fred wilson writes about amazon’s new HITs service which allows companies to pay people to provide useful data and content…it brings up the interesting question of which incentive a community driven service is appealing to ‘help the communty’ or ‘help myself’.
Amazon has built a service called Mechanical Turk, or mturk.com, that is like eBay for HITs.
If you need a human to do something, you send a HIT request (via the mturk API) to Amazon.
The HIT is displayed to the masses, who then complete them, and get paid for doing so.
This is an attempt to automate peer production and add a payment system on top of it.
i’m very skeptical even though it’s so cool to see. as fred points out, community contributed content relies on personal incentives to drive good results. in some cases, this is entirely self interest, and that usually works best – ebay and craigslist. in a few cases there seems to be a more ‘help the community’ ethos which can work like delicious.
at tribe, we launched a ‘pay it forward’ program a year ago. the idea, which i was really excited about, was that a lister could name a bounty to be paid to anyone who helped them get a transaction done. we experimented with jobs which already have examples of bounties. at the time, i hoped there could one day be a marketplace with attention driven by individuals forwarding listings to friends they knew were in the market for an apartment, job or car.
our experiment failed. we found that the number of forwards went up by 4x but the number of job applications did not. many tribe users told me they would forward a job to a friend whenever they saw one that might fit to help them out and that a bounty would not make them think of new friends or more likely to send to an existing one. in other words, they said this is something they do out of
‘help the community’ and not ‘help myself’.
my belief is that there is little to no crossover. i have also learned how hard it is to change consumer behavior. if you’re trying to build a peer produced system seems that you have to leverage existing behaviors driven by one of these two incentives and never bet on them changing. amazon’s new services will only work if they can tap into people’s ‘help myself’ incentive and i seriously doubt the dollar incentive will ever work. jobs was the highest bounty incentive we could find at as much as $5k and it failed.
i know there are also a number of startups pursuing this type of system specifically in jobs. i doubt this approach will work but will watch with great interest. maybe it’s a good idea that was poorly implemented by me, which would not be the first or last time that happened!