Letter #122: James Hong (2007)
Founder of HotOrNot | Reinventing HotOrNot
Hi there! Welcome to A Letter a Day. If you want to know more about this newsletter, see "The Archive.” At a high level, you can expect to receive a memo/essay or speech/presentation transcript from an investor, founder, or entrepreneur (IFO) each edition. More here. If you find yourself interested in any of these IFOs and wanting to learn more, shoot me a DM or email and I’m happy to point you to more or similar resources.
Today’s letter is the memo written by HotOrNot cofounder James Hong after returning to the company in 2006. This was a turnaround memo, as HotOrNot had stagnated at the time, and James was returning to try and breathe life back into it. In this memo, he lays out two things that changed that they had to change with (improved startup economics made it harder to retain talent and improved online advertising made free competitors a reality), and how they changed to address those industry changes. He then explains how he thinks about the business, in particular how they see their business is different from how most people see their business, and the importance of their name not only as a brand but a context setter. Finally, he lays out their new product strategy, which is built upon the two principles of connecting people and the sharing of opinions. He proceeds to introduce a new product (hotlists) and then explains how the user will benefit as well as how it will work as a business.
James Hong was the cofounder of HotOrNot, which at one point was one of the 20 most visited websites in the world. It’s also one of my personal favorite companies in Silicon Valley history, with a few of the reasons listed below.
First, they were a constant innovator, having developed a payments API dubbed HONPAL (Stripe before Stripe), a social network photo-sharing site called Yafro (Instagram before Instagram), the “double match” opt-in communication that was the foundation of their Meet Me feature (Tinder before Tinder), a virtual goods-based economy (now popular in gaming, dating, and crypto), and a gamified moderator system (now used by Reddit and Wikipedia). The engineering team was always tinkering with new technologies and ideas, having been one of the first beta testers of Google AdSense and creating the platform that eventually spun out and became Crunchyroll (a wildly popular anime streaming website that today boasts over 100mn registered users and over 10mn paying subscribers).
Second, beyond innovating, HotOrNot’s early success was an inspiration to a Silicon Valley that had just been decimated by the dotcom crash. According to YouTube cofounder Steve Chen, one of HotOrNot’s greatest impacts was as a singular example of a tech startup that was able find success following the crash, reaching profitability through scrappiness and virality rather than burning cash on marketing.
Third, they supported the Silicon Valley ecosystem. With all their success, they could have retreated and gatekept their knowledge and resources. But they did quite the opposite. In fact, they let their engineers pursue their own interests at work and let them have the rights to their work, free and clear, when they left (a la Crunchyroll). They also gave unrelated startups years of free hosting, including Twitter, Bit Torrent, Mochi Media, and Zipdash (which became Google mobile maps).
Lastly, their influence on Silicon Valley is undeniable. Facebook’s origin story starts with Facemash (a pvp HotOrNot), Twitter was initially hosted on HotOrNot servers and their double opt-in DMs were inspired by HotOrNot’s Meet Me functionality, YouTube started off as a video HotOrNot dating service and YouTube cofounder Steve Chen has frequently cited HotOrNot as an inspiration and even almost joined, and Tinder is a near direct copy of HotOrNot’s Meet Me and their super like is a copy of HotOrNot’s virtual flowers).
I hope you enjoy this letter as much as I do!
Related Resources
Social Companies
Meta Platforms Deep Dive (Full report: 166 pages)
Meta Platforms Business History (~18 pages)
Tencent Deep Dive (112 pages)
Letter #66: Pony Ma (2021) - Consumer, Business, and Society — The CBS Trinity
Letter #77: Mark Zuckerberg (2012) - Facebook S-1 Shareholder Letter
Letter #83: Evan Spiegel (2014) - Financing, Hiring, FB, Lightspeed, and an All Hands
Letter
Ever since I came back to HOTorNOT in October 2006, people have been asking me what the company is up to. For the previous 6 years, HOTorNOT had pretty much been steady along the same course: A picture rating site with a dating-like application built on top (without the seriousness of a dating site) that generated healthy amounts of cash for my cofounder Jim and I. For the first 3.5 years, it was Jim and I working about 10 hours a week each, with the company earning many millions of dollars per year.
Then 2 things changed, and we realized we had to change with them:
Startup economics improved, making it harder to keep good people
The Online Advertising Market improved, making free competitors a reality
1. Startup economics improved, making it harder to keep good people
Three and a half years into running HOTorNOT, Jim and I grew tired of having our personal lives dictated by the health of the website. So we decided to hire employees to help us run the cashcow… but as I have mentioned in the past, getting really smart people to be happy running YOUR cashcow in silicon valley is nearly impossible. You have to either decide to grow a large organization and institutionalize things (but also recognize that the average caliber of the team is likely to drop… and these are the people you have to work with every day), or you have to recognize that those people are going to eventually leave. By late 2006, all of our employees wanted to leave to go do their own things. They recognized that the costs of doing a startup had gotten so low, it made less sense for them to stay at HOTorNOT than start their own startup.
This was of course our own doing.. after all, we hired them BECAUSE we thought they were smart and entrepreneurial. Most of our employees were also younger and have nothing to lose, due to the fact that we mostly recruit straight out of Berkeley. We learned a big lesson here: don’t expect smart, young people to do anything that you wouldn’t do yourself. Jim and I both had to admit, if we were 22 years old, we wouldn’t stick around running someone else’s cashcow for no equity, even if we earned salaries 2-3x the normal wage. We either had to let them go or give them equity.
Problem: We needed top talent, but working at HOTorNOT sucked. Employees shared no upside. To make matters worse, they were encouraged not to take risks because Jim and I were overly concerned with preserving cashflow, making the job boring.
Solution: We finally created a stock option plan for our employees. I will explain in greater detail in a future post why we did not give options to employees in the past, but for now I will just mention that we were an S corporation and that that caused complications. Giving the team a large cut of equity has aligned their interests with ours and we’ve seen it breathe an incredible amount of new energy into them… as it should. They are now working for themselves now and not for us, and we think that is a better situation for everyone.
The other thing we did is start encouraging innovation again, rather than squelch it in fear of our any changes hurting the cash cow. Changes are now made to the site that I don’t always agree with, and in some cases I don’t even know about… and as long as the team is measuring and testing the effectiveness of everything, that’s ok.
(as an aside: our original vision was to become an incubator and to enable our employees to work on new ideas, and let them spin those off as separate companies.. basically let our employees graduate into becoming funded entrepreneurs at a time when funding was hard to get. Our first and only attempt at this was back in 2003 when we hoped to work with Steve Chen and Mike Solomon to start Yafro.com, which was going to be a social networking site with media sharing applications built on top.
In the end things did not work out because some members of our board were uncomfortable with the idea of giving the employees of a spinout majority share and control… so Jim and I agreed with Steve and Mike that it was a no go. It’s hard to say what would have happened if things had moved forward with them, but given Steve and Mike’s huge success with YouTube, it is easy to presume that this was a huge mistake. In general, Jim and I both believe (especially now) that it is better to find people you believe in and take a chance on them than in trying to control and own them. People will work a lot harder if they are working for themselves and feel in control of their own destiny than they will for you.
In the future of the web, the majority of value is in innovation and the quality of execution, not in the funding resources a company can provide… giving employees a healthy share, and a majority share in the case of spinouts they are primarily responsible for is not only not a bad idea, it’s the BEST idea. Another example of this is Yelp, which was a spinout of Max Levchin’s incubator.)
2) The Online Advertising Market improved, making free competitors a reality
The second thing that changed was the development of online advertising. When we launched HOTorNOT, we had no choice but to charge subscription fees for it. That was the only way to pay our server bills because CPMs on ad networks for sites like HOTorNOT dropped to about 3 cents (that’s 3 cents per 1,000 ad impressions shown). So we developed a subscription service on top of our dating service, and that quickly became very profitable.
But what happens when an advertising model DOES provide an adequate amount of revenue, even if just for 2 or 3 people? That means it is now possible to offer the same scale of services and still be profitable, entirely on an ad model… and from the customer’s standpoint, if they can do something for free versus the same thing for charge, which do you think s/he is going to pick? This has enabled free sites like Myspace, Facebook, and PlentyofFish.com to pop up, and it is a real long-term threat to most subscription sites.
There is no doubt in my mind that HOTorNOT’s traffic started to drop around 2004 due to free alternatives, primarily social networking sites. It’s not that these services made HOTorNOT worthless to users, it’s just that they had alternatives occupying their time.
While HOTorNOT’s profitability was still extremely strong thanks to a large loyal base of paying users we’d built up over the years, we saw the writing on the wall.
Problem: Free competitors
Solution: We decided we had to stop being conservative in our actions, and in our desire for cashflow. Earnings became like a drug addiction to us, to the point where we stopped innovating and we became more focused on making sure our next dividend check was coming (and to hell with longer term trends!)
So the first decision we made was to go cold turkey and make HOTorNOT free. If Subscriptions were our past and Advertising and Digital Goods were our future, we had to take the plunge at some point, no matter how painful.
This has much broader implications than you think. First, we now had less to lose by being aggressive (countless right moves were killed in the past because we were worried about what it would do to our bottom line). The other thing this did was it enabled us to let users create more user generated content on the site. User generated content is the real power of the web... but when you run a dating site, one of the things you have to do is screen ALL content to make sure nobody is hiding their contact info in their profile somewhere. Screening all content on a UGC site doesn’t scale…. but under an advertising based model, letting users upload more data doesn’t threaten the business model, it helps it. If a user wants to put their contact information in there, so be it.. it probably only helps us now.
Stop Clinging to the Past and Jump into the Future
While these changes may sound small and incremental in theory, in practice they are not. HOTorNOT is a completely different company to work at than it was just 6 months ago.
It used to be a cashcow where the employees didn’t hold any equity, and where innovation generally took a backseat to income preservation. It was so miserable to be at, even the founders left.
Today, it’s a pretty different story. Things are not boringly comfortably, they are more risky and exciting… but that’s ok because employees now hold equity, so they work hard and they think about and build things for HOTorNOT with general guidance but low supervision... They take more initiative now and are encouraged to take risks… and they feel a larger sense of ownership and pride in their work. On the side, they’ve even built 2 major Facebook apps (Moods, which has almost 2 million users in only 3 week, and Pets, which has 200k very active users). Both of these products may be spun-off as separate companies in the future, with employees involved likely keeping a substantial amount of equity in order to give them a majority of the control and a majority of the upside.
Most importantly, working at HOTorNOT is a lot more fun. Traffic has doubled in the past 3 months (doing about 20 million pageviews per day now), people are building cool new things that users love, our newer employees are learning more here than they could ever learn at a big company, and people are now working hard for their own upside.
Making these internal structural changes was essential to reinventing the company for our employees. Once these changes were in place, our newly motivated team started getting to work on reinventing the company for our users. We are now pointing HOTorNOT in a strange new direction that some have called crazy, others have called brilliant, and a few have called both crazy and brilliant at the same time.
ok, enough blabber about why we needed to change, let's talk about what we are doing.
To explain how we came up with our new strategy, I need to first explain how we think about our business. There are actually 2 points I want to cover:
The way we see our business is different than how most people see it
Our name is extremely valuable, not only as a brand but also a context setter.
Point 1: The way we see our business is different than how most people see it
When we first launched HOTorNOT, we used to wonder if the site would die quickly like many other web memes. After all, we saw it as most people saw it:an a fun but silly site where people looked at other people’s pictures and said if they were Hot.
Fun and innovative? Yes. Sustainable? Not as likely.
But then one of our former advisors, the late Richard Newton (former dean of Berkeley Engineering, venture partner at Mayfield, pioneer of SPICE simulators, and advisor to the founding teams of Cadence and Synopsys) said something so insightful it made us realize that HOTorNOT was more than we had originally thought...
“PEOPLE are the killer app,” he said. Richard realized that what we had was a platform that attracted people and could connect them. Around those words, we decided that our company’s goal would be to become “the ultimate people router”.
Our focus on connecting people, not just on showing pictures, would be the core of HOTorNOT’s sustainability. Connecting people around dating was the first obvious extension of our service, so we launched a dating section on HOTorNOT just 3 months after launching the site… but dating is only a subset of the types of relationships people have an interest in, so to keep it generalized we named the section “Meet Me at HOT or NOT” instead of something like “Date Me at HoN” or “Find your soulmate at HoN”. It was our hope that people would use the system not only to find their soulmate, but also just to meet new friends around common interests.
Richard saw the potential of the Internet, and of our site in particular, as a means to deliver something more important to people than even information or search. We want to deliver people, because people are the killer app.
So as I mentioned, how we see our company is not how most people have seen it over the years.
Our company is not about rating people, it is about connecting them.
Point 2: Our name is extremely valuable, not only as a brand but also a context setter.
The word HOT is interesting because unlike most other words that are used to describe something as being really good (like awesome, killer, tight, rad, groovy), the word HOT is not a fad. According to an article I once read, the word hot has been used in that context for at least a millennium. In other languages, the equivalent word for hot is often used in the same way.
The article hypothesized it is based on your human physiological response to anything you find exciting: your heart races, you start to sweat… quite simply, you feel hot. This is why spicy foods are also often referred to as being hot.
In a nutshell, because we are all human, the word HOT is not a fad.
And then there’s the NOT part. Not rhymes with Hot, which makes HOTorNOT insanely easy to remember. Rhyming may very well be the best mnemonic device we have, and it makes things catchy.
All of this makes HOTorNOT a great name… sustainable and memorable. But there’s one other thing that HOTorNOT does really well… it sets the right context.
Countless companies, many of them huge megacorps, have tried copying our picture rating service under different names. AOL tried, Myspace tried. But none of them really got their products to take off. Somehow, for some reason, the name HOTorNOT sets just the right tone of being fun and puts people into the context of sharing opinion on things in a way that the names “rank” and “rate my buddy” don’t.
But is HOTorNOT limited to setting the context around rating pictures of people? No. The word HOT can be used to describe lots of things as being good… a movie can be hot, a car can be hot.. Almost anything can be HOT. So, it turns out, HOTorNOT is a great name to create a context for sharing opinions about ANYTHING.
So this got us thinking... HOTorNOT as a picture rating/dating site could certainly be a profitable business, but an empire could be built on top of the brand HOTorNOT once we start extending it and applying it to everything.
1 + 2 = 3. Our new product strategy
These 2 points (explained above) are the genesis of the long term strategy for our company:
We are about connecting people
Our name sets wonderful context for sharing opinion on anything
Add these together and you get:
HOTorNOT enables people to share their opinion on anything, and helps connect people around those opinions.
(You’ll note that our prior business fits into that category, if you replace “anything” with “pictures of people”.)
We are confident that if we stay true to this mission, we will make products that will impact a large number of people’s lives in an utterly positive fashion.
Plus, the timing is right. Based on the way we plan to monetize these communities, the main thing really missing to support this empire was the emergence of online brand advertising. Last year we realized this last remaining piece was starting to happen.
Overview of HOTorNOT Hotlists
3 years ago, we started testing our current rating interface and had friends rate pictures of shoes on a scale of 1 to 10. Not surprisingly it was pretty boring, so we quickly realized our future might involve building a different interface. To that end, we came up with our new Hotlist product.
The concept: wouldn’t it be cool if users could, as a means of self expression, display pictures of the things they think are HOT on their HOTorNOT profile? Kind of like how people like to put posters of things they like on their walls, or how people like to wear t-shirts with logos of things they like.. it can be anything.. bands, tv shows, movies, clothing brands, colleges, products, even non-tangible ideas!
Each picture (what we are calling “Stylepix™”) can be anything that a person associates with as an element of his or her own style.
In essence, Hotlists are a form of having visual keywords. Visual keywords are different from text keywords (“aka tags”) in that with text, the more information a user adds, the less anyone else wants to read them. With pictures, people are more likely to look, they are able to comprehend the data faster, and they will remember the list better than they would a text list.
For example, here is a screenshot of my HOTorNOT profile with my top 8 “Stylepix” at the bottom. If you clicked on the “Show All” link, it would take you to a page with my entire Hotlist.
One cool part about the system is that when you are looking at someone else’s Hotlist, if you see something that you like, you can add it to your own Hotlist simply by clicking on a plus sign that appears over the stylepix. You can try it out yourself by taking a look at my Hotlist.
We recognize of course that this sort of functionality should not be limited to HOTorNOT users alone so we created exportable widgets too.
By enabling people to add Stylepix to their profiles, people are able to define themselves through a collage of entities that already have well known attributes. In other words, I am expressing my “james hong brand” as a mashup of many other well known brands that I identify with. It doesn’t have to be just brands, by the way… we encourage people to list anything they identify with. If you hate corporations, then be sure not to stick any company brands on your Hotlist. That’s cool with us.
So HOTorNOT is now enabling everyone on the web to define themselves with pictures. As a Hotlist user, all you have to do is browse other people’s Hotlists or search our directory and hit plus signs. Yes, you could do this yourself manually if you wanted to, but doing so is prohibitively tedious whereas building your Hotlist is actually a lot of fun. The beauty is that only one person has to bother submitting a picture to create a stylepix, but everyone benefits. Because of this, the directory is already fairly comprehensive.
How does the User benefit?
1. Hotlists are a means of self expression. In a world where everyone’s social networking page starts off looking exactly the same as everyone else’s, this concept becomes very important. The response in usage we have had so far indicates that many users like this product a lot.
2. By understanding what a user likes and doesn’t like, we are hoping to do some serious data analysis to start letting them know what other things they would probably think are HOT and NOT. In a world where users are willing to tell us more about them, we should be able to give them more tailored information back. General lists are cool, but tailored lists are cooler.
For example, remember how America’s Funniest Home Videos used to always be a highly rated show on Nielsen’s lists? I pretty much hated that show. On the flipside, I loved Veronica Mars, a show that just got canned. People are individuals who have individual taste. Lists should honor and respect that individuality.
In doing so, we are going to connect people not only to other people, but to other things that they would like too.
3. We are building communities around each Stylepix, enabling people to find and communicate with each other.
This is actually the coolest part about what we are doing. It is still super rough around the edges, but I think the concept is demonstrated a bit by what we’ve built already.
Each Stylepix submitted to the system (they are ALL user generated) has a Stylepix page that will become the basis of a community. Want to find a NY Yankees fan who might have tickets to sell? No problem. Want to find a Hot Girl that likes Linux? Believe it or not, also not a problem.
Not only will the Stylepix page connect people to other like-minded (like-styled?) individuals, it will also provide a place for people to talk about the Stylepix and share their opinions. Over time, we plan to do some really cool stuff on these pages to make interaction more fun and more interesting.
Here’s something else a bit different: our goal is not to build these communities for HOTorNOT’s users only. We are happy to overlay these communities on top of existing social networks and connect them all. In fact, we already have over 500,000 people on Facebook are already using the Hotlist product, and they can communicate on the Stylepix page with any other Hotlist user whether she comes from Facebook, HOTorNOT, or anywhere else we distribute the system to. There are a lot of cool things we are going to do with the data in addition to the cool communities we are building, and we are fine with the entry points to the system being distributed rather than centralized.
If you want to try building a hotlist for your Facebook profile, click here.
How is this a business?
Turning away from a subscription model is hard because it is a simpler business. We know we created value for our users because over 15% of free members ended up paying for it. How does our new strategy fit in with making money?
1. Improving Ads on HOTorNOT
First, making this move will enable us to raise the effective CPM rates on our site. If I know you are interested in computers, I can show you techie ads instead of random "punch the monkey" ones. Even better, if you happen to be on a stylepix page, that ad can become even more targeted. Are you on the Verizon Wireless page asking people if Verizon has good coverage in Memphis? Maybe Sprint would like to show you an ad. In essence, HOTorNOT pageviews used to have little context for targeting and few themes for channelization. With this data and with these communities, we will have plenty of both.
2. Use the data collected to build a superior Brand Ad Targeting Engine
The way brand ads are targeted today is based on inefficiencies of historical media platforms. Making advertising decisions based on the information that I am an Asian male, aged 34, living in San Francisco, with a 60% chance of being in the $75-100k income bracket is better than nothing, but still limited. The industry evolved to that standard because that was the best targeting publishers could provide.
We think the future is a lot richer than that, and we think knowing what someone’s style is (what bands they like, what clothing brands they like, what beers they like, what music they like, what whatever they like) can help us give them a better experience. If you are going to be shown ads, the ads might as well be interesting ones that start bleeding into being content. (Ever notice how fashion magazines are often half ads... and how people actually ENJOY looking at those ads?)
We know from experience that brands cluster. People also cluster (remember high school?) And finally, people-clusters cluster around brand-clusters. By utilizing a wealth of explicit user preference data, we think we can ultimately make your advertising-supported experience on HOTorNOT suck less. In fact, maybe we can make it useful and enjoyable. And if we’re really ambitious, we can do it outside the confines of our website too.
3. Helping Buyers. By connecting people who may be able to answer questions about the products they love to people who might be making purchase decisions, we can start helping people who are in the market to buy things. We believe the conversations that will happen in our Stylepix communities can be used to extract a lot of valuable information for people who are trying to decide, for instance, whether the iphone is worth getting now or not.
4. Market Research. There is a large industry centered around giving marketers high level, aggregate data about their brands. Who likes their brand, what other things correlate with their brand, how is their brand trending over time, in specific locations or among a specific demographic. Most interestingly, we can also figure out who the trendsetters are.. which people tend to add the next big things first, and what are they adding now? We think we can bring Coolhunting to a new level.
In closing...
So there you have it! :) We believe the shift we are making is pretty large. It was emotionally more difficult than we thought it would be to pull the trigger on so drastically reinventing our company. But we’ve gotten some encouragement along the way.
We’ve gotten a lot of validation on our new plan from people that we respect immensely. We’ve gotten a lot of interest from big companies that want to get involved, both from the data side and from the community building side. But the best sign of validation comes from our users who are signing up like crazy and telling us they love what we are doing. Fundamentally that is what matters most.
If you are a programmer and are interested in joining us, please check out our jobs page! Not only will you get involved in our plans, you will have the opportunity to learn everything you need to know to create and run your own web company.
Postscript
Roughly 17 years after writing and sharing this memo (and 15 years after selling HotOrNot), James reflected on the note:
The main thing that failed was that even after I posted this, we did not hire enough people unlike us, so we just never had an ops team to keep things going. And we let our employees work on their own startup while still at hotornot, and that was Crunchyroll which started taking off like crazy so it made sense for them to take off which created a problem for us. We were just so tired of hotornot and didn't have anyone to hand it off to, so we just let it go. (we had hired someone previously but he was not a good fit and we were forced to get rid of him by our team). We just ran out of energy caring about the business.
The final lesson that I take from a lot of our mistakes though is this: if you are a very smart and motivated group of people, don't listen to others on what you should do. Just trust your/your team's gut and ignore the "experienced" folks. Advice can be dated as the world evolves, what was once smart moves can quickly become bad advice. Always listen, but also always think things through from the bottoms up. If you're smart, you will figure out the right moves. Have confidence in your ideas and move forward with gusto.
Wrap-up
If you’ve got any thoughts, questions, or feedback, please drop me a line - I would love to chat! You can find me on twitter at @kevg1412 or my email at kevin@12mv2.com.
If you're a fan of business or technology in general, please check out some of my other projects!
Speedwell Research — Comprehensive research on great public companies including Copart, Constellation Software, Floor & Decor, Meta, RH, interesting new frameworks like the Consumer’s Hierarchy of Preferences (Part 1, Part 2, Part 3), and much more.
Cloud Valley — Easy to read, in-depth biographies that explore the defining moments, investments, and life decisions of investing, business, and tech legends like Dan Loeb, Bob Iger, Steve Jurvetson, and Cyan Banister.
DJY Research — Comprehensive research on publicly-traded Asian companies like Alibaba, Tencent, Nintendo, Sea Limited (FREE SAMPLE), Coupang (FREE SAMPLE), and more.
Compilations — “A national treasure — for every country.”
Memos — A selection of some of my favorite investor memos.
Bookshelves — Your favorite investors’/operators’ favorite books.
Sorry, but this guy sounds like an asshole.