or what the packets look like that are being sent, or various other heuristics.

Some of the more difficult attacks were ones that were very distributed, and cloaked, or time-delayed, that kind of thing. Ottawa is the town where there's federal government, so there's a lot of defense there, so there were people willing to pay for this kind of research.

Ravi: I think Nikita Borisov, one of our professors from Illinois...

Reuben: I know exactly who that is.

David: He's actually my graduate academic advisor.

Reuben: That's really funny, he's from Waterloo.

David: Yeah.

Ravi: I was actually referring to him, because I know he's from Waterloo and he did a lot of security research. Obviously, he got his undergrad there. I don't know where he did his Ph.D.

Reuben: I think at Berkeley.

David: Berkeley, right.

Ravi: I wonder if that's a hotbed of security research, the Waterloo area. Waterloo University.

Reuben: Yeah, I think so. The only reason I know about him, actually, is a friend of mine in high school, Benjamin LaHaise, who was a Linux kernel hacker—very, very smart, like wrote some of the original VM stuff, I think, in Linux...sorry, I'm not 100 percent sure about that, actually, but I think he wrote some of the original VM stuff in Linux. He knew Nikita from high school or something, and was just constantly talking about how smart he was. That's why I knew who this guy was.

Ravi: I didn't know Nikita went to high school in Canada as well.

Reuben: I believe so. Small world.

Ravi: Moving on, so you did your master's at the University of Ottawa. What next?

Reuben: So after I finished my master's, basically I wanted to find another job but I wasn't really sure what I wanted to do. I kind of decided at that point that I wasn't as interested in super low level stuff. As much as I enjoyed it, I felt like I...

Ravi: C didn't excite you anymore?

Reuben: Yeah. I felt like I was really good at it, but it was less exciting. I felt like I'd done it and it was a bit of a slog. [laughs]

David: I think you just need a switch, too, after a while.

Reuben: Yeah, I agree actually. So that's what I did. Basically there were some people doing web contracts, really small companies who'd have like a single idea. They'd want to publish a website, and I knew how to do that kind of thing. Sometimes it was existing code bases that were written by nonprofessionals. [laughs] So I'd go in there and fix whatever needed to be fixed. I got a good reputation for myself because I was able to scale sites and just figure out problems quickly, that kind of thing.

David: This was FlipTrack or a consulting company?

Reuben: Yeah, I was starting to work at FlipTrack at that point, but there was predecessor company called I Play Music that created software. I just helped put up their website and that kind of thing.

Ravi: Right, I remember seeing on your LinkedIn profile that you had done a bit of contracting for a variety of companies.

Reuben: Yeah. That was one of them. There were a few other companies that I worked for during that time, just random projects, that kind of thing.

David: How long was this?

Reuben: This was for maybe about a year.

Ravi: After your master's?

Reuben: It was always just contracting. I ended up moving to Montreal at that time because the city was so interesting and there was so much stuff going on. I was working for American companies, so I could work from anywhere.

Ravi: The whole fractional contracting thing is interesting too. This was in like 2008 by this time? Or when were you doing this?

Reuben: I want to say it was 2006, 2007.

Ravi: Yeah. The reason I mention it is I've also done this a bit where I was more or less on the payroll of three companies at the same time working anywhere from 10-20 hours a week for each one of them.

Reuben: Yeah.

Ravi: It was win-win for both parties I think because they got somebody who was fairly current with new technology, but they didn't a full time employee necessarily.

Reuben: Definitely.

Ravi: You got to work on a variety of different things and meet different people all at the same time. It can really build your career, I think. You decided to get out of it ultimately, though. Did you just find that there was a particular company that you really wanted to be with for the long term? Or what happened?

Reuben: Yeah, one of the companies I was working with was starting a new thing, venture funded. They made me a really good offer. It was this amazing opportunity to kind, like, of write the code base from scratch and really work on something new and scale it up. With our initial product I think we ended up getting maybe eight million a month or something like that.

David: Eight million people viewing the site?

Reuben: Mm-hmm. It was mostly slide shows and stuff like that. We were creating a product that did slide shows through Flash and stuff. It was an interesting product to scale. I was mostly working on the backend.

Ravi: It was a content property you're saying?

Reuben: Yeah, but we were getting users to create content.

David: OK. This was FlipTrack or this was the predecessor to FlipTrack?

Reuben: Yeah, FlipTrack. They changed their name to MobLying and started doing social games.

Ravi: Yeah, Ravi and I saw that; we were looking at this last night. I think they started at FlipTrack. They were founded I believe here or in Santa Monica, I think in the Bay Area.

David: So the purpose was user generated content and it's editable? I don't get their...

Reuben: How they make money? [laughs]

Ravi: Yeah.

Reuben: No, it was basically an editor for slide shows. If you had just some photos that you wanted to edit you could go into this web app, written in Flash, pull all your photos from a Flickr account or Facebook or whatever.

Ravi: And you created a personalized slide show based off of that?

Reuben: Yeah. There would be music playing in the background, and the photos were synced to the music. It was kind of cool. You could definitely do some really neat things with it. Then you could take that and embed that on your blog. This was kind of like in the MySpace days where people were really flashy about the content they did.

Ravi: Exploring their artistic sides.

Reuben: Yeah.

David: Not flashy in the sense of Shockwave Flash, but "flashy" like showy.

Reuben: Yeah, showy. We're probably dating ourselves here too.

David: 2007 maybe?

Reuben: Yeah. MySpace was kind of known for this garish look. You know?

Ravi: I'm sure it still is.

David: Yeah, but it was what was in vogue at the time.

Reuben: It was, definitely.

Ravi: Again, I keep coming back to dates. It's interesting to talk about the timeline of this stuff. This was probably 2006 or 2007 maybe?

Reuben: Yeah, I think that's about right, 2007.

Ravi: OK.

Reuben: So then around 2008, 2009 the company pivoted, as many Silicon Valley start-up companies do, for a variety of reasons. The reason I eluded to earlier, fail quickly. When you realize that you don't have a solid business, find one, especially if you have a bunch of money sitting in the bank. So they moved into social gaming. Virtual goods was just kind of like a theory being explored at that point and a lot of other companies were starting up doing that.

Ravi: Virtual goods like within a video game? So like tokens you can purchase and they apply in a game.

David: Reuben, what was the genealogy of this company? Who were the founders and where were they from? Do you know?

Reuben: Yeah. The main founder was Stuart Putney. There was one other guy who quickly left the company and got bought out. His equity got bought out. Stuart worked at Knight Ridder, a publishing company, and I believe there were one or two other important companies that he worked for too.

Ravi: I have heard of Knight Ridder. I just don't remember the context, but I've definitely heard of them.

Reuben: They have a big office in San Jose. I'm not sure whether they're based there, but they own newspapers and stuff.

David: They're a publishing service, or a newspaper.

Reuben: Yeah.

David: When I was in policy debate in high school there were a lot of Knight Ridder citations that we used to have.

Reuben: Oh really?

David: Yeah. Do you know who the investors were off the top of your head?

Reuben: I should be able to remember because there were two, but I don't.

David: OK. We looked this up on the Internet. I think it might have been Mohr Davidow.

Reuben: There we go, Mohr Davidow.

Ravi: I think it was like Motorola too, Motorola Mobility Ventures or something like that.

Reuben: They did invest some also, but Mohr Davidow was the primary investor.

Ravi: I think I was telling you this earlier or sometime back. When I moved to the Bay Area, the two terms I heard for the first time in my life and I keep hearing now because I'm in the midst of start-up culture but I don't work in start-up my self is a) the concept of pivoting. That's something we talked about earlier. Then b) the concept of fail fast. Right? Like if your idea is not working out don't try to force it into success. Go to the next venture or pivot as you would say.

Ravi: I think I was telling you this earlier or sometime back. When I moved to the Bay Area, the two terms I heard for the first time in my life and I keep hearing now because I'm in the midst of start-up culture but I don't work in start-up my self is A) the concept of pivoting. That's something we talked about earlier. Then B) the concept of fail fast. Right? Like if your idea is not working out don't try to force it into success. Go to the next venture or pivot as you would say.

Reuben: Yeah. I think there's a lot of really interesting ideas that I've only found here in Silicon Valley too. They do tend to spread out from here.

Ravi: Like concepts you mean?

Reuben: Yeah, just ideas around work. Work here is so casual but at the same time very merit based. It's not about what you've done but rather what you're doing. There are so many other really valuable things that people do here.

Ravi: Yeah, like what in your opinion?

Reuben: I'm trying to think of some of the other kind of ideas. Each company seems to have a different ethos. Google, for instance, is constantly trying out new products and doing different things. But at the same time, internally, they are known to be a little bit more rigorous with their testing. If anything is going to affect the core products you have to get senior VP to check on it, that kind of thing.

Ravi: They have a really good infrastructure for just testing and make sure that when the products ship out they're decently tested.

David: One thing that stands out in my mind is the emphasis on experimentation.

Reuben: True.

David: There's a lot of a priori knowledge you might thing you have about layouts, typefaces, conversions, languages, colors, everything. I think the emphasis here is always on "show me the numbers" not "I'm the highest paid person in the room so you should listen to me."

Reuben: Yeah, yeah. Google is famous for being data driven, definitely, almost to a fault. I think Steve Jobs and Apple were kind of at the other end of the spectrum in some ways. They had the more artist mentality.

Ravi: They understand the user.

Reuben: Yeah. They wanted to create something. Steve Jobs was saying, "How can users know a better product that they've never seen before?"

David: The classic quote around this being that had Steve Jobs done a focus group in the time of Henry Ford, we would have gotten faster buggies.

Reuben: Exactly.

David: Right? Not cars.

Reuben: Yep. So I think actually like a focus on aesthetics is really important around here. People really do care about what things look like.

David: Especially in San Francisco, in the city. Right?

Reuben: I agree, yeah. There are a lot of people around who probably spend too long looking at themselves in the mirror in the morning. [laughs]

David: [laughs]

Reuben: I think there are a lot of incredibly talented designers, actually, too. I feel like the bar for design here is just at a different level than anywhere else I've experienced.

Ravi: Is that across all the different design disciplines like architecture and graphic design and print publishing?

Reuben: I've really only experienced it with web. I think that probably it's true. I don't really know about architecture, but with print design and stuff I think San Francisco is very innovative too. But with web, you know, they are always a couple years ahead.

Ravi: When you're talking about clean interfaces, Google comes to mind. Right? I mean, when you look at search engines back in the day when I was in high school and I'm trying to figure out which search engine I want to use or which email account I want to register myself initially. They had Dogpile, Alta Vista, Lycos, Excite, Yahoo. Google seemed the cleanest of them all. You get the cleanest interface. For me as a user, I was in high school, I didn't care that much about page ranking algorithm.

Reuben: No.

Ravi: I didn't give a shit about that. I just cared about my results looking clean and clickable.

Reuben: I remember using Yahoo search engine before that, before they moved to Google. It wasn't based on the importance of the page, and that's really, I think, what Google brought. If you search for a company name like Xerox, you would get Xerox as the first hit on Google whereas no other search engine could post that. You think it's so basic now, but back then, it was a revolution.

David: I think the other thing that Google has done that a lot of people don't give them credit for, they have an incredible computation platform as well. They probably have more compute power, I would imagine, than probably any entity in the world.

Reuben: Yeah.

Ravi: I wouldn't doubt that, for sure. So you were at FlipTrack for some time, 2007. You were there for a couple years if I'm not mistaken, and then it pivoted. Was that reason you decided to move out of it and try for a new venture? What happened then?

Reuben: I think it was mostly a positive change. I was looking for something else. I felt like my skills were stagnating there. I'd been there almost three years.

David: What kind of work were you doing? Just so we know.

Reuben: I was doing web back end stuff in PHP. I was also running the servers and stuff like that.

David: So was this a traditional LAMP stack?

Reuben: Yeah.

David: So Linux, Apache, MySQL, PHP?

Reuben: Yeah.

David: OK. And where was the site hosted? Did you have your own servers?

Reuben: We had them at Peer 1 which is a hosting company.

David: OK. Was that co-location or...

Reuben: No, we didn't have our own servers. They were managed servers.

David: I see. So you paid another company to manage the servers for you?

Reuben: Yeah.

David: OK.

Reuben: But I was writing a lot of the backend code that dealt with the database and just made sure that things were efficient. I did work on featured products and stuff like that. But, for instance, when we started developing games, we needed a server side to store all the data and do all sorts of...replicate logic so that people couldn't cheat and stuff like that. So we'd do stuff in JavaScript on the front end but then a whole bunch of back end stuff too. We started using Memcache, for instance, to keep hot objects quickly between web sessions and stuff.

Ravi: I guess after this, you decided to move on. You were enjoying the work you were doing but at the same time you were looking for new ventures, if I'm not mistaken.

Reuben: Yeah, I had a friend who was working at the company, WikiHow. He was the primary engineer there. There were basically two engineers, and they knew that one was going to start looking for work fairly soon so they wanted another one. I definitely wasn't sold on working there at first, but after talking to them, I really liked the CEO's philosophy. I thought the COO was really, really smart and talented too. I decided to get into that company basically because of the management.

Ravi: I guess now you could describe for the purpose of the audience, what wikiHow does, how it's different from Wikipedia and other such content media sites. What was the philosophy that you really liked?

Reuben: WikiHow is based on MediaWiki which is a software platform that Wikipedia uses. Each article on WikiHow is a how to article, and it's written in a very similar way to Wikipedia. It's collaboratively written. We have a community of people that use tools and generally just participate in the community there. It's a very kind of work-oriented ethic there. People really like doing small tasks to get things done. Some people make major edits to articles, but it's, in general, just a really interesting site and community.

David: What motivation is there for people to contribute to this site? Do you really know?

Reuben: All the content, like yours, is licensed under Creative Commons. So you don't necessarily feel like you're giving something to some behemoth corporation. People definitely don't feel like they're being exploited. The company exists kind of for the community, but they make money off advertising that helps sustain the company.

David: Reuben, on that point for a minute, when Ravi and I were looking at the Wikipedia page of the company the other night there was a quote that I noticed that I wanted to ask you about. It said that the founders want to create this site definitely as a community resource. From what I understand, there's some option to turn advertising off because they really want to create like a really good user experience.

What is the relationship of the website with advertising? I think they view it as an intrusion into the user experience more than anything.

Reuben: That's true, but like so many sites where people get a lot of value out of the site. Like Facebook you're contributing your data to this site, and ads are probably much more intrusive there than on wikiHow. On wikiHow, the vast majority of traffic comes from search, just like Wikipedia. People search for something, how to boil an egg, and this is the site they go to. They go there. They read.

But just like ads all over the Internet, sometimes they go to the ads. People are quite used to it. Some people like it. You find relevant ads to what you're doing. That's basically the relationship.

David: Yeah, the reason I ask is I think if advertising is targeted very well, it ceases to be an intrusion and it can actually be quite helpful.

Reuben: I agree.

David: That's good. I'm not sure how much you can talk about this, but who are their ad partners? Do they have an ad network that they operate themselves?

Reuben: We use Google Ads.

David: OK, sure.

Ravi: You've been with wikiHow for about three or four years now?

Reuben: Three years, close to four, three and a half.

Ravi: What is the development process there at wikiHow? And, to begin with, how many engineers are there and what's the structure like?

Reuben: We're completely open source. We don't take community contributions because we're a very small engineering team. We feel like we don't have the time and resources to review community contributions. We'd love to at some point, actually, but we're not there yet. We have five engineers working on the product right now.

David: That's impressive for the level of traffic you guys have.

Reuben: Yeah. The how-to space is actually like not a bad space to be in. Some of the bigger competitors are like Demand Media, which is a public listed company. They run eHow. Answers.com is another public company. About.com owned by New York Times, there's a lot of kind of bigger content players in that space.

Ravi: I've heard of Howcast as well. I remember talking to people from...I don't know if you're familiar with them.

Reuben: Yeah, video, right?

Ravi: Yeah. They have how to sites but their websites they have in video format.

Reuben: OK.

Ravi: I don't know if they're a competitor to you guys.

Reuben: Certainly. YouTube is. People go there for how-to queries. I don't know if you guys follow this, but Google is good at sending people to their own properties. They didn't used to be as aggressive, but with Facebook and other walled gardens kind of becoming more common they've kind of adopted a bit of that mentality and want to keep people on their own properties.

David: I haven't noticed that too much. Of course an example of that is like Google Plus results being put in a search result.

Ravi: For example, they decide to have Google drive, they don't want people going to a property they can't control like Drop Box or Sky Drive or whatever else people tend to use.

Reuben: Other examples are just the creation of images and image search. You're able to go there, find your image, select it while still staying on the Google site. You'll just kind of see the site you're getting the image from is faded underneath.

David: I have noticed that on Google images. The image is front and center. But I guess they don't want to send you off the page.

Reuben: It's a better user experience, you know?

David: It is.

Ravi: And Google actually has? Had a competitor to Wikipedia, called Knol.

Reuben: Yeah. Other really simple examples are like movie times or weather where the stuff shows up directly in the result. You could argue that it's a better experience, the same with local too.

Ravi: Flight prices too.

Reuben: But then there's the other argument that they're just taking business away from sites like Yelp or the weather network.

Ravi: Coming back to Wikihow, you just mentioned you guys are five engineers and five developers right? What's the development process like? Do you guys use Agile or some other methodology?

Reuben: No. We do check-ins every week where people do that, but we're a very small, kind of productivity-oriented team. Our company culture has kind of been created around...

Ravi: When you say check-in, do you mean like check-in to production? Release/production?

Reuben: No, I'm sorry. When I said check-ins I meant like verbally sitting around a table, telling what you're doing.

Ravi: Oh, I see.

Reuben: The actual development we tend to actually check in and push code ourselves multiple times a day constantly. We have a whole community that is right there to jump on us if we make any big mistakes, and sometimes we do. [laughs] But this is OK. It's, in my opinion, a great way to develop software.

David: Do you have product managers at the company?

Reuben: We have one kind of central product manager, and she's also the COO. She is awesome. She's really, really good at telling when someone can kind of do something themselves. She's excellent at just giving as much responsibility as someone needs or wants.

David: I see. How do you prioritize features? Is it by user request or like...I imagine driving traffic to the site is pretty important. Right?

Reuben: Yeah, definitely. There are a bunch of things. There is strategic stuff that we want to do too. An example of a feature that I worked on was there's this thing on the site called the community dashboard where it shows almost real time what various things people have been working on, whether they've been patrolling recent changes or whatever. This product was just kind of like an idea at first. Oh, we can do this. Just kind of gather all the stats and present them to the users. Now it's something that people really enjoy watching, and it's just kind of a general console for the site to find out what's going on.

In terms of choosing which features we want developed, there is definitely a huge community component. What the community wants and bug fixing and that kind of thing is important, but we're always thinking about new things.

For instance, new ways to get new readers on mobile, that was one of the first things I did when I came to the site, create a mobile site. I was shocked that they didn't have one. [laughs]

Ravi: Like a mobile-friendly site, not an app?

Reuben: Yeah, like an iPhone friendly site. I was just working at a company that did mobile web stuff, so it was pretty second nature to me and really frontend performance-focused too, which is really important in that industry.

David: Looking at your resume I saw there's a lot of optimization work there.

Reuben: Definitely, yeah.

David: You talked a lot about Varnish, that the cache?

Reuben: It was a big deal for the site because we replaced Squid which was an old reverse proxy. It was a difficult project, actually, to implement Varnish. There were a lot of quirks about the configuration. It's almost a fully formed language, but it's been excellent to us.

David: The site's a PHP site, right?

Reuben: Yes.

Ravi: Do you guys still use the LAMP stack?

Reuben: No, we actually use Nginx and PHP-FPM as the application stack and then MySQL on the back end.

David: I see. What are some of the key challenges? Scale is huge I'm sure.

Reuben: Yeah.

David: Asset delivery is big too.

Reuben: Yeah.

David: Do you use any kind of CDN or no?

Reuben: We do.

David: Oh, you do?

Reuben: CDN was already in use, but another one of the early things I did there was really tighten up the front end performance. That meant moving a lot more assets to the CDN and just kind of...a CDN is a great way to both scale stuff. It's very expensive to lease these servers. Whereas CDNs are already at scale so they're excellent. Also they're a lot faster because the servers are quite a bit closer to the user.

David: I was just discussing this at lunch today at Crittercism, where we currently don't use a CDN but we have a lot of EC2 hardware that's leased. How do you think about the economics of using CDN? What are the trade-offs?

Reuben: I think that generally CDNs are going to save you money if you're a small company. I think of serving bits, whether it's images or videos, I think is a really solved problem. It's only when you're at scale serving huge amounts of it that it makes sense to actually manage that stuff yourself.

David: I see. Is the advantage there for a CDN, is it more about performance or cost or a bit of both?

Reuben: It's both. Like whenever you're considering an engineering problem such as running the equivalent of your own CDN, you have to, I think, consider the amount of time you're going to be spending maintaining it, writing code, fixing problems, whatever. Your time is usually a lot more valuable, and personally I'd prefer spending my time writing features and doing that kind of thing and dealing with other scale issues. Application scale issues I find much more interesting than scaling bits and adding new machines.

Ravi: For the purpose of the listeners, CDNs are "Content Distribution Networks," right? Something like, for example...

Reuben: Content delivery networks.

Ravi: Content delivery? Sorry. Something like Akamai, for example, a company that does this. Essentially what they do, if I'm not mistaken, is they take your assets, like your pictures and your static web pages, and they store them locally on servers worldwide, so that, for example, if a user from Russia decides to go to WikiHow, when he tries to access the page or the images he might get the results from a server potentially in Russia. Is that...?

Reuben: It's perfect that you said "vikihow", because that's exactly how they would pronounce it, too. [laughter]

David: Nice. So what are some of the other challenges? I mean, we talked about...asset delivery is big. What's their topology? Do they own their own machines at wikiHow?

Reuben: We don't own our own machines. We have managed servers that we basically lease from Rackspace. We have yearly contracts that we sign with them. And every year, too, we refreshed the latest thing.

David: Yes. GitHub does it that same way, too. They use a lot of leased infrastructure from Rackspace. I know Rackspace is considered -- it's a premier alternative, and it's really good for people who have fairly demanding hardware requirements.

Reuben: Yeah.

David: From what I understand. I mean, the other option would be like using a Platform as a Service provider like Google Apps or Google App Engine.

Reuben: Yeah, or Amazon EC...or, well, AWS is a huge one, too.

David: Yeah, that's more infrastructure, though, right? I'm thinking, you know, platform competitors more like Heroku or Azure -- stuff like that.

Reuben: Yeah, there's companies built on top of Amazon, like one that I used at a previous company was a company called RightScale, which is essentially -- it's, in my opinion, very similar to having a server set up like at Rackspace, where you get into the box and whatever -- you can control these, bring them up, but RightScale was essentially just a way of bringing up new servers. I think Amazon has a competing product, but as the web load on one or your cluster goes up, it can just start new servers.

David: Yeah.

Ravi: So why use one over the other? Like...why use Rackspace over AWS and vice versa.

Reuben: Rackspace has a really awesome fast internal network, so you don't have to worry about latency in the network, like you do with Amazon. These app infrastructures, like Heroku and stuff like that, I think when you get down to the optimization layer, they're connecting to the database and stuff like that isn't...it's not necessarily the fastest, and you really have to kind of write, in my opinion, write your app from the ground up there rather than port something over there. So, for instance, we're based on MediaWiki, which is a lab stock, and it would probably be difficult to make that perform really well somewhere else.

David: Does MediaWiki write to the files of some...?

Reuben: It does, like images and stuff like that. It manipulates stuff over NFS.

David: That's interesting. One of the major design tenets of Heroku is very stateless web server.

Reuben: That's awesome, yeah.

David: So if you touch any kind of bits, the expectation is that you put all your blobs on S3 or a blob store. You know, all your data is going to go in some kind of relational or NoSQL store. It's a pretty tight set of constraints. I think it makes it difficult to run certain classes of applications. WordPress is another good example of this. I know that there have been people that have tried to run WordPress on Heroku, and they've had a tough time.

Reuben: Yeah, yeah, it makes sense to me. I think that MediaWiki -- they're fairly stateless, in the sense that the web servers are stateless. Like we get requests coming in. We have two front end caches that run Varnish. We have six application servers sitting behind them. Any given requests, like of anything, whether it's an Ajax call or whatever can go through any path to get to the database or the image store.

David: I see...Is then is all on the West Coast?

Reuben: No, actually it's on the East Coast. I believe the data center's in Virginia.

Ravi: Coming back to challenges -- one thing I've always wondered...when I was reading about Wikipedia, or trying to figure out why people would want to just contribute -- they contribute pages and pages of information -- but I guess it's more the sense of belonging to a community.

Reuben: I think our community is probably a lot more like that. Wikipedia, a lot of it's the "I'm the smartest guy in the room," too. Apparently, it's kind of a caustic community, sometimes. That's what I've heard, but there are a lot of incredible people on there, too.

David: When Ravi and I were preparing for the interview last night, we were sitting talking about the role of community with WikiHow. You've used that word a lot, "community."

Reuben: It's a huge part of it.

David: You guys think about this a lot, obviously. Do you have community managers?

Reuben: We have one community manager -- I believe that's her title -- but it's more just, she's...when things go crazy, she helps sort it out. She helps shepherd the new people along, and stuff like that. That's her main job, and I think that's what she spends a lot of time at, too. You think about sites like Yelp. That's another example of a very community-driven site. I feel like they've defined their culture just by being kind of fun. When you want to contribute to their profiles, they're very jokey, and the language is very informal.

David: They're kind of mean, sometimes, Yelpers.

Ravi: Really?

Reuben: Sure.

David: Oh, yeah. It's a big part of the culture. Being really hard on restaurants and everything.

Reuben: It's funny, though, because if you want to be an elite Yelper...

Ravi: "Elite Yelper."

Reuben: Yeah, well that's a big deal, right? For people that spend a lot of time on there. If you want to be that, it's apparently very frowned upon to give certain businesses bad reviews. [laughs] This is what I've heard.

Ravi: Wait. It's frowned upon to give certain businesses bad reviews?

Reuben: Yes. They have ways...Yelp, I think they've kind of drafted off Google, in a sense. Like, they're trusted to not modify search results, and to do what's right for the user, but Yelp is an advertising-based business, just like Google. I think Yelp, even more so, really blatantly...businesses that pay the $30,000 to advertise with them, they help them bury bad reviews.

David: Yeah. I think the term "editorial independence," or the "Chinese wall" comes up in journalism a lot. When you have newspapers that have a lot of advertising dollars flowing in, but they don't want editorials or other columns to be written about their advertisers.

Ravi: I'm actually not aware of that. So Yelp provides businesses a fee that they can pay, and then they say, "We'll edit your reviews so that all the bad apples are gone"?

David: No, I don't think it's that.

Reuben: It's not that blatant. It's not that blatant, no.

David: It's just, there's an inherent conflict of interest, when you have a site that purports to be providing objective advice about different establishments, but you're also in the business of advertising those establishments and taking money from them.

Ravi: I see.

Reuben: There's a lot of ways that they -- as you know -- a Yelp rating is very, very key to where you are sorted and searched, and stuff like that. Examples of how they might "help out" businesses are...if you decide to throw a Yelp Elite event there, it's kind of expected that everyone's going to give a four or five-star rating to that place, when they go home.

They also close the accounts of people who haven't been active for some amount of time, and usually those are the people who go on, complain about one business because something really crappy happened, or they imagine that something really crappy happened, and then leave.

There's a bunch of different ways, but they have all the reasons for doing that stuff, too, right? There's reasons why you don't want complainers on sites like that, too, and why you want...if there's enough Yelp ratings of a place, then you can probably trust it. It's just the ones that have like 10.

David: It's a hard problem.

Reuben: It is.

David: Craigslist is also very community-driven, and it's very difficult to balance the interests of everyone in those communities together.

Reuben: Definitely.

David: It's hard.

Ravi: Coming back to wikiHow and to the community itself. I think you might have touched on this a bit earlier, but what's your guys' plan to increase the number of communities, or increase the traffic to the site?

From what I know of Wikipedia is, you guys have, monthly, about 35 million, if I'm not mistaken. What are you guys' growth plans, to increase traffic?

Reuben: Do you guys know about the Panda updates to Google, that happened last year?

David: I don't, no.

Ravi: No, I don't either.

Reuben: Basically, what happened was Google made a change to the algorithm, and it really affected a lot of search results, including us. Essentially, a whole site would get penalized if they decided that the bar of quality was too low on that site.

David: Oh, I actually did hear about this. I think, what do they call them, link farms?

Reuben: Yes.

David: They were trying to combat link farms with this.

Reuben: And content farms, too.

David: Content farms. Maybe that was the term I was thinking of.

Ravi: What's an example of a content farm?

Reuben: Yeah. Perfect example. They actually pioneered the idea, which isn't a bad idea, but it's writing very, very specific articles to get data about what people are actually searching for, like "How to buy a car in Tucson, Arizona."

David: I think Associated Content might be one, too.

Reuben: Mm-hmm. Yahoo! Bought them.

David: Yeah.

Reuben: Like all of our competitors were basically doing this similar stuff. They'd pay people very little to write articles generally, like $30 an article or something like that to do very basic stuff. Then they'd put a whole bunch of ads around it. There would be some content there. There would be some value, but there were some egregious examples of horrible content too like how to do CPR. "First, put the person on their back. Second, press on their chest." You know? [laughs] Like just really bad stuff.

David: The idea is that it's junky content, right.

Reuben: It's really focused around key words and just matching what the person searched, because that's what Google did so well. They were just kind of feeding into Google's algorithm. Oh, people are searching for this. But Google decided, smartly, that they should be sending people to high quality pages so they started penalizing domains that had too much crappy content.

Ravi: So how do they determine the quality of a page?

Reuben: They obviously don't talk about their specific algorithms.

David: Right.

Reuben: At the time it was very, very opaque. But now people believe that there are actually people who rate content, a whole system of people who actually have like 100 questions that they'll answer about a web page, about sites on a site. Then there is this machine learning based on those answers to find out what sites kind of match this profile.

Ravi: They can determine the quality of an entire domain based on a few pages.

Reuben: Yeah, if there are like a whole bunch of spelling errors, for instance, on the site or bad links. There's a bunch of different things.

Ravi: I think I remember a talk about this. One of their key engineers in search was this guy called Matt Cutts.

Reuben: Yeah.

Ravi: He gave a talk about why they're trying to move links to content farms and such. To me at that point, obviously, it didn't make much sense. I thought content farms were obviously a good thing. Right? It didn't make much sense to me, but you have people writing localized information for users, so...

Reuben: Unfortunately, it's just done so cheaply that the quality was very poor. Now if you want to survive in the SEL business, it's really about writing high quality content and getting people interested and staying on your site for longer than your competitors.

Ravi: So is that what you guys are focusing on because of these set of changes by Google?

Reuben: We have always focused more on quality of content which is really lucky for us.

Ravi: [laughs]

Reuben: But, you know, the viewership is only one part of the site. That part is what makes money, but when you make, for instance, the mobile site that helps the viewership, but it doesn't necessarily focus on the community and writing new articles. We really spend a lot of time working on tools to help people create content on the site. That's one side definitely, the kind of SEO aspect. I think that there are other people that do that a lot better than us, and there are definitely other people who do ads better than us I'm sure.

Ravi: Something I'm learning as well, and David and I talked about this. There are definitely people who are champions of SEO. They look at your page or if you're a client they'll tell you how to optimize your SEO rating.

Reuben: I think before Panda, SEO always felt like a whole bunch of tricks that you do. You feed a whole bunch of stuff into the meta headers or stuff the URL full of key words, but I think that's changed a little bit. The fundamentals are still important, but it's much more difficult to tell someone, "No, you have to just make really good, relevant content to the search." That's harder. [laughs]

Ravi: Yeah.

Reuben: That's why video sites have done quite a bit better too. Like for certain search topics people actually need video...others they don't. Others they want standard book content. If you go somewhere and find out how to boil an egg, you don't really want to watch a three minute video before you do it, right? You want to just look at the steps, refresh your memory, and then go do it.

Ravi: Howcast actually has video for every single...

Reuben: Oh, they do?

Ravi: How to tie a shoelace.

Reuben: For people who can't do it, it's great.

Ravi: In case you hadn't noticed.

Reuben: Yeah. We were made fun of because we had an article called "How to Peel a Banana" on our site. There was a journalist who was like, "This site is just as bad as any of the other content farms." I can't even remember what they were saying. But basically our CEO responded. He's like, "Did you actually look at the article? There are like 15 different ways to peel a banana on here." [laughs] I'm exaggerating, but there are. You know? You can do it from the other end. You can crack it in half. Looking at it, it was a fascinating article. [laughs]

David: I'm sure it would be a great topic of conversation at parties. [laughs]

Reuben: Exactly.

Ravi: Hey guys, I know how to peel a banana. Consider the following, you can actually break it up in half.

David: Yeah.

Reuben: Who brought this guy? [laughter]

David: Reuben, I'm sure you'd be fun at parties considering how much knowledge you know from your own site.

Reuben: Yeah, that's what I do all day basically is just write content.

David: Nice.

Reuben: I'm shaking my head.

Ravi: Yeah, I think we're pretty much out of time. A parting question, though, I always like to ask guests...You've been in the industry for about nine years now. You graduated in 2003. I was wondering if you had any sort of advice for people who are either new to the industry or coming from colleges now or people who are in high school going to colleges or people who are in Canada thinking of moving to the Bay Area...what do you think?

Reuben: I've always felt like there needs to be more people going into computer science and engineering, especially in western countries. I really think that this is the future of jobs, and a lot more people have a knack for that than what they think. Just like mathematics. When you're young, you kind of like there's the stars of the class, and then other people are like, "Oh, I don't want to spend time on that" or whatever. There's people who are better. Same with athletics, too.

But I really feel like it's a very rewarding career choice. Great for your own personal growth. All of the jobs in our industry are awesome, right? Like half of them you can do from a cafe. [laughs]

It's just a ...

Ravi: You don't need to wear a tie to work.

Reuben: Yeah. A couple of the engineers we have, work in the Northeast. Yeah, like, I guess I'm focusing more on people who wouldn't otherwise be interested in computer science, but I just would encourage more people to go into it, because it's something that's learned, and I think it's really important to changing the culture, so that maybe it's a little less geeky, too. I think that's part of the reason why a lot of people are kind of turned off by it.

Ravi: I've actually heard of the programming profession being turned into less and less of a geeky profession nowadays, because of the whole concept of "brogramming." I don't know if you guys have heard of this, but there's essentially -- you should look it up. There's an entire culture, like frat boy culture of programming associated with...

Reuben: The Facebook movie, like "Social Network," or like...

Ravi: A lot like it actually. "Hey, guys, I've just checked in. Let's take a shot." Literally -- I kid you not. So you should look it up. If you don't know about this, you're definitely too old, Reuben.

Reuben: Yeah, I think so. [laughter]

Ravi: I think I'm the youngest among us. I feel very happy with myself.

Reuben: Yeah.

Ravi: Yeah, so that's pretty much all the time we have today, so thanks a lot for chatting with us, Reuben. Appreciate it.

Reuben: Great. Well, thanks for having me here.

Ravi: Thanks. [silence]