Is Amazon’s Mechanical Turk a Big failure?
More than a year ago, at the first Seattle Mind Camp I attended a presentation (I believe that it was the first public one) of Amazon’s Mechanical Turk. At that time, it seemed like a great idea and we had a good sessions with lots of suggestions for usage of it. The service (you can see it here so I am not going to explain what it is) seemed to contain all the necessary features for a success, Web2.0 like features like crowdsourcing (or the Sawyer Effect as I recently heard from someone at work refer to it) and leverage of the huge infrastructure that Amazon already has in place. Few days ago, a former colleague send me a link tothis pretty funny (but mostly useless) application of Amazon’s Mechanical Turk. That prompted my curiosity about how Mechanical Turk is doing so I went to their site to see how many innovative applications people were building on top of it and to my surprise, there are only 31 HITs available and most of them seem to be generated by Amazon itself. Considering that each task on a HIT goes for a very low price in the pennies and that Amazon takes an even smaller cut of those profits, one might conclude that Amazon is just making few hundred dollars on the service, certainly not something to be proud of it. I have not seen too much advertisement of it so that might be a reason but since it certainly hit all the right Web2.0 places when it was launched, I am surprise that the usage of the application is so low. For this service to make any sense to a company of the size of Amazon (and to be useful to its users), you need to have millions of applications and they seem to be really far from it. I wonder how other similar Web services initiatives from Amazon like S3 or the recently launched EC2 are doing or going to do.
CD
