Facebook, College Degrees, Money, and Persuasion

(I posted this before on an internal blog at Schlumberger, and I just realized that if I removed one or two comments, it could be made public.

There is no moral to this story — I think it's just interesting that Facebook is the object of a college course, that it allowed some students to make money, and that a key issue is how people persuade their friends to install a new application. OK, now please excuse me while I go and reject one more invitation to be a "vampire" or to play "Texas Hold'em Poker.")

On December 12, while I was in California for a meeting of the SOA Consortium, Prof. BJ Fogg hosted the final presentations by the students in his Stanford Computer Science course CS377 entitled "Creating Engaging Facebook Applications."

As the name indicates, the focus was not on the technicalities of using the Facebook "platform" to create an app, but on how you make people (a) come back, (b) invite their friends. In other words, what influences virality? This is consistent with the fact that Fogg is an experimental psychologist by training, not a computer scientist. But regardless, I am fascinated by the topic, given the recent uptake of social networking in professional environments, including Schlumberger, which has about 4,000 employees on Facebook right now.

Prof. Fogg had asked me to be a judge (we happen to indirectly know each other), which gave me an obligation and a chance to look at the apps in more detail than I might otherwise have done.

He gave an introductory talk on "Persuasive Applications and Metrics: the Stanford Experience with Facebook." He said that he came to Stanford in 1993 as a psychologist, and since then he has been looking into how computers involve and persuade users. "Facebook is the #1 persuasive technology of 2007." It demonstrates "mass interpersonal persuasion." Could this principle then be used to entice people to solve larger problems? I noticed that he avoided the question of whether the same phenomenon might be used negatively.

The process followed by sucessul applications goes through these steps:

• Acquisition: get an address to which to send an invitation
• Activation: the target visits the site and registers
• Retention: the user keeps coming back
• Referral: the user invites friends to use the same application or site
• Revenue: some mechanism is used so that these visits generate money

At each step, you lose some people. The key to success is to maximize the conversion percentage from each stage to the next.

When Fogg announced the course, about 100 students showed up at the first session. After they understood the focus of the course, 73 remained, and they teamed up to create almost 30 apps. Those apps were, altogether, downloaded 10 million times within 10 weeks. 5 of the apps achieved more than one million users each, and were placed in the "Facebook 100" top downloads. 10 applications had more than 100,000 users, and 20 more than 5,000 users. At the end of the course, there were 925,000 daily users of one or more of the applications. Students tended to spend 20 hours per week on this course, even though it is only a 3- or 4-unit course that should occupy much less time.

Some students included a way to monetize their application (probably by placing sponsor ads) and made real money. Rumors has it that one team essentially got enough money to pay for this year's tuition, but this may be exaggerated.

Here are some remarks about the applications that struck me as... being worth remarking on. I'm passing on a lot of applications that are just variants on the "Poke" widget or the "SuperWall" widget in Facebook, which are redundant ways to say hello or send a photo.

• PhotoGraph: there are 4.1 billion photos on Facebook, so how can you search for anything? The app builds a collaborative filtering mechanism, where people who browse photos can select which of 3 proposed photos they want to see next. The app captures these links to "thread" the photos together and propose this sequence to others, so that navigation becomes more relevant.


• Polls: this allows people to poll their friends about any chosen topic. Usage was modest until they made the polls focused on rating your friends. In the end, this app was installed 61,000 times.


• SuperStatus: elevates the posting of status to a sort of "microblogging" by allowing the comments to be longer, and allowing crosstalk.


• SocialBuzz: this allows people to use their friends as sources of information and recommendations, but also to enter feedback on a recommendation. As a result, the app keeps a record of how much you trust each source in each area (you may trust someone'a advice about computers, but not about restaurants, or vice versa), and this affects the recommendations you see the next time. This seems very relevant to Knowledge Management in an enterprise, and I talked to the creators of this application after the presentation.


• OneVoice: allows people to "video jockey" YouTube videos.

An interesting remark by some students was that "Facebook is the most convenient and respectable way to stay connected with other people." This confirms to me the way Facebook has differentiated itself from MySpace over the last couple of years, with MySpace trending "down-market" while Facebook, true to its Harvard origin, took on a definite "college-educated" flavor.