Ethics & Wonder/Amazon's Mechanical Turk/Kiva
Posted by Mike Damphousse on Sat, Nov 29, 2008 @ 08:05 AM
Having already experimented with outsourcing some database work to India, we thought we should give a try to Amazon's Mechanical Turk. From the FAQ page:
"Amazon Mechanical Turk is a marketplace for work that requires human intelligence. The Mechanical Turk service gives businesses access to a diverse, on-demand, scalable workforce and gives workers a selection of thousands of tasks to complete whenever it's convenient.
Amazon Mechanical Turk is based on the idea that there are still many things that human beings can do much more effectively than computers, such as identifying objects in a photo or video, performing data de-duplication, transcribing audio recordings, or researching data details. Traditionally, tasks like this have been accomplished by hiring a large temporary workforce (which is time consuming, expensive, and difficult to scale) or have gone undone."
So I decided to to load up a database project. The project is comprised of thousands of repetitive tasks and some research that anyone with basic business background can accomplish. We get these projects every week and have interns on staff to tackle them. I figured if an intern can do it, I'm sure some stranger a world apart can do it.
I proceeded to loaded the project which consisted of: look at the data, research a piece of the data, post the results. It probably takes 1-2 minutes per piece of data. Getting the project ready was fairly easy using the site's WYSIWYG form builder and data mapping. I posted a price of $0.15 per piece of data researched (in mturk terminology, each task is a Human Intelligence Task, or HIT). For this project there were 1,000 HITs so $150 was available. You hit "Publish" and it is live. 48 hours later, the project was complete. Interestingly, most of the work was completed by the same three people, yet an additional 5 participated in a smaller way. Also, most of the work was done in the middle of the night. The workers were most likely from the other side of the world. Total cost: $150 even and the workers invested 25 hours. That's $6 per hour. Comparitively, this would have taken one of my interns six or seven half days and cost me possibly more than twice as much. It's off-shoring for the masses. It works and it is cheaper, and the networking of multiple people get the tasks done much faster. Still checking on quality, but thus far the random spot check looked good.
That was the Wonder, now for the Ethics. Is what I paid for the work a fair price? If the person is working in India, for instance, the average white collar wage is $2 per hour. In the US, a white collar, non-managerial worker, earns $11 per hour. So I think my $7 per hour is fair, considering cost of living differences and the fact that the workers earned more than they could have working full time for a company. mTurk also gives them the ability to work their own schedule, be entreprenuerial, and in this case make more than they would have otherwise.
A friend asked me what my interns would do now that I off-shored their work. Well, I have some proactive marketing projects that might not have been funded had it not been for this project, so I'll assign it to them. The mturk workers win, my interns get more challenging work, and my company gets additional marketing. I think this is good for everyone.
I'm still not sold on the fair trade aspect of this, although the economics look fair. So as a feel good, I made another kiva.org micro-loan to a woman in Equador starting an egg business. I'm sure she will learn how to market with courage! Kiva is so addictive.