Archive

Posts Tagged ‘research’

Will Facebook dominate Social Commerce?

February 13th, 2011

Facebook is a monster of a company, no doubt. It seems that world domination is its ambition, and it is well on its way towards that goal. But will it dominate social commerce, as it has so many other areas?

It turns out that social networks like Facebook “are not structurally well-suited to be networks of sellers,” according to a fascinating study done at Columbia (download pdf). The problem? Social networks like Facebook and Twitter end up being too clustered. When I click on a high school classmate from Hunter, I inevitably find that we have 60+ friends in common.

Why is this a problem? If you start traversing the social graph from my profile, you’ll likely to get stuck in a Hunter cul-de-sac. I love my classmates, but after a while, it’s a deadend that eventually gets boring. And empirically data from a social shopping site in France support this explanation.

The solution is to foster a network that looks more like a web than a cluster. More links are better, but you want to make sure that they take to different places. In my opinion, this structure can be induced through proper incentives, but it does not occur naturally. Herein lies the opportunity.

But it is likely to be tricky, because such “strategic attempts to alter a social network’s structure can lead to unintended consequences.” Unfortunately, the researchers don’t provide much detail into what alternative incentives are worth trying or what consequences they are concerned about.

Since the opportunity involves creating a new network, it seems to me that it is not only large but defensible. A difficult problem, perhaps, but what a nut if you can crack it! Given how different it is than building a global social network, I think it is one that Facebook is unlikely to crack by itself.

1. technology , , ,

Are we “a nation of cowards”?

February 21st, 2009

Charles Blow accuses most citizens of being biased against blacks in today’s NYT. Moreover, he accuses the vast majority as “cowards” because they are biased, but they won’t admit it. To support this accusation, he sites evidence from a controversial new measurement called the IAT, or “Implicit Attitude Test.” It measures how closely you associate concepts in terms of response time.

You can try the test out yourself on this site. It’s pretty crazy, and you will almost certainly come out with a slight bias against blacks no matter how hard you try.

But the question remains: does that measure have any real significance? Does it imply that we treat people with racial bias? These questions, applied to gender rather than race, are a central part of my dissertation.

In short, I find that implicit attitudes do matter, but it depends on the circumstances. Implicit attitudes affect behavior unconsciously, so they are indeed pernicious. When all else is equal, they probably influence decisions. But they do not overwhelm conscience decision making, or overwhelm more relevant information.

For example, race probably factors heavily in low information races. When you do not know about a minor local race (say, city council), you may unintentionally favor a white candidate over a black candidate. But in high information (e.g., the Presidency), race probably places only a small role.

I plan to post more about this topic in the future, but I wanted to react to this interesting item that appeared in today’s paper!

2. politics , , , ,

Olark Livehelp