Subscribe to my blog

James Hong

Friday, July 13, 2007

Who says a Facebook strategy is pointless?

Thanks to my buddy who has access to Comscore's MediaMetrix, it's nice to see the effects of all our hard work!

A significant portion of our Facebook traffic is actually from our Hotlists product rendered as a Facebook Application.. It sure would be interesting to start targeting brand ads based on specific user brand preference data ;)

Right now we don't have any in-house ad salespeople.. anyone interested? contact us at hotornotjobs (at)

we are also looking for a few more good coders.

Monday, July 09, 2007

Reinventing HOTorNOT, Part III

This was actually the beginning of Part II, but I realized it was a bit too long and so I cut this part out.. In the interest of keeping it, I've made it a separate post. This gives the higher level explanation of how we came to our strategy


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:

1. The way we see our business is different than how most people see it
2. 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.

Monday, July 02, 2007

Reinventing HOTorNOT, Part II

1 + 2 = 3. Our new product strategy

ok, enough blabber about why we needed to change, let's talk about what we are doing.
These 2 points are the genesis of the long term strategy for our company, which I explain in part 3 of this post.

1. We are about connecting people
2. Our name sets wonderful context for sharing opinion on anything

Add these together and you get

3. 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. Here is an example of one:

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.

This post was part 2 in a series, to learn more about our company's changes, be sure to see part 1 too.