Live from AppDevCon in San Francisco

I’m here at AppDevCon in San Francisco, a gathering of Facebook developers, entrepreneurs, and investors hosted by SocialMedia‘s Seth Goldstein at the offices of Fenwick & West.

Today’s agenda:

12:30 – 1p: What is Engagement and why is it so important?
— Dave McClure (500 Hats)
— Seth Goldstein (SocialMedia)

1p: Creating, Spreading and Scaling Multi Million User Facebook Apps
— R. Tyler Ballance, (Slide)
— Blake Commagere (Vampires / Zombies)
— Dave Genztel, (SocialMedia)
— Jia Shen, (RockYou)
— Joe Winterhalter (Quizzes)
— James Hong (Hot or Not)

2p: Facebook Advertising Models
— Aryeh Goldsmith (Mybucks)
— Sourabh Niyogi (Appsaholic)
— Scott Rafer (Lookery)
— Narendra Rocherolle (fbExchange)
— Matt Sanchez (VideoEgg)
— Sundeep Ahuja (Appfuel)

3p: How to Value Facebook Apps?
— Eve Phillips, (Greylock)
— Keith Rabois, (Slide)
— Naval Ravikant, (Hitforge)
— Angela Strange, (AppFactory)
— Susan Wu, (CRV)

What is Engagement and why is it so important?

12:46pm: Dave McClure: Bake in a metrics system from day 1. Hypothesize the customer lifecycle and refine. Identify 5-10 conversion steps and put somebody on measuring it.

12:52pm: Dave McClure: Let’s not screw up this new ecosystem and fail to provide advertisers with effective metrics. We need some other usage metric than just users – like page views.

Creating, Spreading and Scaling Multi Million User Facebook Apps

1:04pm: Tyler Ballance (creator of Slide’s Top Friends, now with 12 million users): We’ve got a little social network inside your social network. Your top friends are the users you interact with the most. It’s really difficult to scale something after building it in your dorm room. Very difficult to keep persistent database connections open, very difficult to memcache very large amounts of data. Slide and RockYou have been able to do cool things because we’ve got caching servers. If you can afford an iframe, do it. If not, you’ve got FBML, but we all know that it breaks. They had the image caching collision issue where I had friends asking me why they showed up as furniture.

1:09pm: Jia Shen: We have to change things all the time. We’ve pretty much stayed away from all the caching stuff that Facebook offers. What we’ve done instead is build an infrastructure that allows us to regenerate all the profiles pretty quickly. They’re pushing FBML changes today and I’m just reading the post now.

1:19pm: Blake Commagere: Causes is an app with a lot of depth that I’m proud of. I also did Zombies, Vampires, and Werewolves, but they’re more like games. We think people want to compare themselves to others. That type of application is inherently more interesting. Collecting points for trivial activities within games – people love to get rewarded and compare. Let people show off what they’ve done. As you’re figuring out what you’re going to do with your apps, your database server is always going to be your constraint. We leverage memcache, we leverage Facebook’s caching. We chose Ruby on Rails because we could build our apps much faster with limited resources. Focus on your database and make sure it’s solid. We use MySQL on a machine with 32GB of RAM. Our app servers are cheap, but they do great. You don’t need 100 app servers to serve millions of users.

1:25pm: Joe Winterhalter: My inspiration came from what’s been successful on MySpace. Quizzes is about self expression. Create an incentive for you to have your friends involved. Most of our growth still comes from invites.

1:31pm: James Hong: We started out trying to create a bunch of different apps, now we’re focused on engagement. HotOrNot, HotLists, Pets, and Moods. Pets is driving millions of canvas page views a day. In our experience, database servers are not the bottleneck, Facebook’s proxying is. We’ve been using a lot of ajax. This does create issues if you want to refresh ads with ajax.

1:35pm: Dave Gentzel: Every single web page that I have made in the last 10 years that has taken less than 2 days to make has worked. Anything that has taken weeks or months has bombed. If your app doesn’t work well enough to succeed after a week, adding more complexity is not going to help. The massively successful apps will be simple, sticky apps that just work better. Apps that don’t involve others won’t work. Vampires was brilliant. Instead of inviting your friends, you’re “biting” them. Within Happy Hour, we do type-in invites instead of checkboxes because the people you type in are the people that are much more likely to pay attention to your request.

1:38pm: Jia Shen: News Feeds convert the best. It’s a black box, but if you can get there, you’ll convert. Do whatever you can do to get in the news feed. Requests and notifications are also good, but they convert less.

1:45pm: James Hong: Three things came to my attention this morning that make me think we’re on to something huge. 1) One of our applications that we spent a week building and haven’t touched since is making $1,000 a day with AdSense. You don’t need to have a big company to build an app to make a living. 2) Roelof Botha is in the room (from Sequoia, who invested in YouTube), and he’s here looking for the next big thing. 3) Bill Gates has rated 16 people on HotOrNot’s Facebook app (this is real). It feels like the early days of the Internet.

1:55pm: Advertising on Facebook – Making money with ads vs Protecting the user experience:

Blake Commagere: These kids don’t need to refinance their mortgage. If I’m just going to break even, I’m not that interested in it.

Dave Gentzel: Browsing MySpace is like browsing an ad-fucked website, so don’t go that far.

Jia Shen: Advertising on Facebook is ultimately not that different than advertising on the rest of the web, or MySpace. But your app is your baby, so it’s your choice. VideoEgg has some high CPM stuff. Do a bunch of A/B testing. I have heard rumors that Facebook is going to tax FQL queries.

James Hong: When you add ads there is some user leakage, but they’ll come back if you have a good product. If you leave for Dr. Pepper you’ll come back if you get messaged by a hot girl. I expect CPMs to go up 2x in the next year or so.

Tyler Ballance: We’re using our smaller apps as testbeds for advertising without sacrificing users on our big apps. You can do much more targeted advertising on Facebook because you know their network, etc. If ad networks can do it properly, it works much better.

Facebook Advertising Models

2:25pm: Scott Rafer: most correlations that ad networks could theoretically draw require tens of thousands of rules. The bulk of the Facebook crowd won’t be monetized all that much to begin with because they don’t spend that much. Most of the early monetization will be helping people do something like learn how to buy their first car, like Ford is doing now. Less profile inference and more not screwing up privacy concerns.

2:27pm: Matt Sanchez: We’re focused on bringing brands into social environments like the Facebook platform. We’re selling at $10 and sharing 60% with the developers. Early and need more advertisers. Selling in the UK, Canada, Australia, and the US.

2:30pm: Narendra Rocherolle: fbExchange is a hybrid between a text link network and a link exchange. The first step for new apps to get some exposure. Completely inventory constrained.

2:35pm: Aryeh Goldsmith: We started Matches and other successful apps and were about to start an ad network, when my partner said we shouldn’t because Facebook might do it. We have built a link exchange however. Now we’ve built AceBucks, a virtual currency. If you integrate a currency or points system within your app, your users now have more reason to come back and be involved. It also changes how people use your apps – whereas people wouldn’t pay $1 to do something in your app, they may pay $10 AceBucks which they can get for free by doing something else in your application.

2:39pm: Seth Goldstein: We launched “Monetize Your Apps” within Appsaholic last week. You can pick your app and ad format. We’re doing social advertising. We also have Promote Your Apps, and just this week started showing live bids. We are experimenting to create a quality score. You can buy clicks from us with PayPal.

2:44pm Seth Goldstein: Facebook says they want third party ad networks to exist, but that they may launch an advertising marketplace at some point themselves. There’s still a lot of ambiguity. Facebook understands that they could enable better behavioral networks than anyone else.

2:45pm Matt Sanchez: Media buyers want more simplicity. Age, gender, location.

2:48pm Sundeep Ahuja: We have to be careful around which data is private and what’s not.

2:52pm David Henderson: The young media buyers know about Facebook. But it’s still new. We have to give the value proposition clearly to buyers. People in agencies don’t understand what the Facebook Platform is. A major beer brand had a problem with the fact that we couldn’t guarantee that 70.1% of the audience was over 21 (can’t guarantee that in the feeds).

2:58pm Keith Rabois: People say they like transparency, but ultimately they really don’t.

3:04pm Lee Lorenzen: With Adonomics we’re hoping to measure all the way from installation to canvas page views. We encourage developers not to sell their apps for less than $1 per user. We think users are worth $500 to Facebook. If Facebook grows to 200 million users, I think they’re worth $100 billion. The social operating system is a big curveball that has been thrown at Google. It could be a chance for Microsoft to get back in the game.

3:10pm Roelof Botha: Brand marketers don’t innovate very quickly. Early search marketers on Google were not the Fortune 500. I encourage developers to create more measurable performance-based ad products because performance-based advertisers innovate more quickly.

How do Investors/Acquirers Value Facebook Applications?

3:22pm Angela: Web sites have spent more energy attracting user than a Facebook app has. Therefore, you could say that a website user is more valuable. At the same time, with a Facebook user comes their relationships on Facebook, so you could say that a Facebook user is more valuable. At Bay Partners, we value apps according to growth, category leadership, engagement (which is currently subjective, hopefully Facebook will help), team, and business model possibilities.

3:23pm Naval: I think Facebook users are worth more than website users. Most websites inflate their user numbers. The value of a Facebook user is going up – at first users were being sold for $0.10, now it’s $0.60-$0.70. TripAdvisor just bought Where I’ve Been for $1.50/user.

3:24pm Keith: Whoever is more upstream has more valuable users. App users are derivative of Facebook users. It just can’t be the case that Facebook users are more valuable than destination site users.

3:26pm Susan: It’s still so nascent, there’s still divergent opinions, it remains to be seen. I think it completely depends on the application. I’m a big investor in online games and virtual worlds, or synchronous social media. The type of engagement you get with synchronous social media is much different than asynchronous social media.

3:28pm Eve: The classic destination site is starting to disappear.

What business options do developers of successful apps have?

3:30pm Eve: Entrepreneurship is fundamentally about the entrepreneur. Some want to build a lifestyle business and others want to be more aggressive.

3:35pm Keith: There hasn’t been a technology company to go public without venture capital in the last 20 years. For owners of other destination sites Facebook is a tactical play. But I’ve looked at thousands of apps and I wouldn’t invest in any of them (in the VC sense).

3:40pm Susan: Resoundingly we prefer to invest in apps that could exist outside the Facebook user. It’s up to the CEO to determine the ROI of users in every channel – it completely depends on the relationship you have with your users. There are very few apps on Facebook today that would be good standalone businesses. There are inklings of micro-transaction or virtual currency platforms. There are more sophisticated business plans that are trying to create more persistent or new types of relationships with users.

3:43pm Naval: The apps that win aren’t necessarily going to feel more complex, they will feel more right. Also, I will take a contrarian view and actually bet on apps that will only live in Facebook. I think about it like an OS – would I invest in applications that only exist in Windows? Yes. Facebook however doesn’t want to let apps have out of band relationships with users.

3:46pm Roelof Botha: There is opportunity to take the torrent of users and built a sizeable business off the platform, but only a small number of firms will be successful in doing that.

What about other social networks?

3:49pm Angela: LinkedIn announced that they were opening their platform a week after Facebook did…in 9 months. Maybe they should make their API compatible with Facebook’s.

3:50pm Naval: I think it will be difficult for others to emulate the Facebook API, there are just different features. MySpace will take a more controlled approach. LinkedIn has a different demographic. This is Facebook’s game to lose.

Is there a market for liquidity for apps not at venture scale?

3:58pm Keith: We only buy developers. We hire them, not just apps.

3:59pm Angela: Users is just a baseline, we look at a lot of engagement factors.

4:00pm Naval: For those who said that Paypal didn’t get a good valuation because they were too dependent on eBay, we would all be happy with that outcome. It may not be enough for a half billion dollar VC fund, but for someone smaller there’s a lot of value to be created.

4:02pm Keith: Maybe in certain verticals, like travel or movies, you could sell those, but not What’s Your Stripper Name.

4:04pm Susan: There are segments where there are attractive exits for small development teams.

4:04pm Eve: Although there may be an arbitrage opportunity for individuals, it doesn’t make sense on the venture side, we have to look at value creation. We are willing to make seed level investments though.

4:05pm Naval: $15m exits can apply here.