Web 2.0 is People!
I’m at ETech where much of the discussion has centered around user attention, and the applications built through user contributions.
A number of people, including Clay Shirky have drawn a comparison between Wikis, blogs and other social applications and the 1973 movie Soylent Green in which Charlton Heston makes the horrid discovery that Soylent Green is people. audio. When Wink first started referring to “People Powered Search” internally before we launched, we joked about the Soylent Green comparison.
One might say that the people in Soylent Green made the ultimate commitment to “user generated content”.
There’s another aspect to the movie that is perhaps a better analogy. Since resources such as energy and food are scarce, people have to power their own homes. In the famous scene where Edward G. Robinson tells Charleton Heston what it was like in the old days, he is keeping the lights on by riding a stationary cycle. (Of course one wonders at the wisdow or efficiency of expending the human energy derived from a few green biscuits on keeping a 100 W bulb lit.). The Matrix approach to humans-as-batteries is so much more efficient from all angles.
That is an example of a system “powered” by people, rather than made of people. On Monday evening Bruce Sterling referred to the “Mechanical Turk” - an 18th century chess-playing machine that was later discovered to have a person inside, and only played chess by direct action of the person inside.
There’s was discussion last November about Amazon’s new Mechanical Turk project to have humans act as Web services, doing tasks that humans find easy but are difficult for users. Now that’s “people powered”.
The Mechanical Turk
A more useful view of people powered applications is when people contribute to a social application voluntarily. There’s some intrinsic personal motive (like keeping track of a site they’ve found for social bookmarking, or getting better movie recommendations, for movie reviews), coupled with other motives such as a sense of contributing to the whole, or some attention for their efforts, such as the effort to get to the top of the listings on Digg or Memeorandum.
In social search there is a hope that the collective efforts - and collective discoveries - of users can lead to a better result. The idea that tagged Web pages could add to a search result is becoming popularized. There’s considerable value in a tagged page.
Think of the search process. You start with a term that describes what you’re looking for. Say “diving Thailand“. You look at search results, filtering through pages that aren’t what you’re looking for - hotel ads, sites that haven’t been updated in 5 years, etc. - and modifying your search term. Say you realize that SCUBA Thailand is a better search. Or perhaps you realize you want to dive on the West side of southern Thailand. You might refine your search to diving Phuket, or “diving Koh Lanta”. Or you might decide that a particular dive site is the best for you, such as Hin Mouang. The other thing you are doing is following promising links along the way.
After you’ve done all this reading you don’t want to forget the site you’ve discovered. You bookmark it with your favorite service, using tags such as “diving Thailand” and “Hin Mouang”, so you can get back to it, or email it to a friend. There’s considerable effort embedded in that bookmark. You applied human heuristics and pattern matching to narrow down your search to the best site you could find, and then labeled it. The fact that you labeled this site, and that tens of thousands of others have tagged millions of sites, is strong evidence of the relevance of those sites, and potentially time saving tools for others.
Further, if there is something new, such as the discovery of a new reef, which actually happened recently in Thailand. The results can adapt quickly as people discover, tag and block that fact. Or if people want to know if dive operators are back to business as usual after the Tsunami this is information that is conveyed through user reviews and blogs, and not necessarily on the home pages of commercial diving Web pages.
Others who come along searching for diving in Thailand will have the benefit of the effort expended by the first person. The system can give credit to the first person, or can even allow him to assemble a collection of resources on Diving Thailand. So it’s people doing what they’re good at to help the machine.
I’m still looking for an analogy that better describes this kind of constructive user input than Soylent Green does.
Tags: people powered, social applications, soylent green, mechanical turk, Wink, social search
Comments are closed.