Adam L Penenberg's book Viral Loop examines the spread of "virals" - ads, films and emails that get forwarded by so many poeple that they become global phenomenoms - and how they have become the business model for some of today's most celebrated companies.
Here, Adam intervews James Hong, a co-founder of 'HotOrNot', one of the first viral phenomenoms.
PENENBERG: Can you tell me a little bit about your background?
HONG: I basically grew up in the Bay Area and went to Berkley as an electrical engineering computer science major. I then worked at HP doing sales engineering and then product marketing for test equipment, totally unrelated to computing or Internet. In 1997 I left HP and did my MBA. I started a venture-funded company towards the tail end of school, but it was like a B to B…
PENENBERG: What’s was it called?
HONG: It was called Red Ladder. There were a bunch of construction industry oriented vertical B-to-B market places. But, I quickly decided that verticals are more about the industry and not as much about the technology or the business model so I pretty much realized that I would getup in the mornings pretty uninspired. So, I basically told my co-founder that he could have the stock that I didn’t invest back because I just had to get out of there.
PENENBERG: So, you sold it?
HONG: No. As a founder, most of my shares were still vesting so I just gave the shares back.
PENENBERG: So you walked away from the company that you founded?
HONG: Yeah. I walked away from that and got together with a bunch of friends. I had a friend who was graduating from his PhD program, I was hanging out with Jim, my co-founder of Hot or Not and another friend from college and we all got together and quit what we were doing and said we were going to do a startup. We came up with a bunch of ideas and we ended up focusing on doing a dating site but a next generation one. It would be more interactive as opposed to doing something you could just do on paper. So we came up with a contest…
PENENBERG: When was this?
HONG: This was 1999. We came up with the contest idea because Jim was friends with a guy who was on an MTV reality show who was really popular at that time. He was this character named “Yes”. It was a reality show and his friend was on it and we thought it would be cool to promote the dating site. So we said that if you joined our contest, you could win a date with “Yes”. Then we said well that if you refer your friend to the contest they will get an entry into the contest but you will get another entry so the more people you refer the more likelihood you would win.
PENENBERG: That’s virality, what made you think of virality?
HONG: Well we were broke at the time and we had friends that started a company called AtWeb. They now run “SimplyHired.com”. The advice they gave us was not to go crazy raising money. So we took their advice to heart but back then the only example you had for viral marketing was Hotmail. We called the company Viral Promotions, which is actually not a good name for a company so we renamed it “Gill Spring”. We were basically convinced that on the web, commerce would fall into various categories. The large drop-ship type of goods, we figured would be commoditized due to price comparison. Then we said that there were the niches. It’s funny, I just pulled up my old business plan for it as I was cleaning out the HotOrNot office and I had terms like the social database and like I had the idea that the things that will make money are niche products that are hard to find. So we are thought that maybe viral promotions are good for those kind of companies because typically niches are based around hobbies or interests. If you are a horseback rider, you certainly have friends that are also horseback riders but average friends of yours are not going to want to know about where they can buy custom-made leather saddles. The idea was a more efficient way of passing around awareness of these companies via leveraging your existing customer base to get more customers. We created this platform, launched it, and never raised any money but within a month we had about 4% of Yahoo stores using it. It was off to a very good start and the concept seemed to work.
PENENBERG: Was it viral?
HONG: We were measuring the virality but free is very viral. Everything else is somewhat viral but not to the same degree.
PENENBERG: When was this?
HONG: We started in November 1999, launched by December or January and probably decided to shutter it in March or something like that. The first thing we tried to do was sell it. We got offers in the $2 to $3 million range from HomeStore.com, it was more like a bonus and they wanted to lock us out for five years. We moved to LA and not even the cool parts of LA. We ended up just like scraping it and then we all went off in separate directions. One guy started working on a company called ZipDash. The idea was what you would do once everyone was connected. I am really excited about the iPhone now because I am actually seeing where there’s a platform now so a lot of those ideas are going to come back. At this point, I started working on another company, more like a hobby, called Xmethods.com. We were the first Yahoo type-site for web services – a directory of publicly available SOAP web-services. It’s the concept of APIs and all that stuff. In the process of talking about web-services we came up with the idea of HotOrNot and it was actually very similar to something we had talked about before when we were talking about the dating site. The original dating site with collaborate filtering. You would come along and say who you think is attractive and then we are going to do correlations to show you new people.
PENENBERG: That looked like those people?
HONG: Yes. Effectively it’s the same vision for HotOrNot, you come in and you give your opinion on these people. We never really got to fully working it out because we came up with the idea for viral promotions and thought that was better. The idea is that as we watched who you like, based on that profile and those data points, we can help someone else find the same girls if you like the same ones but we are not going to waste your time showing you people you will not like. Then we could start a two way matching too.
PENENBERG: Its like E-Harmony for looks?
HONG: Yeah. E-Harmony is brilliant.
PENENBERG: How did HotOrNot then become a business?
HONG: Well, out of desperation. This was 2000; we had launched it in October of 2000 as a joke…
PENENBERG: How did you come up with the HotOrNot idea?
HONG: I mean the idea for HotOrNot did not come in one meeting; the factors that lead to it were over years.
PENENBERG: What were they?
HONG: The original dating site idea, the concept of the collaborate filtering. I was obsessed with Blind Date, a reality TV show. HotOrNot in its purist form is reality programming. Do you know Dave Viner? He had a phrase called the ‘two way web”. We were thinking a lot about viral marketing. I tried to dissect what causes things to grow back the because the things that are causing virality today are based on the communication platforms back then. We did a lot of thinking about it and what drives it.
PENENBERG: Who is “we”?
HONG: Mike Chu and Mark Crady, who were the ZipDash guys, and then Jim Young. We also tried to dissect what motivated people to spread things. There is a book called Influence. It’s my favorite book. I read negotiation books and the laws of power. I’m very much into psychology and what motivates people. So, Dave Viner was talking about the “two-way web”. The idea is that everything in the past is one way, like classifieds. eBay is two-way. So the concept of HotOrNot was kind of reverse engineered, saying that “hey, you know when I give my picture it’s an expression of myself and when someone votes on it, it’s a communication back.” So, you get a group conversation going around each number. After you rate someone we will show you how many people rated that person and when that person last logged in to check their score. The idea is if you stop checking your score we stop showing your picture because you have left the room. In the meeting where we came up with the idea, we were hanging out at our house and Jim and I had gone to a party the previous weekend and he said that this girl was a “perfect ten”. I’m always fascinated by the fact that Jim and I have totally different taste in women so I kind of doubted she was a perfect ten.
PENENBERG: Was she a perfect ten?
HONG: I never saw her; she was just a random girl at a random party. That’s why he was like, “did you see that girl?” Maybe he was hoping I knew who she was or something. We thought this had all these elements, back then, of being, in reality not viral but having strong word of mouth. If you want to get to technical definition, you have probably heard it already, there is nothing in our site that forced virality but back then, whether it was viral or word of mouth it was always based on the content. That definitely changed when Plaxo came out as far as I can tell.
PENENBERG: What did HotOrNot look like?
HONG: Exactly what it looks like now. We did not change HotOrNot virtually in any way but we added some other functions.
PENENBERG: How did you seed it?
HONG: At first we put up our pictures and we put up some pictures of girls. I didn’t go to Sports Illustrated. We quickly came up with the tagline that HotOrNot was fun, clean and real. Incidentally, you will notice that early in YouTube they took that tagline because YouTube initially started off by making HotOrNot with video. We launched Monday, October 9th 2000. We launched it by 2pm and I emailed 42 of my friends. It’s funny because the number 42 keeps coming up in my life; my current addressee is 42, 42 is the meaning of life if you have read “The Hitchhiker’s Guide to the Galaxy”. I emailed 42 of my friends and before I knew it we were starting to get not only more people coming to the site but submissions.
PENENBERG: How hard was it to submit a photo?
HONG: It became really hard. Initially, it was easy because you just submitted it to an upload just like any other site but that caused problems because of the bandwidth. I went roller-blading in the TellMe parking lot across the street. And I just went up to some random guy and said “hey, have you seen this website where you can rate people if they are hot or not?” He said no so I roller-bladed home and I started watching our logs and I was searching for TellMe IP’s and ten minutes later one came in and he went through a bunch and then all of a sudden a bunch of TellMe IPs came in. Apparently, he sent it on an internal list so I had about 37,000 unique IPs come to the site by the end of the day in 10 hours.
PENENBERG: How fast did the picture database fill?
HONG: I think we got about 200 pictures our first day. I figured we needed 30 to 40 pictures a day to keep it fresh so we were thought this was awesome but quickly the problem became the costs because we were doing this on a borrowed machine in a data center with real costs. So, bandwidth cost was only one problem, the other problem was that we were on a borrowed machine. Jim was grad student at the time so we drove to Berkeley at 4am and we reinstalled the site onto a machine that I got for free. It was the lowest end machine you could ever have and we basically hid it under his desk and stacked books in front of it so no one could find it.
PENENBERG: What was he studying there?
HONG: He got a PhD in electrical engineering and computer science.
PENENBERG: So 40,000 IPs the first day, then how much did they grow up to?
HONG: On the 8th day we broke a million pages a day. We did 1.8 million that day. So the school told us to get off.
PENENBERG: 1.8 million in how long?
HONG: 1.8 million pages per day, a week later. Two months later, we had broken into Nielsen’s top 25 advertising domains on the web.
PENENBERG: Did you have advertising on the site?
HONG: We added advertisement network on the site. The first one was 24/7 Media. We tried to get into Double-Click but at that time everybody was ramping down so we couldn’t get in. We started a moderation system where our users moderated the pictures for us. So the way we handled the scaling of the computing and bandwidth was we called Rack-Space which is the largest managed host company back then and basically cut a deal with them. So, we got the machines for free, we got bandwidth for free but the next thing we did is we stopped allowing people to upload their pictures to us and started saying, “hey, go to Yahoo or GeoCities and submit your picture there and give us the URL of the image.” We were basically stealing bandwidth. A month later, Yahoo turned off the ability for anyone to do that but then they white listed us. We did a deal with OFoto and said what if we changed it to, “if you didn’t have digital cameras go to GeoCities but if you have a camera go to OFoto and do it”? So OFoto started paying us for sending them people who had cameras so we started making money from it. By year-end we had made about $100,000 dollars. It was all profit because at that time, we weren’t paying for our machines or our bandwidth so there were no costs. The day after we launched it was in Sloane.com and by the end of the week we were in every major European paper and by two weeks we were in the New York Times.
PENENBERG: What did you do after that?
HONG: We realized that we had to make money and that all the press is going to die at some point, which is why we added that dating component. It was meant as a retention tool because we are getting all these users but once the press dies our users are going to drop and it will die really fast. You could see the press spikes but over time the press spikes started getting drowned out by our consistent users.
PENENBERG: How did the dating work?
HONG: Instead of rating them from 1 to 10 you can add a short profile and you can say Yes or No to people. and basically one person at a time we would show you people you can say yes and no. So it’s basically going through Facebook and going yes, yes, no, yes, no, no.
PENENBERG: What was the button that you had under your site that said about dating?
HONG: As you are rating people, the ones who opted into the dating personals, we would frame it with a colored border and at the top it said “Click here to meet me.”
PENENBERG: So they had to register?
HONG: If they had their picture in there then they were already registered.
PENENBERG: And you did it off Yahoo?
HONG: Yes it was still “go to Yahoo, submit a picture, come back and give us the URL.”
PENENBERG: When you were first there, how much revenue did you make?
HONG: We launched the dating service in February or March, but it wasn’t a fee service, it was free, remember we are just doing it for retention purposes.
PENENBERG: But you had advertisements.
HONG: Well, the advertisements were dying. We realized pretty quickly that we would need to make money some other way than advertising. It was clear from the very beginning that CPMs were drying up so we knew we had to come up with something else, we just didn’t know what.
PENENBERG: How much did you start charging?
HONG: We were charging $6 dollars a month.
PENENBERG: That’s very low, a lot for you but low for the users. So, when did you launch the dating?
HONG: It was in February of 2001 and we started charging around April.
PENENBERG: How did that go?
HONG: I think our rate was doing about $300,000 a year or something like that. By the end of the year we did $600,000 I think, then we just doubled every year for a long time.
PENENBERG: So you were making millions within a couple of years?
HONG: Oh yeah, even the run rate at the end of the first year had us doing probably $1 to $2 million in revenue and growing on a nice cliff.
PENENBERG: So, what are you doing now?
HONG: I am just taking a break for now.
PENENBERG: When did you do sell the company?
HONG: We sold it in February. We ran it for four years and we had hired employees and then we realized that we had to manage those employees. We both had made enough money from it to retire already. I was telling my girlfriend at that time that I’d probably be happier if HotOrNot just died because even though it didn’t take that much work you are always worried about your baby. We did minor tweaks here and there, you know like, does it work better? Do people vote more when the icon is up on top or on the side?
PENENBERG: What’s the answer?
HONG: I think on the side because the picture shifted up which means less scrolling and more interaction. Basically the net conclusion is anything that makes it easier for the person to click to get to the next section, the less friction, which will make them stick around much longer and pass it around much more.
PENENBERG: Can you tell me how much you sold it for?
HONG: I can’t say exactly what it’s sold for, there is number that floated on TechCrunch which is not exactly right but the magnitude is right. They said 20. We didn’t sell it for 100 but we had already made pretty decent amounts just up front.
PENENBERG: Yea you made more in revenue while you ran the company. So, when YouTube was first conceived, it was conceived as a video HotOrNot?
PENENBERG: You know that as a fact?
PENENBERG: They talked to you about it?
HONG: I think Steve mentioned it in The Times magazine when Steve was “Man Of The Year.” Jim is very good friends with Steve. We are also very close to PayPal and we were all in Mountain View and all the guys used to hangout at the same café we hung out at.
PENENBERG: What is the name of that café?
HONG: Dana Street Roasting Company.
PENENBERG: So Steve was at PayPal at that time and so was Chad and they teamed up to create YouTube later. When was this that you were hanging out at this café?
HONG: We have always hung out at that café but I think we were hanging out with them around 2003-2004, right around when Friendster came around.
PENENBERG: Are you planning to start a viral company again?
HONG: Yeah, sure, but what does it mean to start a viral company? Lets just put it this way; I hope to build a product compelling enough that people will want to spread the word but in the sense of it, I aspire to create something that is viral by nature as a communication method, not spamming.
PENENBERG: Who is guilty of that?
HONG: It’s your definition of spamming. But who is guilty of address book scraping? Every decently sized social network on the planet. If you look at how all these companies have behaved, they got big and fat first by any means necessary and once they have the luxury of scale they can back down and preach ethics. With MySpace it got them to critical mass where press started writing about them and artists started having MySpace accounts like UB40. I am involved with Slide and I know the RockYou guys pretty well.
PENENBERG: What are you doing with Slide?
HONG: I am an investor in it and I go in there once a week and kind of help out with a product a little wherever I can.
PENENBERG: What do you say to people like RockYou, who had trouble rising?
HONG: Well, that’s just a function of the economy right? They are not growing like they used to but they are not disappearing either because to some degrees users are still using them. It was almost like they were taking the ability to go from zero to a million users for granted because basically Facebook did all the heavy lifting for them. It’s so easy to start a company now because the capital costs required are so low that any guy or girl can just do it. You can get an account on a managed host for very little money, or with Google application engine which is free so. What that means now, it’s like blogging, and you will see a lot of crap out there too.
PENENBERG: Here is an American question for you. If you look at some of the greatest innovations of the 20th and 21st century, they came out of America. Why America?
HONG: The first thing is cultural like especially in enclaves like Silicon Valley where you are encouraged to do it, and I also encourage people to do it even if I am not sure if they should. They are so cemented to be conservative on the other side.
PENENBERG: So in failure you learn something?
HONG: Yeah. If you fail so what, try again right? So one reason is just being ok with people failing and supporting them and trying. In a lot of countries you are expected to get it right. So that’s one reason. The other reason is that we have enough mobility. Basically people who are hungry can still succeed, if they work hard. I am not saying that people who are at the bottom and hungry, that its not harder for them. Like if you are too rich you are not as hungry and if you are too poor the odds are very low that you will get the good education needed but its kind of like that sweet spot in the middle where you get hungry and capable and we allow social mobility, we allow people to try things. In Europe social mobility is a lot tougher because it’s a society that was built on a legacy of structure and class. It’s a lot more difficult to be an entrepreneur in Europe I think, I lived in London for a half a year.
PENENBERG: When did you live there?
HONG: In 1998 during my MBA. I did an exchange thing to London Business School. I don’t think LSE has an MBA program. They have a Masters program in economics; it’s a real degree as apposed to an MBA. You know China though, people are taking risks; they are the hungriest country I’ve every seen. A lot of the companies out there are funded VCs because there are first generation VCs out there and many of them were educated out here and they are not really creative themselves but are just trying to copy business models from here to there. The largest market capital Internet company in China is a company called 10 cents. They copy IM but they are entirely making their money off virtual goods.
PENENBERG: What do you mean by virtual goods?
HONG: Buying items for your avatars and that kind of stuff. All the virtual goods get money from your users for virtual items.
PENENBERG: So how are you liking the iPhone?
HONG: Yeah, love it. It’s just so awesome.