When in Rome: Newcomers on Facebook

by Jesse Farmer on Sunday, April 5, 2009

A teammate of mine recently sent me a link to a paper called "Feed Me: Motivating Newcomer Contribution in Social Network Sites" and I thought it was worth discussing. The paper was jointly authored by Moira Burke, a PhD student at Carnegie Mellon, and Cameron Marlow and Thomas Lento, two research scientists at Facebook.

The Chicken and the Egg

The root question addressed in the paper is this: what motivates newcomers to contribute to social networks? For social networking sites getting users to contribute is one of the primary problems, right after how you acquire new users.

Let's dive right in and look at their hypotheses, methodology, and conclusions.


The authors took all users who joined on a random weekday in March 2008 — amounting to about 140,000 users — and tried to predict their long-term sharing habits based on the experiences they have in the first two weeks. Specifically, they looked at how users interacted with photos.

The paper outlines four hypotheses:

  1. Social learning: Newcomers whose friends share more content will go on to contribute more content themselves.
  2. Singling out: Newcomers who are singled out in content will contribute more content.
  3. Feedback: Newcomers receiving more feedback on their initial content will go on to contribute more content.
  4. Distribution: Newcomers whose initial content is distributed widely will go on to contribute more content.

Conclusion: When in Rome...

The authors also broke down the newcomers into two categories, early uploaders and non-early uploaders, depending on whether or not they uploaded more than one photo in the first two weeks.

The two factors that correlated with long-term photo sharing for early uploaders were whether your friends were also sharing photos in your firs two weeks, and whether people commented on your photos. Surprisingly "singling out," i.e., getting tagged in photos, had no statistically significant effect.

Singling out, however, did work for non-early uploaders, suggesting that people can be cajoled into uploading photos by tagging them, but that people already uploading photos to Facebook won't upload any more than they were before.

In short, newcomers are susceptible to peer pressure.

How is this useful?

The upside to this paper is that it gives a clear picture of what is worth measuring. Getting a user to upload a photo doesn't just mean one more photo on the site — some percentage of their friends will upload a photo, too.

What's more, you can enter into a sort of feedback loop. The paper didn't address whether "social learning" also correlated with increasing auxiliary activities like feedback, but imagine this: more photos uploaded means more comments, which in turn means more photos. Is it possible to make this cycle self-sustaining?

The downside is that this doesn't help with the chicken-and-egg problem. What happens when a user comes to the site and they have no friends? There are some public spaces on Facebook, but most social networking in that vein are dominated by interactions among friends.

Overall one of the most detailed papers analyzing data from a huge social networks. Leave a comment and let me know your thoughts, especially if you know any other papers of this kind!