Some of the most promising new companies are in the business of doling out work, particularly work that requires almost no skills. Rising-star startups like Microtask and Crowdflower are being hired by other companies to break large projects into tiny repetitive tasks that are then distributed to a workforce across the globe. Incredibly cost effective, according to the New York Times Microtask charges approximately $0.0005 (or $1 per hour) with a single task taking about 2 minutes, and to ensure the accuracy, each task is distributed to several workers.
Though they perform the same service, the two startups crowdsource differently. Last week, Microtask announced that 25,000 volunteers had contributed to the company’s first major project, Digitalkoot – a project to digitalize Finland’s national library archives. Rather than source workers from a community of library enthusiasts, Microtask developed a videogame to utilize gamers. According to the New York Times, players of Digitalkoot type words that character recognition can’t identify, a task that helps moles build bridges in the game. In one month, volunteers have completed 2 million individual tasks over the course of 1,700 hours. Another source of labor used by Microtask is outsourced call centers. When the phone lines are slow, workers are turning their attention to completing Microtask assignments.
Crowdflower, on the other hand, boasts an on-demand labor force of 1 million. Crowdflower clients decide on the best channel for their project. Their work-force options range from the Crowdflower global workers, a Facebook game application, or Amazon’s Mechanical Turk which doesn’t assign workers tasks but rather lets them choose ones they would like to complete. Gauranteeing a 90% accuracy rate, Crowdflower manages its nebulous workforce by distributing correct and incorrect data into the work flow. After a project is completed, Crowdflower does a quality control check, weeding out workers who failed to identify the inaccurate data and retaining the rest.
With efficiency and cost effectiveness nearly guaranteed, the micro-sourcing trend is a hot investment – Crowdflower announced this month that it had received an additional 7 million in investments. But there is an obvious downside to micro-sourcing: like traditional outsourcing, the work could become exploitive. When large projects are reduced to tiny tasks, it is hard to image them as anything but mind-numbing. How can micro-sourcing ensure that its workers are receiving reasonable pay for the less-than-pleasurable work? Are more games such as Digitalkoot the answer? Should micro-sources expand on the method employed by Amazon’s Mechanical Turk, and let workers choose their tasks? What other ways can we employ micro-sourcing methodology without falling into traditional outsourcing exploitation?