As we initially wrote in September, Stanford University students had a unique opportunity to take a cutting-edge class this fall: “Create Engaging Web Applications Using Metrics and Learning on Facebook.” The class, taught by Dr. BJ Fogg of Stanford’s Persuasive Technology Lab with the assistance of Dave McClure, quickly swelled in ranks as students jumped at the chance to learn how to build and market Facebook applications.
While the quarter is now wrapping up, two teams in the class have achieved unexpected levels of success: both KissMe (by Joel Darnauer, Eduardo Abeliuk, and Chris Mocko) and Send Hotness (by Joachim De Lombaert, Ed Baker, and Alex Onsager) have been installed by over 1 million Facebook users in under 30 days, each with over 100,000 daily active users (in fact, Send Hotness just reached 2 million installations and currently has over 300,000 daily active users).
We spoke with KissMe team member & Stanford undergrad Chris Mocko and Send Hotness team member & Stanford GSB student Ed Baker about their experiences in CS 377W and the success of their apps.
IF: Why did you decide to take this class? What was your background with social web apps?
CM: Throughout the summer, I never really had too much interest in Facebook applications and only supported the applications because of my addiction to Scrabulous. I did not comprehend the possibilities made available by the Facebook API, nor did I have any experience in other social web apps and widgets. I did know, however, that when I received an e-mail from my department mailing list announcing the class, I had to do anything, even drop a course required for my major, to be a part of the class.
I did not come in with much experience in web development, but I had a very strong interest in learning everything and anything about the subject, and what better way to learn than by pairing up with the smartest students in the country! I have been fortunate to be paired with two excellent partners, Joel Darnauer and Eduardo Abeliuk, who have been essential in setting up the servers and the database and implementing the code.
EB: I decided to take this class partially just because it sounded cool, but also because I was impressed with the professors and with the other students in the class. I helped virally tune a Facebook app called Compare People over the summer, and as a result they grew from a couple thousand users to millions of users.
I have been practicing viral tuning for several years now and have been amazed by the viral metrics of Facebook apps versus stand-alone sites. Not only is it easier to get a higher viral factor, but the periods are much shorter, which result in an even higher “effective” viral factor. Alex and Joachim had built a really cool Graffiti app before taking this class, so I knew they’d be good partners to team up with.
IF: What metrics of your success are you the most proud of, and why? What did you do (tactics, strategies, clever tricks) to get there?
CM: We are most proud of reaching one million users. We remember during the first weekend after our application’s launch being thrilled when we reached 500 users. To think that we could reach one million installs in a month, with no prior experience and no investment to “buy” installs, boggles our minds. A metric that we are currently tracking and very intrigued with is our daily active users because this number stagnated for a couple of days, but is currently on the rise again! Getting users is one thing, which we owe to the application’s inherent virality, but keeping these users engaged is a much more challenging task.
We attribute the majority of our success to KissMe’s simplicity. The name KissMe is simple but memorable, descriptive, and provocative and the application itself is basic, but intuitive to use. Too many applications out there have lengthy instructions that bore and confuse the user, and too many actions that force the user to think and choose. Another great feature of KissMe is that by sending a kiss, the user is inviting a friend to join the application without making the user feel like he is spamming a friend with another “invite” message.
EB: I think I’m most proud that we were able to get 1 million installs in less than three weeks. We didn’t use any specific “clever tricks”, but it basically all comes down to math formulas. It’s a mix of art and science — we are very scientific about the way we track our viral growth metrics, and we try to be more “artistic” about the way we actually affect those metrics. The face that Joachim and Alex have been able to iterate quickly on new ideas has also been key.
Joachim and Alex: We’re thrilled by the rate at which our application has grown. Joachim and I treated this very differently than Graffiti – rather than building every feature we could think of, an area we focused on instead was acquiring as many users as possible. We hit a million users this past weekend, and it looks like we’re on track to hit 2 million by this coming weekend.
IF: What do you think were the most important points you learned in class?
EB: BJ and Dave have been great professors. I think the most important thing they have emphasized in the class is how you should keep your app simple in order to make it as easy as possible to get users engaged. We have also really enjoyed all of the guest appearances from experienced entrepreneurs in the Valley who have given our class some good tips.
CM: The lectures have been very different than what we expected. Since this course is listed in the computer science department, we assumed each lecture would have a technical focus. In reality, the lectures have instead been devoted to teaching the class the fundamentals of building engaging applications so that we can apply our new-found knowledge to a plethora of other disciplines. BJ has not given us the secret formula to creating a great Facebook application, but rather promoted a series of procedures that lead to user engagement in any field. One of the useful strategies BJ has discussed is getting qualitative feedback. While delving into the numbers that our database and Google Analytics provides is quite useful, BJ promotes interacting with users to discover problems, hear suggestions, and define a focus. We often assume that we share the same opinions as all Facebook users, but of course the surveys and pair-wise testing we have done prove otherwise.
Dave McClure has been an unbelievable mentor as a man who has been through nearly everything and knows almost everyone. His AARRR model (Acquisition, Activation, Retention, Referral, and Revenue—see 500 Hats for more!) has greatly influenced how we measure our success and develop our application. Just as important, though, every lecture and lab session he brings with him his portfolio of entrepreneurial success that has provided great guidance for our group especially as we decide on the future of KissMe.
IF: What were the key mistakes you made early on, and what did you do differently as a result?
EB: When we were first brainstorming app ideas, we came up with some ideas that we thought would be cool, but they were much more complex. These apps would have taken longer to develop, and because of their complexity, it probably also would have been more difficult to get users engaged.
CM: Our biggest mistake was not thinking down the road. We had no idea that our application could and would grow so quickly, and we did not plan accordingly. The scaling issues we experienced were far greater than we expected, and for about a week we were running into server problems and optimization issues with our SQL code which slowed down the site, and on occasion shut it down completely. Now that we have migrated our code to Amazon’s EC2 servers, we think most of our scaling issues will be easier to fix down the road.
IF: What are your goals with the app after the class ends?
CM: Now that we have found some initial success with our application, we are definitely interested in developing more social network applications in the future in addition to improving KissMe. We are not sure whether we are going to turn our unofficial partnership into a business or try to sell our application to the highest bidder, but this is something that is definitely on our minds right now as the quarter wraps up.
EB: We don’t want to stop at one app. We have ideas for several more apps that we hope to implement in the next few weeks. Please stay tuned!
IF: Have you spoken with any potential investors?
CM: We have had a couple of meetings with investors and some unofficial discussions with VCs, but we have not formulized anything yet. At first, we were most interested in creating a viral app and establishing a large user base, knowing that more parties would become interested as the KissMe brand became more familiar. Now that we have exceeded expectations, it is time to turn from a purely technical focus to a more business-centered approach. Thankfully, some of our teachers and classmates have been through similar experiences already, and should be able to provide us with quality, trustworthy advice.
EB: Fortunately our revenues already far exceed our costs, so we don’t think we need to raise any money at this point in time.
Finally, I asked BJ Fogg to share his thoughts and reflections on the first edition of this pioneering class.
BJ: When planning the course, Dave and I wondered if any team would get more than 100K installs during the quarter (students would have only six weeks to do so). We thought 100K would be amazing. Now two teams have over 1M installs in about four weeks. And it seems that half a dozen teams may break 100K installs soon.
Creating virable apps for Facebook is definitely “learnable.” Our class doesn’t just have one hit app. We have these two big apps and more successes ramping up. So I believe people can learn to create viral apps; it’s not pure luck.
The key to success hasn’t been my expertise or teaching style. Instead, the successes come from giving motivated students excellent resources. This includes the many people who have helped with the class this quarter. We have lots of folks to thank — that’s for sure.