Social Network App Study Details App Engagement Patterns

This is a guest post by Atif Nazir, a PhD candidate at the University of California, Davis in computer science. In late 2008, Atif’s group conducted a study on the usage of social network applications by developing several applications and social games, deploying them across Facebook, and measuring the results. We invited Atif to share his most interesting findings here – you can also find a link to his full paper below.

Over the past two years, the Facebook application economy has evolved into a multi-million dollar business. Developers have recently started shifting focus to social gaming applications, but a majority of social applications are still gifting applications. A research study I recently conducted provides some insights into the nature of user activity on on these two genres of applications on Facebook by considering user activity from three applications: Fighters’ Club, Got Love? and Hugged.

Of these, Fighters’ Club is a  gaming application, while the others are utility (gifting) applications. (Fighters’ Club and Got Love? are now owned by SGN.) The study uses Fighters’ Club as a representative of social gaming applications, while Hugged and Got Love? are typical gifting applications.

Our study found that a small number of users contribute most of the user-user interactions made through the applications, and that for gifting applications, about 80% of users respond to Facebook requests in the first 50 hours after a request is sent.

For games, a small fraction of users are responsible for most resource consumption (bandwidth, CPU) on application servers, and on average, the number of activities per user, total number of Facebook friends, as well as number of Facebook friends that use the same application  is more than four times higher on games than on gifting applications. This means that gamers tend to form Facebook friendships with other gamers, which enhances user engagement on games. Because of this, warm-up time for social games tends to be higher than for gifting applications.

Our study further observes that user data processing is a performance bottleneck for applications, but found that activity for more than 85% of users on social applications was inter-connected. However, on games, user activity is “more strongly” connected than on gifting applications (details), hinting at higher overall user engagement in gaming applications.

Moreover, our study did not find a strong correlation with user locality (countries, Facebook networks) and activity on applications. This means that simply segregating user data for more efficient/intelligent data processing is not possible by relatively intuitive metrics such as by country or Facebook networks.

The study provides a number of other interesting findings for developers and social network researchers. Details for all the findings can be found here.