(Editor's note: read this Sunday's "The Year In Ideas" issue of The New York Times Magazine to learn more about HP Labs social computing research in a section called "Social Media as Social Index")
After recently publishing work on content monetization and patterns of influence on Twitter, HP Senior Fellow and HP Labs Social Computing Research Group director Bernardo Huberman has turned his attention to explain the science behind a previously unquantifiable social media insight, which he calls ‘the paradox of cooperation’ in a new paper.
The paper starts by asking: why do people willingly contribute to crowd-created sites like YouTube, Digg, and Wikipedia when they can receive plenty of value from those sites without ever contributing a thing?
Apply a classic ‘tragedy of the commons’ frame to that question and it’s clear that the most rational thing to expect of everyone visiting the sites is for them to freeload. And yet people do contribute, making these ‘crowdsourced’ sites hugely successful.
Citing HP Labs research that analyzed tens of millions of Wikipedia edits, Diggs, Gnutella downloads, and YouTube videos, Huberman concludes that “the puzzle is explained by the fact that those contributing to the digital commons perceive it as a private good rather a public one.” And the private good they receive is attention.
Look at the data another way, though, and you get to another insight: that the more attention people get for their posts, the more they contribute. That’s an observation that has consequences for the people who own, run or want to create competitors for crowd-sourced media.
“While any user can contribute to these forums,” Huberman explains, “a disproportionately large percentage of the content is submitted by very active and devoted users, whose continuing participation is the key to the sites’ success.”
The takeaway: even though successful web sites like YouTube, Twitter, and Digg are built on the notion that anyone can contribute, they should be looking for ways to actively nurture their relatively few “superusers” if they want to remain successful into the future
Research by the numbers
Huberman’s paper cites research conducted by HP Labs (and others) analyzing such phenomena as peer-to-peer file sharing, YouTube, and Twitter.
Some of the raw data analyzed in this research:
- 3,100,464 files from 33,335 hosts on Gnutella (“Free riding on Gnutella”)
- 50 million edits to Wikipedia on 1.5 million articles (“Assessing the value of cooperation in Wikipedia”)
- 9,896,816 YouTube videos submitted by 579,471 users (“Crowdsourcing, attention, and productivity”)
- 59,853,763 Diggs made on 2,676,160 stories (“Feedback loops of attention in peer production” [PDF])
Why is HP conducting research on social media?
HP believes that information is becoming the greatest resource we have for addressing problems in business and society. Social media is increasingly becoming many people's interface to IT, and these media interactions produce an enormous amount of data. However, data isn't necessarily information as it contains a lot of noise. Creating software, hardware, and services that can automatically analyze enormous and noisy data sets to help people make informed decisions is an extremely challenging technical task and an area of focus at HP Labs.
Has HP published other social media research?
This is not the first report that Dr. Huberman and his fellow researchers have published on the subject of social media. Earlier this year, for example, Dr. Huberman and Dr. Asur at the Social Computing Lab released a report that found Twitter to be a surprisingly accurate predictor of the box office success of Hollywood film releases. For more of HP’s research in this area, visit the Social Computing Research Group.