ReadWriteBuilders is a series of interviews with developers, designers and other architects of the programmable future.
Whether you know it or not, you’ve probably used Disqus. (If you’ve ever commented on a ReadWrite story, you’ve definitely used it. It’s the system we use here on our site.) According to Daniel Ha, 27, his company handles an intense load from 3 million websites, 200 million commenters and 1.2 billion unique visitors.
It’s strange to think then that the man responsible for stewarding that enormous pipeline of drivel and genius doesn’t actually care about blog comments. What he does care about are communities. In 2007, he and childhood friend Jason Yan founded Disqus (pronounced “discuss”) to unite the disjointed conversations that were taking place online. And today, his website commenting service powers much of the conversation pulsing through today’s Internet.
ReadWrite sat down with Ha to find out more about what it took to build Disqus, his take on the competition and what he sees as the future of blog comments.
A Silicon Valley Legacy
ReadWrite: What was it like growing up around here?
Daniel Ha: I grew up in Silicon Valley. My dad was a big early adopter. He worked in tech on the finance and biz side. One of the first people in his big family—he’s one of 10 siblings, and he’s one of the younger children, and he always has to be the first one to have a computer at his home, so my aunts and uncles could come see that. He’d always adopt new technology and play around with that. And he’d always teach me those things.
My dad taught me to work on cars and also how to build computers. So a lot of that initial passion came directly from him. And months into him teaching me those things, then I would spend all my time doing that, when he’s actually out at work. And he’d come back and I would teach him what I learned.
Once I turned 9, I started to learn how to program. That was almost purely driven by me loving video games. Programming was sort of the “taking your toy and try to see if there are Lego blocks behind them." I started to learn how to program by doing small game modifications. I learned more languages by doing tutorials, and by the time I was in middle school, between 1997 and 1999, I was working on a couple of different websites and web projects. It’s also when I met my co-founder.
RW: Jason Yan [Disqus co-founder and CTO].
DH: Yes. We became friends—me and him and a couple of other guys—because we liked to play the same video games and we liked the Internet—which is a funny thing to like today. But it was novel at that age and at that time, when you would really embrace it and do stuff on it, actually contribute and run chat rooms, IRC channels or participate in these discussion forums or build websites. That’s what we did together.
RW: You lost touch, though. And then reconnected again in college, at the University of California, Davis, right?
DH: Yeah, exactly. So we met when we were 13, in the same small town of Milpitas. We were families one and two on the sign-up list for broadband Internet when it came to town. We signed up months early because I was 12, and I’d pester my parents about this. I was excited to get it. He was the same way. He was adamant about getting Internet at home.
One of the first things we did together was a simple music site. We’d create accounts to upload music illegally and share with friends. This was before peer-to-peer music sharing, pre-2000. It was very popular at school. We’d seed it to our friends, they would request music, and we’d scour the Web for them. And we’d be involved in these chat rooms where people would trade files with each other.
RW: So how do you go from creating a music-sharing site to tackling blog comments? How did that happen?
DH: When we originally started off, we didn’t know anything about blogs. Still don’t care that much about them. What we do care about, the thing we started off with, is the notion of improving online communities. It was a very broad, pie-in-the-sky concept.
Growing up, respectively, Jason and I cared about these online communities. We ran discussion forums, we wrote software for them, we knew how powerful they were in connecting people—kids who lived in the Bay Area could connect with folks out in South Korea who are passionate about some topic. And that was something that was always just so cool to me.
I didn’t treat it like a social network. It was much more. Like, anything that I wanted to talk about, I could connect with folks [about it]. I used my online handle, and we’d spend time getting to know each other as people—without ever needing to reveal who I was, where I went to school, what my job was, who their families are.... The only thing we cared about was that they cared as much about football as I did. That was the magic of online communities to me.
So Disqus started out with that premise.
RW: And the name…
DH: I opened up Photoshop and typed in a bunch of different letters and vowels, to see what shapes they’d make. I was more interested in what the logo would look like than what the product actually did. And we ended up with the name Disqus. It was a nonsensical word with a .com available.
The only argument Jason and I had was, I really hated the name. I thought it was ridiculous because it had a “q” in it, it’s really hard to pronounce and it would not be memorable. But we were talking for a while and we’d already bought it. It was $40 and we figured, why buy another one?
RW: So you were locked in.
DH: We kept it. We always like the idea of discussion, and it sort of connected well. So we just ran with that concept.
Developing Disqus 1.0
RW: What were some of things you were hoping to do with the first version of Disqus?
DH: The initial vision was of portable online communities and identities. I’d create a persona because of what I’d been contributing to the community—the insights I delivered and the stories I told. And I’d have the things people cared about in these communities: a high post count, an early join date, a recognition in the people who have added me as a friend. Some sort of real, solidified place in that community. And then when I went to other communities, I wouldn’t be able to port that clout with me.
If you enjoy something new ... well, you are a different persona in a new place. But there are some accomplishments that I’d like to bring with me. The fact that I’m not just a fly-by-night troll, someone who’s a committed member of a community—I’d like to carry that over.
Essentially we built forum software, so users like me could pull all the discussion forums they were participating in, from message boards to chat rooms, Google groups, Yahoo groups at the time. Pull that in, manage them, and have one product that makes it easy for me to participate in these places.
We took what we had built for the forum to the website. I was trying to blog at the time. We were finished with our second year in undergrad, and [Jason and I] were scrounging for ideas. We were still in school; it was something that we did on weekends and after class. We’d go to each other’s apartments and work on this.
RW: What was it like, developing it? What were your primary challenges?
DH: It’s funny, the first thing we built for Disqus was actually our own tutorial. We were learning a new web framework, and we called it Disqus, after the name of our startup. We started building this as a way to learn Django, which is the webframe that we used.
We’d been building our previous projects using PHP, and I started to learn Python [programming language], which is very easy to pick up. We were just playing around with stuff, but hadn’t really built anything on the Web with Python yet. And Jason said there was this completely new framework coming out called Django, and it’s supposed to make Python very good on the Web. So the first pieces of code for this company was a tutorial we were putting together for ourselves.
The biggest challenges have been pacing along with that framework’s development. It was new when it started, now it’s really big. And Disqus has become the largest deployment of Django on the Web, in terms of traffic and usage.
A lot of challenges had to do with scale. Our application relies on being embedded on other people’s websites. That means we absorb all of their traffic. So every website can manage that themselves, but for us to absorb everyone’s makes it unpredictable. And it was a little costly in the beginning. There were key events that happened—celebrity deaths, elections, anything that resulted in a bunch of traffic…
RW: … and that meant everyone wanted to talk about it.
DH: People wanted to talk, and we'd scramble and have our mini “fail whale” [like Twitter] and try to figure out what’s going on. It took us a while. It took us a couple years to feel really confident in our ability to build infrastructure.
I don’t profess that we invented any sort of way for people to post comments. And we haven’t invested in the R&D for the best possible transaction of someone posting a comment on a webpage.
But I’m really proud of two things: our attention to user experience, or the fact that what we’ve done has really created an experience that people know, and hone it in a way that makes people want to use it more. And two, [that we] really think about our scale, and do it in a universal way that will get the behaviors that we want across sites, with support on the technical side.
RW: That’s a big challenge. You deal with a billion unique visitors, and 20 million comments. Sounds like you conquered the scaling issue.
DH: Yeah, it took us a while. Every time we had a big scale, model scale, we’d tell everyone we’re doing awesome. And a couple days later, it would start to break a little and we’d patch it together again. Today, we’ve rebuilt everything a couple of times already. [But] we’re really confident in the way it’s growing.
We’re doing eight to nine billion page views per month, but about 1.2 billion uniques. Half of those people are the ones who are actually interacting with Disqus. And of that, maybe 200 million or so are actually contributing something, and 50 million are active in a certain way. So the numbers start to get more granular as you go down, and our goal right now is to really own that entire user acquisition funnel and understand how do we get something closer to the 1.2 billion number, which is a far reach.
RW: Comments are a hot topic. YouTube sparked user backlash when they switched to Google Plus, Popular Science got rid of comments, others are handing them off to Facebook and Twitter. Some people might wonder if comments as we know them are broken.
DH: Well I think comments as a concept is a very loaded word, and we certainly know that. Because comments is a unit of transaction. We don’t want to improve comments, we want to build an environment so that better comments can exist.
[Here’s] an analogy: I use Instagram often. There’s going to be crappy photos indefinitely. You’re not going to solve the problem of people taking bad photos. What products like Instagram and social networks really focus on is the ease of creation, editing and sharing. We see our challenge the same way. We build our environments for these communities, but software can’t solve for everything.
There are going to be sites like Popular Science. There’s two parts of that spectrum: the media companies that treat it like a defensive measure, and the ones who look at it as their audience and their community, where they have the opportunity to talk to them in a very authentic and candid way… [For them] comments can become this huge asset, a huge competitive edge that no one else will have.
You mentioned YouTube, and that’s a very interesting example, because I’m really confused by their move there. I understand they’re trying to push Google Plus. But YouTube is not simply a video watching website. A lot of its early growth and soul came from its communities. These video bloggers, these people who created their livelihoods and following by building content on YouTube, there’s a big sense of personality there.
The comments are awful, that’s clear. I’m not arguing for that. But you don’t solve it by making sure everyone uses their program. Moving in that direction takes away from the communities that are good.
RW: Do you think software can actually deal with trolling successfully?
DH: Disqus is very good at that. We focus on it, if there are community managers and there is a community that wants to be a good community. Disqus embodies all the great characteristics and the personalities of good communities in software. That’s our goal, and that’s what we’re continuing to build.
RW: You recently released “embedded rich media” [which displays images, videos and audio files directly in comments via URL links]. LiveFyre has had this type of functionality for a while now. Isn’t it the same, and why did it take so long?
DH: I believe it is different. The product is, honestly, better. I think it works in a much more designed fashion. Other products out there that have done media did it in a haphazard way. You throw media in, and it sort of reacts off of it. [For us] this is one part of a full vision of how media is going to be used on Disqus. As for why it took so long … there’s no special reason. We just took so long on that.
RW: You once said that Disqus has the potential to shape the Internet. What did you mean by that?
DH: We’re a product that exists in a lot of different places. We touch a lot of people. A lot of those people may not interact with Disqus directly, or know they’re using Disqus. That’s a challenge and an opportunity for us.
People like to describe the Internet as the Wild West, with rules to create. It’s not as true anymore. It’s sort of like an Industrial Age where the infrastructure has been created and there’s a lot of opportunities on top of that. I think that’s where we are.
RW: What’s next for Disqus?
DH: We see a future where people can refer to Disqus when they want to discover topics. We actually do that pretty heavily through email, for example. We send them daily digests, and people find topics and click through to find new content from the publishers using Disqus. And we’re driving significant traffic with that, so there’s emphasis on discovery as a whole.
Other things that we’re doing is making sure people can use Disqus directly. You can’t wake up in the morning and say, “I really want to talk about badminton today!” and then go open Disqus and talk about badminton. But that’s been our focus. That will become a reality in the near future.
RW: If disjointed forum conversations are like Online Comments 1.0, and Disqus and its contemporaries are like Comments 2.0, what do you think Online Comments 3.0 will look like?
DH: It’s going to look different. Like the whole idea of it being a secondary, tagged-on part of a webpage needs to be inverted a bit. The way people discover content today, and the way that people care about certain pieces of content—a lot of it’s driven not only by social, but a lot of it by discussion.
Social’s more like friends or family recommending things, whereas discussion’s more like, “I want to find more information about this movie,” and I go to a place where all the big critics and fans are talking about it. I may not know them, but because they are passionate geeks around that community, they’re going to share a lot about it.
Images courtesy of Disqus