Hacker News new | past | comments | ask | show | jobs | submit login
Duncan Watts vs. Malcolm Gladwell: Mathematician turned sociologist challenges "Tipping Point" idea. (fastcompany.com)
28 points by fiaz on Feb 19, 2008 | hide | past | favorite | 7 comments



I think it greatly depends on the market rather or not influencer's have an impact. The average person isn't buying toothpaste becuase someone famous is, but they do drive escalades and hummers because they saw it done by "influencer's"


"The average person isn't buying toothpaste becuase someone famous is, but they do drive escalades and hummers because they saw it done by "influencer's"''

I think Watts' point is that the influencers are not part of some special core sect of mavens and connectors, but simply people within one or another social sphere who, for whatever reason, happen to get attention. It's more random that Gladwell suggests.

(I tried reading the book, but just didn't have the attention span and it went back to the library before I finished it. I think I skim-read about half of it.)


Watts book Six Degrees is pretty interesting to read. There's a lot of good material in it that will leave you with lots to think about. It's a bit dense in places and could have used a better ordering of the materials, but I'd still recommend the book.


One thing Watts does not account for in his simulation ... in real life, influentials are disproportionately likely to be connected to other influentials.

And just empirically we can see the influentials do have an effect. No one can really deny the power of celebrity endorsements. Now what can be questioned, though, is just how many connections do you need before you can be considered an influential, and what magnitude is their influence? From Watt's experiment we can probably conclude that 40 is not enough to do a lot.


Bit of an ego blow to those styling themselves as "influentials."


'Bit of an ego blow to those styling themselves as "influentials."'

I wonder if Galdwell's conjecture gained traction because, deep down, most people really like to believe that some people are simply special.


Watts' claim that trend-spreading data doesn't show the influentials pattern certainly seems plausible. The article places naive emphasis on his models, but those can be tweaked any which way. The really impressive finding is his repeat of Milgram's degrees-of-separation experiment.

I was going to add the music-ranking experiment to this list, but on second thought that one (while interesting) doesn't address the influentials theory at all.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: