Futuristic Play by Andrew Chen

Eyeballs versus dollars: What should startups focus on?

When you read about websites in the mainstream press, you often hear about two kinds of consumer Internet properties. First, there are social, high-pageview sites like YouTube, Facebook, and MySpace. Secondly, you also have high-revenue/transactional sites like Monster, eHarmony, and PayPal. The former often have high pageviews but low(er) revenue streams, whereas the latter have directly monetizable activities.

It's interesting to examine these two groups in more detail. For every web property, there's a drive to maximize two things:

1. The total pageviews generated
2. The amount of revenue per pageview

Unfortunately, the two are often inversely correlated. Let's discuss why.

How do you generate high pageview counts?
What are internet activities that people do every day? It's a short list. Here are a couple broad categories I'd definitely put on the list:

• Communication: Email, IM, forums, etc.
• News: blogs, news content, etc.
• Entertainment: music, videos, casual games, etc.

For me, my daily activities include going to Gmail, reading my RSS feeds, and occasionally going to YouTube to check out some new videos.

Unfortunately, the very stickiness of these sites drives down the value of each pageview. The reason is that ultimately, the highest value users are the ones are the ones about to buy something online. The closer you are to the transaction, the more value it is. Thus, e-commerce, travel, mortgage sites, etc are all fantastic at driving high monetization.

When you're in the context to communicate or entertain, the furthest thing on your mind is to buy something. Thus, the CPMs on MySpace and Facebook are quite low, often way under $1 CPM.

How do you generate high revenue per pageview?
As mentioned earlier, the easiest way to generate revenue is to be close to the transaction. Another way to describe this is to capture "purchase intent." If you can grab people right before they are ready to buy, then you can be a gatekeeper for online revenue. This is why businesses like comparison shopping engines or eHarmony or Monster are great at pocketing money directly from each transaction.

Yet at the same time, being close to the transaction creates the inverse issue, where the site won't be something you use every day. Or you might use it in bursts, during a short transition point in your life (like job-hunting or finding a significant other), but nothing outside of that.

The clash of stickiness versus intent
So how can you reconcile this conflict between stickiness and intent? It's definitely very hard, and companies try to solve it in different ways.

When you're a site with minimal stickiness but great monetization, you end up trying to become more sticky by adding things like news or social features or entertainment.

When you're a site with great stickiness but minimal revenue, you end up trying to capture more intent by segmenting out pieces of your site to carve out commercial portions. That might look like adding a "shopping tab" or adding a search bar, etc.

Check out military.com for a site that successfully blends the two. The top bar has access to Travel, Shop, Finance, Careers, and Education, which are all monetizable things. Yet the stickiness of the site comes from the daily news and social functions.

Is search an exception?
One might say that search engines are actually an exception to the rule - actually they are not! Check out the top search queries across a bunch of different sites here. Of course, another source is the Google Zeitgeist.

What you'll find is that the majority of searches are actually not very useful for search engines. Things like "sex" or "britney spears" are not actually very valuable searches.

Instead, you can think of search as a service that costs money for 80% of queries, which are on things like news, celebrity gossip, and other random topics. It's the 20% of queries with commercially relevant value that also have advertisers bidding on the terms that drive all the revenue.

What's the right approach? Focus on eyeballs or focus on revenues?
Ultimately, it seems like there are two concrete ways to approach the issue if you want to build a billion dollar, venture-fundable business:

1. Build an eyeball company around a monetization engine
2. Build a monetization engine on a large, sweeping high-value consumer need
3. Build an eyeball company so huge that you win the category and get brand dollars

If you're going to build a high-pageview "eyeball" company, then you'll want to make sure you have a general sense on how to monetize it. Whether it's semi-regular lead generation or virtual goods or whatever, you'll mostly subsidizing social/news/entertainment activities with some revenue activities thrown in

If you're doing a monetization like Monster or eHarmony, just make sure you build it in a category where everyone cares about the issue. Then although you won't be sticky, you'll get enough people churning through the system that they hand you lots of money. Also, in this case, you probably don't want to use advertising dollars to support your business - charging by the transaction is probably more efficient, and people will likely pay up.

And finally, if you are hell bent to be one of the biggest websites in the world, you can eventually break into brand advertising through your sheer mass. But realize that this takes an enormous amount of venture funding to get to scale, and selling brand ads is like selling enterprise software - long sales cycles, relationship driven, and generally difficult.

Article in the NYT on games for older adults

Link: Video Games Conquer Retirees.

The first time she lost at Bookworm, Sister Jean-Marie Smith recalled, “I stood up and said, ‘Me and this computer are going to have a talk.’ ”

I really love the fact that folks in casual games are so successfully bringing the fun of video games to older folks. In particular, this is difficult for the reason that you're often designing games for a much different audience.

People creating websites would be smart to try to learn a bit from the people building these types of properties online.

Now just wait until you get a ClubPenguin for older adults going... MMOGs for seniors ;)

How do you find a badass co-founder?

Finding a co-founder is damn hard
In the last couple months, I've been keeping an open eye out on finding a high-quality co-founder for the startup I'm doing as part of my EIR gig. Ultimately, the scarcest commodity in the entrepreneurial community is NOT venture capital money - there are billions out there - but rather very high quality people. In particular, the highest quality people out there turn into co-founders, so that's incredibly important.

In particular, a co-founder's able to help balance you out, especially on mood. So if you are both in a room, the startup is on the rocks, and you say, "god we're fucked!" then sometimes your co-founder will say, "well, why don't we do X." The same will happen vice-versa, which is great.

How many co-founders?
2-3 founders maximum. I think once you get beyond that, you're diluting the group of talent in place. Ultimately, there's a huge distinction between founders and employees, and you have to choose carefully. Beyond 3, the equity structure gets messed up too - you take a round or two of VC money and you own a very small piece of the company.

Of course there are exceptions like VMWare, which had 6 co-founders that all did well. But the norm seems closer to 2-3.

What defines a good co-founder?
Short answer is, I have no idea :)

Long answer is, I've done a lot of talking and thinking about the issue, and I think I know what is good for me (and maybe me only). Ultimately, you are looking for a guy with the following:

Complimentary in skills, but from the same cloth in attitude and culture

On the skills front, because I'm more of a business-y person, I'm looking for someone who is very technical. Also, because I'm more of an unstructured creative thinker, it might be useful to meet someone who is more structured and detail-oriented. A big piece of this is also a Mr. Inside versus Mr. Outside designation. Who's in charge of talking to customers, partners, and potential investors? That might be one guy, whereas the other is more focused on internal operations. This might hold true even as the company scales up.

The other side, which is about attitude and values, is much more difficult. If you are looking to found a company, and you have an idea that you're driving, that says a lot of things about you already. You're probably driven, have a vision for where you want things to go, and are self-motivated enough to get things off the ground. You may also be someone who can convince people to follow you, or give you money, or whatever.

My questions for values/culture
For me, I've been thinking about a series of questions related to culture and values. Here are a selection of them:

• Let's say you wanted to start a new company? How would you do that?
• Tell me about a major disagreement you had recently - describe what happened?
• How would you approach hiring people?
• What's your long-term goal with your career? Where do you want to be in 20 yrs?

... and etc. Lots of questions you'd ask an employee, of course.

I think you'd also ask a couple questions as you observe the guy:

• If you put them in a room with 5 peers, would they emerge with the 5 guys signed up to follow them?
• Would you feel comfortable introducing them to everyone you know?
• If you say something they disagree with, how long does it take before they push back? How hard do they push back?
• If you guys disagree on their side of the complimentary skills, what happens? What happens if it's on your side of the domain expertise?

Peoples' views on this are going to be different, but in general I'm going to be looking for the guy who can sign up the 5 guys in a room, who's great to introduce to everyone at all levels, who pushes back hard and immediately, and doesn't care if its on your side of the skillset or theirs. I think all of these things define a strong leader who's a peer, rather than an employee.

I'll write more on this topic later, as it's a critical one, but would appreciate comments in the meantime.

Engaging travel site?

As a followup to my post on what's broken with online travel, a reader (thanks Gabe!) sent me a link to a site called coolcapitals.com. Try it out by clicking through, selecting a city, and navigating through the site a bit.

Overall, I like the general concept of making the experience stylized and fun. I *hate* the background noise and airport motif. That's exactly the stuff that people don't enjoy about traveling. I'd like to see more focus on pictures, more richness, and so on.

Another issue, of course, is that this product won't scale well since it's all being built by hand. I'd prefer a more automatic process for this, as a company.

Crazy Alexa growth for file upload sites

Has anyone been noticing the crazy rapid growth of all the file upload sites? I originally started seeing Rapidshare and Megaupload in the forum communities, but now it's spread all over the place.

Check out the Alexa stats for Rapidupload, Easy Share, SpeedyShare, Rapidshare, Quicksharing, Gigasize. Most of these are ranked ~1000, and Rapidshare is ranked #21 on Alexa. I wonder if somewhere in here is the next Photobucket.

Here are two examples:

Any ideas on what the business models could be for these sites, beyond AdSense? I think sticking with pure AdSense will lead to nice lifestyle businesses, but not venture fundable ones.

Quick comment on broad reach ad companies

Recently, Jeremy Liew of Lightspeed wrote a great blog called Three ways to build an online media business to $50m in revenue. In fact, it even received some coverage from the New York Times, which you can find here.

First off, let me encourage every Web 2.0 entrepreneur to read the blog for a sobering analysis. But let me state that the analysis merely shows the EXTREME endpoints, and not the typical situation.

Jeremy's calculations basically come down to dividing $50MM against a $1 CPM, a $5 CPM, and a $20 CPM. Here's an example:

1. Be a site with a broad reach (say general social networking, communications, news). At large scale, without a great deal of targeting possible, a startup’s “run of site” or “run of network” advertising might be able to get to the $1 RPM range (Revenue per thousand impressions, including CPM, CPC, and CPA models). To get to $50m in revenue you would need 50 billion pageviews in a year, or just over 4 billion per month. According to Comscore, Bebo had the 10th most Pageviews in the US in Janurary 1007, with 3.4bn, so you would need to be bigger than that.

Great thinking. But keep in mind that the typical site has inventory at $20, $5, and at $1 CPM, and your mix leans one way or the other depending on how contextually relevant you are. And oftentimes, the top 5% of inventory actually acounts for 40% of the revenue, because the $20 stuff is worth so much. The way this happens is that when you're a social networking site, you still start to carve off areas that seem contextually relevent enough.

So for example, you might carve off music or games or movies or whatever, and sell those at a high CPM. Or you could even artificially limit the # of ad positions on a place like your homepage, so that it commands a premium.

I think a lot of people are mistakenly taking the calculations as saying that if you have a Web 2.0 community, you'd have to be really huge to be worth anything. I think that's basically half true - you do have to be huge, but there are strategies to claw up into the higher CPMs.

What's broken with online travel?

Take a moment to think. When you think of the word "travel," what do you imagine?

I've asked a lot of people this, they typically say something like this:

When you ask them to elaborate, they say:

Aaaahhh. Beach. Relaxation. Escape. Fantasy. "Me time." Or, if you have a significant other, "Our time."

In fact, if you browse the Flickr pool for travel, you see a lot of beautiful pictures of exotic cities, breathtaking environments, etc. (If you have a moment, I encourage you to click and watch... it's really great!)

So taking a page out of my previous analysis on online dating using the MDA framework, you could ask the following questions:

• What emotions do you want people to feel, when it comes to online travel?
• What are desirable dynamics and behaviors, when it comes to those emotions?
• And finally, what do you need to build to accomplish these things?

I'd encourage you to read the previous posts on dating and the MDA framework, otherwise this post will make little sense :) Let's analyze the current stage of the art.

What emotions do Expedia/Travelocity/etc. convey?
Let's face it - Modern travel websites are all about logistics. What flight do you want? Where do you need to go? What dates? They're optimized for business travelers that need to get from Point A to Point B, without a big fuss. So when you think of Expedia, you can think about airports, planes, taxi rides, and all the things that get you from Point A to Point B.

And in a way, that's advantageous for them - it means that Expedia is about a transaction. The emotions that it conveys are about efficiency, cost effectiveness, and Getting Things Done. Which is great, when you're someone that needs to get from one place to the other.

But how many people, when you ask them about "travel," say they think of: Itineraries, cost effectiveness, efficiency, etc.? Let me argue, though, that for most consumers, travel is not about that. For most people, it's about getting away.

What emotions do you want to amplify, in a travel site?
As I mentioned before, my thesis is that most consumer travelers care most about the fantasy of travel, not the logistics. So ideally, you're looking to trigger emotions like:

• Wow, wouldn't it be great if... (Fantasizing)
• Ooooh, that's so pretty! I could just curl up and watch... (Relaxation)
• Honey, it'd be so fun if we... ("Our time")
• God I hate my job, I could really get away... (Escape)

If you had people looking at a series of pictures or imagining what they could be doing, then you have succeeded.

Like all experiences, "travel" starts earlier than when you board the plane. A very smart friend of mine, Kevin Lee of Y!, pointed out that the act of buying magazines and books to do research was part of the experience. Doing those things created the fantasy.

What dynamics do you want to create?
Ultimately, if you are succeeding in getting people to experience these emotions, you might argue that they'd do it all the time.

For example, with fantasy, they might create little vacations they probably wouldn't do, just for fun. For example, making an Antarctica trip. Or an African safari that was too expensive, just for the help of it. Or perhaps a girlfriend would plan a romantic getaway for a boyfriend, sending it out of the blue as a substitute for daydreaming.

For relaxation, maybe people would come to the site and make little trips just to soothe them and look at pretty pictures. If they were stressed out from work, maybe they'd come and check out other peoples' vacations, just to keep their mind off things.

if people are making little vacation slideshows and sending it around, or visiting regularly and planning random trips to all over the place, then you know you've really engaged them.

What mechanics drive the dynamics?
Now comes the part when you start analyzing the actual features and functionality that drive the dynamics and ultimately the aesthetics of the product. As an aside, it's clear that a lot of nerds start here when they start building products - it's easy to fall in love with technology that way. But when you do that, all you are doing is leaving the aesthetics to chance! They will automatically grow out of whatever features you put in place, and you may or may not understand why your site attracts an audience. (Or doesn't attract an audience)

Based on what I've written so far, I think it's clear that part of what is needed is some sort of immersive travel fantasy process. So you know that some sort of slideshow, or movie, possibly with music, would be great. You might also even place some avatars into the slideshow, to increase the fun. Video would work for this, as well. So you know that this is the end product.

You also need to help people constrain the fantasy, so that they can use it as a practical planning tool. That means when you show a picture of Angkor Wat, you'll want to link to Wikipedia information on it (or whatever). And you'll want to show a visual representation of the itinerary, maybe Indiana Jones style. Similarly, you'll want to take into account information like: how many days are you traveling? What region of the world? What's the budget? What's the schedule?

A skeleton of a travel product idea?
Rather than point out a problem and not try to fix it, I wanted to take a crack at the START of a conversation at fixing it. I'm not a travel industry insider, so I don't understand the logistics, but I think the "emotional mismatch" problem is a great one to try to solve. Particularly for a nerd, since it requires thinking about people rather than logistics and technology.

Just brainstorming randomly, perhaps you could have a site where you start out by selecting one picture out of several pairs, similar to what you do on LikeBetter.com. You could use this process to elicit travel preferences, such as:

• Do you like to travel to party? Or to relax?
• Do you prefer tourist traps or authentic holes-in-the-wall?
• Are you an independent traveler, or do you like to go in groups?
• ... etc.

You could then collect some more mundane information like, dates, budget, region of the world, etc. You'd try and make this as visual and engaging as possible, in alignment with the overall feel of the site. So region would be selected through a fancy map with pictures, or whatever.

Based on that information, with preferences and constraints in mind, the system would then generate a series of dreamy slideshows that describe potential itineraries. The user would flip through them, with the option to yay or nay, and the system would adapt. You could regenerate ideas over and over again, to see lots of variations. You could fade the edges of the slideshow, and have them come in and out slowly.

At the same time, you'd need to connect this "toy" with the real world. So under every picture, you'd need to let the user drill in deeper for information. Or combine one travel idea with another. Or understand the pricing differences, or constraints like schedule and so on. And of course, you'd want to have a "Buy" button.

And once you're done traveling, you can upload your pics onto the site and create the ACTUAL slideshow of what you really saw. Perhaps these pictures might get recycled into the entire system, for other people to use.

[Optional pet feature: You might have a little dotted line cutout for an avatar, where you can place you, your friends, boyfriend, or whoever might be suited for being part of the slideshow. Maybe the system could even generate little sayings with justifications like, "Wow this is sooo relaxing!" Ever play with an avatar system? It's fun... make one of you or a friend here.]

How do you measure success?
If a concept like this were successful, you could imagine that before Spring Break, college kids would be sending out travel slideshows like crazy to each other, trying to come up with the best ideas. Or a couple best friends now living far away might send slideshows to fantasize about relaxing together. Or a boyfriend might surprise his girlfriend by pitching her a vacation idea.

In essence, if people start to use the site for fantasizing rather than pure logistics, you know you were able to capture the aesthetics of travel better than what exists out there today.

Making user profiles more fun

Link: Imagini VisualDNA.

Just caught this on delicious even though it's a couple days old. Pretty fun way to capture data about the user - using pictures rather than form fields.

You could imagine an entire eHarmony or dating system built around this.

Archive of a couple older posts

I recently had a spate of new readers - if you guys haven't read through the blog, I wanted to pull out a couple of my older posts that might be interesting to peruse:

Advertising
Because I've spent a couple years in advertising at Revenue Science, I often like to write about random stuff happening in the online ad world.
How to create future value in online ads
The pyramid of advertising and modeling ad revenue
Standardizing ad units and PayPerPost

Company analysis
Sometimes I wrote about companies, and every once in a blue moon, I am right :)
User behavior challenges and BitTorrent
Business models for Federated Media
Are you misusing Alexa numbers? (Probably)
eHarmony's matching algorithm and other places to use it
Glam.com and building vertical networks

Meeting new people in Silicon Valley
Since moving down from Seattle 2.5 months ago, I've had to re-establish a network down here. It's been a lot of fun, and I've written down a couple observations.
Meeting lots of people through LinkedIn
Silicon Valley people versus Seattle people
10 tips for meeting people at conferences

Hope you like these, and if you ever have something you want an opinion on, write me at voodoo [at] gmail.

What's broken about online dating?

I had previously written about game design through the MDA framework. I wanted to analyze something more deeply - online dating is something I think is super broken, so let's look at it more closely.

Online dating is a maturing industry
The first fact about online dating is that it's a maturing industry. It started out growing exponentially, but it's now slowing down to single digit growth levels. It's down to around 9% now, according to CNN. Ouch. That's not what you want to see in an internet segment. I think that has to do with the limited zero-sum group that is really into "dating" sites, rather than social sites.

"Dating sites" primarily cater towards an older audience
Think about it: Bars are places for single people to hang out, and maybe you might meet someone there. Great, that's appealing to everyone! But when you start talking about the crowd that goes to "singles mixers" and "speed dating," you're primarily talking about a much older crowd that's looking to pair up. One issue here, obviously, is that as soon as you slap the "dating" label on something, you're automatically appealing to people ready to settle down rather than to hang out with people casually.

What are the aesthetics desired for dating?
So, going back to the MDA framework, you should start at the question of, what are the feelings and emotions you're trying to trigger when it comes to dating sites.

For the younger group, dating is about emotions that are fluttering, up and down, cat-and-mouse, etc. And after enough time, you like the person, you end up giving up pretenses and the shield drops. So a lot of the excitement of dating comes from the thrill of the chase, and then settling down into a real relationship.

For the older crowd, I'm guessing (since I'm not part of this group), that it's much more around safety, life-long matches that are highly personalized, and making it a comfortable experience. This group still wants a little bit of the cat-and-mouse, but they don't want to make it *too* exciting. I won't address this group much since I can't speak for the userbase.

What's wrong with online dating?
If you guys have ever used a typical dating site, it seems like the aesthetics are completely broken. By paging through a huge group of profiles, and then clicking to message them, the feeling is much more "transacitonal" in nature. It feels like you're buying a microwave from Walmart.com than it feels like a social experience.

It seems that "winking" and some in-between interactions are a good place to go to create emergent playfulness, which is where the thrill comes in. I love the fact that on Facebook, you can give an anonymous flower to someone. Or that on MySpace, you can send a message to a friend-of-a-friend asking about something random, and then the game of reading into the message can start.

Where's a good place to start fixing the problem?
So starting with the aesthetics I named above, I think the dynamics you want to create are ultimately around playfulness, mixed messages, push-pull, and all the other great cat-and-mouse games.

Here are some random ideas I've just brainstormed out:

• Remove the "dating" label from the site, but make it more about hanging out and being social.
• Make a very long ladder of interactions for people, from winking to poking to giving gifts to asking questions to open ended messaging.
• Give people an excuse to hang out OTHER than dating - be it casual games, chat, watching videos, etc.
• Set them up for dating success, even if you're not a dating site: Provide mixed gender balances in chat rooms, or group people by location and age, etc.

Ultimately, if you can get a gender balanced group of people who are all local to each other to talk and have fun, I think you're going to find that a lot of them will automatically start dating.